rpms/systemd
6
0
Fork 0
Code Issues Pull Requests Releases Activity

import systemd-239-74.el8_8

Browse Source
This commit is contained in:
CentOS Sources 2023-05-16 06:14:02 +00:00 committed by Stepan Oksanichenko
parent d8ae20ffcd
commit 020374c810
75 changed files with 5989 additions and 39 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

59
SOURCES/0832-ci-lint-add-shell-linter-Differential-ShellCheck.patch Normal file
View File

@ -0,0 +1,59 @@
From 5f69ba3919d32ed93c68bb6b8b70a516f2bb56a8 Mon Sep 17 00:00:00 2001
From: Jan Macku <jamacku@redhat.com>
Date: Tue, 16 Aug 2022 14:34:49 +0200
Subject: [PATCH] ci(lint): add shell linter - Differential ShellCheck
It performs differential ShellCheck scans and report results directly in
pull request.
documentation:
https://github.com/redhat-plumbers-in-action/differential-shellcheck
(inspired by commit
https://github.com/systemd/systemd/commit/3f3c718e79abdac698ae90de5cd4c0560a0a75d4)
RHEL-only
Related: #2122499
---
.github/workflows/differential-shellcheck.yml | 31 +++++++++++++++++++
1 file changed, 31 insertions(+)
create mode 100644 .github/workflows/differential-shellcheck.yml
diff --git a/.github/workflows/differential-shellcheck.yml b/.github/workflows/differential-shellcheck.yml
new file mode 100644
index 0000000000..fa94679b51
--- /dev/null
+++ b/.github/workflows/differential-shellcheck.yml
@@ -0,0 +1,31 @@
+---
+# https://github.com/redhat-plumbers-in-action/differential-shellcheck#readme
+
+name: Differential ShellCheck
+on:
+ pull_request:
+ branches:
+ - master
+ - rhel-8.*.0
+
+permissions:
+ contents: read
+
+jobs:
+ lint:
+ runs-on: ubuntu-latest
+
+ permissions:
+ security-events: write
+ pull-requests: write
+
+ steps:
+ - name: Repository checkout
+ uses: actions/checkout@v3
+ with:
+ fetch-depth: 0
+
+ - name: Differential ShellCheck
+ uses: redhat-plumbers-in-action/differential-shellcheck@v3
+ with:
+ token: ${{ secrets.GITHUB_TOKEN }}

34
SOURCES/0833-meson-do-not-compare-objects-of-different-types.patch Normal file
View File

@ -0,0 +1,34 @@
From deb09b3bd826571149f6b018f3a3ff8a33cd104b Mon Sep 17 00:00:00 2001
From: Yu Watanabe <watanabe.yu+github@gmail.com>
Date: Thu, 28 Jun 2018 16:09:04 +0900
Subject: [PATCH] meson: do not compare objects of different types
This fixes the following warning:
```
meson.build:1140: WARNING: Trying to compare values of different types (DependencyHolder, list) using !=.
The result of this is undefined and will become a hard error in a future Meson release.
```
Follow-up for f02582f69fe1e7663a87ba80bd4f90d5d23ee75f(#9410).
(cherry picked from commit 48f5da19b6e8f0d05f5217bc9856093d354ce5d0)
Related: #2122499
---
meson.build | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/meson.build b/meson.build
index 6729a9ea5e..af4cf331da 100644
--- a/meson.build
+++ b/meson.build
@@ -1165,7 +1165,8 @@ substs.set('DEFAULT_DNSSEC_MODE', default_dnssec)
dns_over_tls = get_option('dns-over-tls')
if dns_over_tls != 'false'
- have = libgnutls != [] and libgnutls.version().version_compare('>=3.5.3')
+ have = (conf.get('HAVE_GNUTLS') == 1 and
+ libgnutls.version().version_compare('>=3.5.3'))
if dns_over_tls == 'true' and not have
error('DNS-over-TLS support was requested, but dependencies are not available')
endif

48
SOURCES/0834-journal-remote-use-MHD_HTTP_CONTENT_TOO_LARGE-as-MHD.patch Normal file
View File

@ -0,0 +1,48 @@
From ea9b3a664f5e67d0ee6b0bf6ca362835ae11fedc Mon Sep 17 00:00:00 2001
From: Yu Watanabe <watanabe.yu+github@gmail.com>
Date: Mon, 20 Dec 2021 20:48:32 +0900
Subject: [PATCH] journal-remote: use MHD_HTTP_CONTENT_TOO_LARGE as
MHD_HTTP_PAYLOAD_TOO_LARGE is deprecated since 0.9.74
(cherry picked from commit 30df858f43b14a55c6650b43bea12cbf2cc0bc67)
Related: #2122499
---
src/journal-remote/journal-remote-main.c | 2 +-
src/journal-remote/microhttpd-util.h | 10 +++++++---
2 files changed, 8 insertions(+), 4 deletions(-)
diff --git a/src/journal-remote/journal-remote-main.c b/src/journal-remote/journal-remote-main.c
index 47fe9d7433..bcaa370099 100644
--- a/src/journal-remote/journal-remote-main.c
+++ b/src/journal-remote/journal-remote-main.c
@@ -304,7 +304,7 @@ static int request_handler(
/* When serialized, an entry of maximum size might be slightly larger,
* so this does not correspond exactly to the limit in journald. Oh well.
*/
- return mhd_respondf(connection, 0, MHD_HTTP_PAYLOAD_TOO_LARGE,
+ return mhd_respondf(connection, 0, MHD_HTTP_CONTENT_TOO_LARGE,
"Payload larger than maximum size of %u bytes", ENTRY_SIZE_MAX);
}
diff --git a/src/journal-remote/microhttpd-util.h b/src/journal-remote/microhttpd-util.h
index 26909082a1..dd0ca1d9bd 100644
--- a/src/journal-remote/microhttpd-util.h
+++ b/src/journal-remote/microhttpd-util.h
@@ -38,9 +38,13 @@
# define MHD_HTTP_NOT_ACCEPTABLE MHD_HTTP_METHOD_NOT_ACCEPTABLE
#endif
-/* Renamed in µhttpd 0.9.53 */
-#ifndef MHD_HTTP_PAYLOAD_TOO_LARGE
-# define MHD_HTTP_PAYLOAD_TOO_LARGE MHD_HTTP_REQUEST_ENTITY_TOO_LARGE
+/* Renamed in µhttpd 0.9.74 (8c644fc1f4d498ea489add8d40a68f5d3e5899fa) */
+#ifndef MHD_HTTP_CONTENT_TOO_LARGE
+# ifdef MHD_HTTP_PAYLOAD_TOO_LARGE
+# define MHD_HTTP_CONTENT_TOO_LARGE MHD_HTTP_PAYLOAD_TOO_LARGE /* 0.9.53 or newer */
+# else
+# define MHD_HTTP_CONTENT_TOO_LARGE MHD_HTTP_REQUEST_ENTITY_TOO_LARGE
+# endif
#endif
#if MHD_VERSION < 0x00094203

71
SOURCES/0835-Fix-build-with-httpd-0.9.71.patch Normal file
View File

@ -0,0 +1,71 @@
From ca86de228e19cea268ec3eeabc9097d7c28fbf24 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Zbigniew=20J=C4=99drzejewski-Szmek?= <zbyszek@in.waw.pl>
Date: Tue, 30 Jun 2020 09:56:10 +0200
Subject: [PATCH] =?UTF-8?q?Fix=20build=20with=20=C2=B5httpd=200.9.71?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
The return type of callbacks was changed from int to an enum.
(cherry picked from commit d17eabb1052e7c8c432331a7a782845e36164f01)
Related: #2122499
---
src/journal-remote/journal-gatewayd.c | 4 ++--
src/journal-remote/journal-remote-main.c | 2 +-
src/journal-remote/microhttpd-util.h | 6 ++++++
3 files changed, 9 insertions(+), 3 deletions(-)
diff --git a/src/journal-remote/journal-gatewayd.c b/src/journal-remote/journal-gatewayd.c
index 54446ff7b5..3ff05a4d72 100644
--- a/src/journal-remote/journal-gatewayd.c
+++ b/src/journal-remote/journal-gatewayd.c
@@ -338,7 +338,7 @@ static int request_parse_range(
return 0;
}
-static int request_parse_arguments_iterator(
+static mhd_result request_parse_arguments_iterator(
void *cls,
enum MHD_ValueKind kind,
const char *key,
@@ -795,7 +795,7 @@ static int request_handler_machine(
return MHD_queue_response(connection, MHD_HTTP_OK, response);
}
-static int request_handler(
+static mhd_result request_handler(
void *cls,
struct MHD_Connection *connection,
const char *url,
diff --git a/src/journal-remote/journal-remote-main.c b/src/journal-remote/journal-remote-main.c
index bcaa370099..a1008db6eb 100644
--- a/src/journal-remote/journal-remote-main.c
+++ b/src/journal-remote/journal-remote-main.c
@@ -241,7 +241,7 @@ static int process_http_upload(
return mhd_respond(connection, MHD_HTTP_ACCEPTED, "OK.");
};
-static int request_handler(
+static mhd_result request_handler(
void *cls,
struct MHD_Connection *connection,
const char *url,
diff --git a/src/journal-remote/microhttpd-util.h b/src/journal-remote/microhttpd-util.h
index dd0ca1d9bd..792c07ac20 100644
--- a/src/journal-remote/microhttpd-util.h
+++ b/src/journal-remote/microhttpd-util.h
@@ -51,6 +51,12 @@
# define MHD_create_response_from_fd_at_offset64 MHD_create_response_from_fd_at_offset
#endif
+#if MHD_VERSION >= 0x00097002
+# define mhd_result enum MHD_Result
+#else
+# define mhd_result int
+#endif
+
void microhttpd_logger(void *arg, const char *fmt, va_list ap) _printf_(2, 0);
/* respond_oom() must be usable with return, hence this form. */

303
SOURCES/0836-ci-replace-LGTM-with-CodeQL.patch Normal file
View File

@ -0,0 +1,303 @@
From 2b1dbcab1af1a22f3a46fa23aa551a7394673938 Mon Sep 17 00:00:00 2001
From: Frantisek Sumsal <frantisek@sumsal.cz>
Date: Thu, 15 Sep 2022 15:29:23 +0200
Subject: [PATCH] ci: replace LGTM with CodeQL
As LGTM is going to be shut down by EOY, let's use CodeQL instead.
This is loosely based on upstream's CodeQL configs with some minor
tweaks to avoid backporting tons of unrelated commits.
rhel-only
Related: #2122499
---
.github/codeql-config.yml | 12 ++++
.github/codeql-custom.qls | 44 ++++++++++++
.../PotentiallyDangerousFunction.ql | 3 +
.../UninitializedVariableWithCleanup.ql | 16 ++---
.github/codeql-queries/qlpack.yml | 11 +++
.github/workflows/codeql.yml | 68 +++++++++++++++++++
.lgtm.yml | 37 ----------
7 files changed, 146 insertions(+), 45 deletions(-)
create mode 100644 .github/codeql-config.yml
create mode 100644 .github/codeql-custom.qls
rename {.lgtm/cpp-queries => .github/codeql-queries}/PotentiallyDangerousFunction.ql (93%)
rename {.lgtm/cpp-queries => .github/codeql-queries}/UninitializedVariableWithCleanup.ql (86%)
create mode 100644 .github/codeql-queries/qlpack.yml
create mode 100644 .github/workflows/codeql.yml
delete mode 100644 .lgtm.yml
diff --git a/.github/codeql-config.yml b/.github/codeql-config.yml
new file mode 100644
index 0000000000..7c01d32caa
--- /dev/null
+++ b/.github/codeql-config.yml
@@ -0,0 +1,12 @@
+---
+# vi: ts=2 sw=2 et:
+# SPDX-License-Identifier: LGPL-2.1-or-later
+name: "CodeQL config"
+
+disable-default-queries: false
+
+queries:
+ - name: Enable possibly useful queries which are disabled by default
+ uses: ./.github/codeql-custom.qls
+ - name: systemd-specific CodeQL queries
+ uses: ./.github/codeql-queries/
diff --git a/.github/codeql-custom.qls b/.github/codeql-custom.qls
new file mode 100644
index 0000000000..d35fbe3114
--- /dev/null
+++ b/.github/codeql-custom.qls
@@ -0,0 +1,44 @@
+---
+# vi: ts=2 sw=2 et syntax=yaml:
+# SPDX-License-Identifier: LGPL-2.1-or-later
+#
+# Note: it is not recommended to directly reference the respective queries from
+# the github/codeql repository, so we have to "dance" around it using
+# a custom QL suite
+# See:
+# - https://docs.github.com/en/code-security/code-scanning/automatically-scanning-your-code-for-vulnerabilities-and-errors/configuring-code-scanning#running-additional-queries
+# - https://github.com/github/codeql-action/issues/430#issuecomment-806092120
+# - https://codeql.github.com/docs/codeql-cli/creating-codeql-query-suites/
+
+# Note: the codeql/<lang>-queries pack name can be found in the CodeQL repo[0]
+# in <lang>/ql/src/qlpack.yml. The respective codeql-suites are then
+# under <lang>/ql/src/codeql-suites/.
+#
+# [0] https://github.com/github/codeql
+- import: codeql-suites/cpp-lgtm.qls
+ from: codeql/cpp-queries
+- import: codeql-suites/python-lgtm.qls
+ from: codeql/python-queries
+- include:
+ id:
+ - cpp/bad-strncpy-size
+ - cpp/declaration-hides-variable
+ - cpp/include-non-header
+ - cpp/inconsistent-null-check
+ - cpp/mistyped-function-arguments
+ - cpp/nested-loops-with-same-variable
+ - cpp/sizeof-side-effect
+ - cpp/suspicious-pointer-scaling
+ - cpp/suspicious-pointer-scaling-void
+ - cpp/suspicious-sizeof
+ - cpp/unsafe-strcat
+ - cpp/unsafe-strncat
+ - cpp/unsigned-difference-expression-compared-zero
+ - cpp/unused-local-variable
+ tags:
+ - "security"
+ - "correctness"
+ severity: "error"
+- exclude:
+ id:
+ - cpp/fixme-comment
diff --git a/.lgtm/cpp-queries/PotentiallyDangerousFunction.ql b/.github/codeql-queries/PotentiallyDangerousFunction.ql
similarity index 93%
rename from .lgtm/cpp-queries/PotentiallyDangerousFunction.ql
rename to .github/codeql-queries/PotentiallyDangerousFunction.ql
index 39e8dddd13..63fd14e75f 100644
--- a/.lgtm/cpp-queries/PotentiallyDangerousFunction.ql
+++ b/.github/codeql-queries/PotentiallyDangerousFunction.ql
@@ -46,6 +46,9 @@ predicate potentiallyDangerousFunction(Function f, string message) {
) or (
f.getQualifiedName() = "accept" and
message = "Call to accept() is not O_CLOEXEC-safe. Use accept4() instead."
+ ) or (
+ f.getQualifiedName() = "dirname" and
+ message = "Call dirname() is icky. Use path_extract_directory() instead."
)
}
diff --git a/.lgtm/cpp-queries/UninitializedVariableWithCleanup.ql b/.github/codeql-queries/UninitializedVariableWithCleanup.ql
similarity index 86%
rename from .lgtm/cpp-queries/UninitializedVariableWithCleanup.ql
rename to .github/codeql-queries/UninitializedVariableWithCleanup.ql
index 6b3b62f8bc..e514111f28 100644
--- a/.lgtm/cpp-queries/UninitializedVariableWithCleanup.ql
+++ b/.github/codeql-queries/UninitializedVariableWithCleanup.ql
@@ -50,16 +50,16 @@ class UninitialisedLocalReachability extends StackVariableReachability {
* fun(&x);
* puts(x);
*
- * `useOfVarActual()` won't treat this an an uninitialized read even if the callee
+ * `useOfVarActual()` won't treat this as an uninitialized read even if the callee
* doesn't modify the argument, however, `useOfVar()` will
*/
override predicate isSink(ControlFlowNode node, StackVariable v) { useOfVar(v, node) }
override predicate isBarrier(ControlFlowNode node, StackVariable v) {
- // only report the _first_ possibly uninitialized use
+ /* only report the _first_ possibly uninitialized use */
useOfVar(v, node) or
(
- /* If there's an return statement somewhere between the variable declaration
+ /* If there's a return statement somewhere between the variable declaration
* and a possible definition, don't accept is as a valid initialization.
*
* E.g.:
@@ -71,7 +71,7 @@ class UninitialisedLocalReachability extends StackVariableReachability {
* x = malloc(...);
*
* is not a valid initialization, since we might return from the function
- * _before_ the actual iniitialization (emphasis on _might_, since we
+ * _before_ the actual initialization (emphasis on _might_, since we
* don't know if the return statement might ever evaluate to true).
*/
definitionBarrier(v, node) and
@@ -92,14 +92,14 @@ predicate containsInlineAssembly(Function f) { exists(AsmStmt s | s.getEnclosing
* for this check to exclude them.
*/
VariableAccess commonException() {
- // If the uninitialized use we've found is in a macro expansion, it's
- // typically something like va_start(), and we don't want to complain.
+ /* If the uninitialized use we've found is in a macro expansion, it's
+ * typically something like va_start(), and we don't want to complain. */
result.getParent().isInMacroExpansion()
or
result.getParent() instanceof BuiltInOperation
or
- // Finally, exclude functions that contain assembly blocks. It's
- // anyone's guess what happens in those.
+ /* Finally, exclude functions that contain assembly blocks. It's
+ * anyone's guess what happens in those. */
containsInlineAssembly(result.getEnclosingFunction())
}
diff --git a/.github/codeql-queries/qlpack.yml b/.github/codeql-queries/qlpack.yml
new file mode 100644
index 0000000000..a1a2dec6d6
--- /dev/null
+++ b/.github/codeql-queries/qlpack.yml
@@ -0,0 +1,11 @@
+---
+# vi: ts=2 sw=2 et syntax=yaml:
+# SPDX-License-Identifier: LGPL-2.1-or-later
+
+library: false
+name: systemd/cpp-queries
+version: 0.0.1
+dependencies:
+ codeql/cpp-all: "*"
+ codeql/suite-helpers: "*"
+extractor: cpp
diff --git a/.github/workflows/codeql.yml b/.github/workflows/codeql.yml
new file mode 100644
index 0000000000..c5426d5686
--- /dev/null
+++ b/.github/workflows/codeql.yml
@@ -0,0 +1,68 @@
+---
+# vi: ts=2 sw=2 et:
+# SPDX-License-Identifier: LGPL-2.1-or-later
+#
+name: "CodeQL"
+
+on:
+ pull_request:
+ branches:
+ - master
+ - rhel-*
+ paths:
+ - '**/meson.build'
+ - '.github/**/codeql*'
+ - 'src/**'
+ - 'test/**'
+ - 'tools/**'
+ push:
+ branches:
+ - master
+ - rhel-*
+
+permissions:
+ contents: read
+
+jobs:
+ analyze:
+ name: Analyze
+ runs-on: ubuntu-22.04
+ concurrency:
+ group: ${{ github.workflow }}-${{ matrix.language }}-${{ github.ref }}
+ cancel-in-progress: true
+ permissions:
+ actions: read
+ security-events: write
+
+ strategy:
+ fail-fast: false
+ matrix:
+ language: ['cpp', 'python']
+
+ steps:
+ - name: Checkout repository
+ uses: actions/checkout@v3
+
+ - name: Initialize CodeQL
+ uses: github/codeql-action/init@v2
+ with:
+ languages: ${{ matrix.language }}
+ config-file: ./.github/codeql-config.yml
+
+ - name: Install dependencies
+ if: matrix.language == 'cpp'
+ run: |
+ echo "deb-src http://archive.ubuntu.com/ubuntu/ $(lsb_release -cs) main restricted universe multiverse" | sudo tee -a /etc/apt/sources.list
+ sudo apt-get -y update
+ sudo apt-get -y build-dep systemd
+ sudo apt-get -y install libfdisk-dev libpwquality-dev libqrencode-dev libssl-dev libxkbcommon-dev libzstd-dev
+
+ - name: Build
+ if: matrix.language == 'cpp'
+ run: |
+ # EL 8 systemd fails to build with newer gnu-efi (3.0.13 on Ubuntu Jammy ATTOW)
+ meson build -Dlibiptc=false -Dgnu-efi=false
+ ninja -C build -v
+
+ - name: Perform CodeQL Analysis
+ uses: github/codeql-action/analyze@v2
diff --git a/.lgtm.yml b/.lgtm.yml
deleted file mode 100644
index fe93957b67..0000000000
--- a/.lgtm.yml
+++ /dev/null
@@ -1,37 +0,0 @@
----
-# vi: ts=2 sw=2 et:
-
-# Explicitly enable certain checks which are hidden by default
-queries:
- - include: cpp/bad-strncpy-size
- - include: cpp/declaration-hides-variable
- - include: cpp/inconsistent-null-check
- - include: cpp/mistyped-function-arguments
- - include: cpp/nested-loops-with-same-variable
- - include: cpp/sizeof-side-effect
- - include: cpp/suspicious-pointer-scaling
- - include: cpp/suspicious-pointer-scaling-void
- - include: cpp/suspicious-sizeof
- - include: cpp/unsafe-strcat
- - include: cpp/unsafe-strncat
- - include: cpp/unsigned-difference-expression-compared-zero
- - include: cpp/unused-local-variable
- - include:
- tags:
- - "security"
- - "correctness"
- severity: "error"
-
-extraction:
- cpp:
- prepare:
- packages:
- - python3-pip
- - python3-setuptools
- - python3-wheel
- after_prepare:
- - pip3 install meson
- - export PATH="$HOME/.local/bin/:$PATH"
- python:
- python_setup:
- version: 3

71
SOURCES/0837-ci-mergify-Update-policy-Drop-LGTM-checks.patch Normal file
View File

@ -0,0 +1,71 @@
From 4c241b812ea79f3faa02c45f95834842c7847b76 Mon Sep 17 00:00:00 2001
From: Jan Macku <jamacku@redhat.com>
Date: Wed, 21 Sep 2022 15:14:26 +0200
Subject: [PATCH] ci(mergify): Update policy - Drop LGTM checks
rhel-only
Related: #2122499
---
.github/workflows/differential-shellcheck.yml | 1 +
.mergify.yml | 28 ++++++-------------
2 files changed, 9 insertions(+), 20 deletions(-)
diff --git a/.github/workflows/differential-shellcheck.yml b/.github/workflows/differential-shellcheck.yml
index fa94679b51..4399f0bc64 100644
--- a/.github/workflows/differential-shellcheck.yml
+++ b/.github/workflows/differential-shellcheck.yml
@@ -13,6 +13,7 @@ permissions:
jobs:
lint:
+ name: Differential ShellCheck
runs-on: ubuntu-latest
permissions:
diff --git a/.mergify.yml b/.mergify.yml
index 3afd04f18e..a5eed6a82a 100644
--- a/.mergify.yml
+++ b/.mergify.yml
@@ -11,16 +11,10 @@ pull_request_rules:
- -check-success=build (stream8, GCC_ASAN)
# CentOS Stream CI
- -check-success=CentOS CI (CentOS Stream 8)
- # LGTM
- - and:
- - "-check-success=LGTM analysis: JavaScript"
- - "-check-neutral=LGTM analysis: JavaScript"
- - and:
- - "-check-success=LGTM analysis: Python"
- - "-check-neutral=LGTM analysis: Python"
- - and:
- - "-check-success=LGTM analysis: C/C++"
- - "-check-neutral=LGTM analysis: C/C++"
+ # CodeQL
+ - -check-success=CodeQL
+ # Other
+ - -check-success=Differential ShellCheck
actions:
label:
add:
@@ -36,16 +30,10 @@ pull_request_rules:
- check-success=build (stream8, GCC_ASAN)
# CentOS Stream CI
- check-success=CentOS CI (CentOS Stream 8)
- # LGTM
- - or:
- - "check-success=LGTM analysis: JavaScript"
- - "check-neutral=LGTM analysis: JavaScript"
- - or:
- - "check-success=LGTM analysis: Python"
- - "check-neutral=LGTM analysis: Python"
- - or:
- - "check-success=LGTM analysis: C/C++"
- - "check-neutral=LGTM analysis: C/C++"
+ # CodeQL
+ - check-success=CodeQL
+ # Other
+ - check-success=Differential ShellCheck
actions:
label:
remove:

4
SOURCES/0832-time-util-fix-buffer-over-run.patch → SOURCES/0838-time-util-fix-buffer-over-run.patch
View File

@ -1,4 +1,4 @@
From 3ec7e3ece5474475b9123a67636f1ba3eec1e3c4 Mon Sep 17 00:00:00 2001 From a521f942d5c304bca7c61bacb3c79e565853718e Mon Sep 17 00:00:00 2001
From: Yu Watanabe <watanabe.yu+github@gmail.com> From: Yu Watanabe <watanabe.yu+github@gmail.com>
Date: Thu, 7 Jul 2022 18:27:02 +0900 Date: Thu, 7 Jul 2022 18:27:02 +0900
Subject: [PATCH] time-util: fix buffer-over-run Subject: [PATCH] time-util: fix buffer-over-run
@ -7,7 +7,7 @@ Fixes #23928.
(cherry picked from commit 9102c625a673a3246d7e73d8737f3494446bad4e) (cherry picked from commit 9102c625a673a3246d7e73d8737f3494446bad4e)
Resolves: #2139390 Resolves: #2139391
--- ---
src/basic/time-util.c | 2 +- src/basic/time-util.c | 2 +-
src/test/test-time-util.c | 5 +++++ src/test/test-time-util.c | 5 +++++

4
SOURCES/0834-basic-recognize-pdfs-filesystem-as-a-network-filesys.patch → SOURCES/0839-basic-recognize-pdfs-filesystem-as-a-network-filesys.patch
View File

@ -1,4 +1,4 @@
From 9274a3bec0d3b259a006574f6bfb559888b04533 Mon Sep 17 00:00:00 2001 From 2fe9fb3e844d7991105c40d4363eed9069a6837d Mon Sep 17 00:00:00 2001
From: Michal Sekletar <msekleta@redhat.com> From: Michal Sekletar <msekleta@redhat.com>
Date: Fri, 18 Nov 2022 16:16:36 +0100 Date: Fri, 18 Nov 2022 16:16:36 +0100
Subject: [PATCH] basic: recognize pdfs filesystem as a network filesystem Subject: [PATCH] basic: recognize pdfs filesystem as a network filesystem
@ -11,7 +11,7 @@ https://software.fujitsu.com/jp/manual/manualfiles/m130027/j2ul1563/02enz200/j15
RHEL-only RHEL-only
Resolves: #2143100 Resolves: #2094661
--- ---
src/basic/mount-util.c | 3 ++- src/basic/mount-util.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-) 1 file changed, 2 insertions(+), 1 deletion(-)

4
SOURCES/0833-core-move-reset_arguments-to-the-end-of-main-s-finis.patch → SOURCES/0840-core-move-reset_arguments-to-the-end-of-main-s-finis.patch
View File

@ -1,4 +1,4 @@
From 944596e66a9cd947545c714c4682681fdbefc09f Mon Sep 17 00:00:00 2001 From 4bb425eea9f3037a583a23d99f15aa71562f2481 Mon Sep 17 00:00:00 2001
From: Anita Zhang <the.anitazha@gmail.com> From: Anita Zhang <the.anitazha@gmail.com>
Date: Thu, 17 Sep 2020 01:49:17 -0700 Date: Thu, 17 Sep 2020 01:49:17 -0700
Subject: [PATCH] core: move reset_arguments() to the end of main's finish Subject: [PATCH] core: move reset_arguments() to the end of main's finish
@ -15,7 +15,7 @@ right before the return, when nothing else will be using the arg_* variables.
(cherry picked from commit 7d9eea2bd3d4f83668c7a78754d201b226acbf1e) (cherry picked from commit 7d9eea2bd3d4f83668c7a78754d201b226acbf1e)
Resolves: #2127170 Resolves: #2127131
--- ---
src/core/main.c | 3 ++- src/core/main.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-) 1 file changed, 2 insertions(+), 1 deletion(-)

91
SOURCES/0841-manager-move-inc.-of-n_reloading-into-a-function.patch Normal file
View File

@ -0,0 +1,91 @@
From 708c394b7ca35fe2328fa0760696ff95caab8ff8 Mon Sep 17 00:00:00 2001
From: David Tardon <dtardon@redhat.com>
Date: Tue, 29 Nov 2022 16:15:47 +0100
Subject: [PATCH] manager: move inc. of n_reloading into a function
[dtardon: This is inspired by commit d147e2b66b4d6b71db1bc59b62286b2eb9c3d29f ,
but it does just the minimal change needed for the next commit.]
Related: #2136869
---
src/core/main.c | 2 +-
src/core/manager.c | 12 ++++++++----
src/core/manager.h | 1 +
3 files changed, 10 insertions(+), 5 deletions(-)
diff --git a/src/core/main.c b/src/core/main.c
index cfa6fec930..c3e2ce8956 100644
--- a/src/core/main.c
+++ b/src/core/main.c
@@ -1131,7 +1131,7 @@ static int prepare_reexecute(Manager *m, FILE **_f, FDSet **_fds, bool switching
return log_error_errno(r, "Failed to create serialization file: %m");
/* Make sure nothing is really destructed when we shut down */
- m->n_reloading++;
+ manager_reloading_start(m);
bus_manager_send_reloading(m, true);
fds = fdset_new();
diff --git a/src/core/manager.c b/src/core/manager.c
index f4611e6f8f..f923cbce37 100644
--- a/src/core/manager.c
+++ b/src/core/manager.c
@@ -1578,6 +1578,10 @@ static void manager_preset_all(Manager *m) {
log_info("Populated /etc with preset unit settings.");
}
+void manager_reloading_start(Manager *m) {
+ m->n_reloading++;
+}
+
int manager_startup(Manager *m, FILE *serialization, FDSet *fds) {
int r;
@@ -1609,7 +1613,7 @@ int manager_startup(Manager *m, FILE *serialization, FDSet *fds) {
* this is already known, so we increase the counter here
* already */
if (serialization)
- m->n_reloading++;
+ manager_reloading_start(m);
/* First, enumerate what we can from all config files */
dual_timestamp_get(m->timestamps + MANAGER_TIMESTAMP_UNITS_LOAD_START);
@@ -3093,7 +3097,7 @@ int manager_serialize(Manager *m, FILE *f, FDSet *fds, bool switching_root) {
assert(f);
assert(fds);
- m->n_reloading++;
+ manager_reloading_start(m);
fprintf(f, "current-job-id=%"PRIu32"\n", m->current_job_id);
fprintf(f, "n-installed-jobs=%u\n", m->n_installed_jobs);
@@ -3211,7 +3215,7 @@ int manager_deserialize(Manager *m, FILE *f, FDSet *fds) {
log_debug("Deserializing state...");
- m->n_reloading++;
+ manager_reloading_start(m);
for (;;) {
_cleanup_free_ char *line = NULL;
@@ -3455,7 +3459,7 @@ int manager_reload(Manager *m) {
if (r < 0)
return r;
- m->n_reloading++;
+ manager_reloading_start(m);
bus_manager_send_reloading(m, true);
fds = fdset_new();
diff --git a/src/core/manager.h b/src/core/manager.h
index 3f2cfc5e2e..adbbb518cb 100644
--- a/src/core/manager.h
+++ b/src/core/manager.h
@@ -386,6 +386,7 @@ int manager_new(UnitFileScope scope, unsigned test_run_flags, Manager **m);
Manager* manager_free(Manager *m);
DEFINE_TRIVIAL_CLEANUP_FUNC(Manager*, manager_free);
+void manager_reloading_start(Manager *m);
int manager_startup(Manager *m, FILE *serialization, FDSet *fds);
Job *manager_get_job(Manager *m, uint32_t id);

59
SOURCES/0842-core-Add-new-DBUS-properties-UnitsReloadStartTimesta.patch Normal file
View File

@ -0,0 +1,59 @@
From 22eb8fbdab14e5b1b11a4d84c83bef97317e1d2a Mon Sep 17 00:00:00 2001
From: Jan Macku <jamacku@redhat.com>
Date: Thu, 2 Sep 2021 16:37:13 +0200
Subject: [PATCH] core: Add new DBUS properties UnitsReloadStartTimestamp and
UnitsLoadTimestampMontonic
(cherry picked from commit 49fbe940a429c3d8807bacdfce03af834275257c)
Related: #2136869
---
src/core/dbus-manager.c | 1 +
src/core/manager.c | 2 ++
src/core/manager.h | 1 +
3 files changed, 4 insertions(+)
diff --git a/src/core/dbus-manager.c b/src/core/dbus-manager.c
index 5b1ed3646e..8a41eda4a6 100644
--- a/src/core/dbus-manager.c
+++ b/src/core/dbus-manager.c
@@ -2486,6 +2486,7 @@ const sd_bus_vtable bus_manager_vtable[] = {
BUS_PROPERTY_DUAL_TIMESTAMP("GeneratorsFinishTimestamp", offsetof(Manager, timestamps[MANAGER_TIMESTAMP_GENERATORS_FINISH]), SD_BUS_VTABLE_PROPERTY_CONST),
BUS_PROPERTY_DUAL_TIMESTAMP("UnitsLoadStartTimestamp", offsetof(Manager, timestamps[MANAGER_TIMESTAMP_UNITS_LOAD_START]), SD_BUS_VTABLE_PROPERTY_CONST),
BUS_PROPERTY_DUAL_TIMESTAMP("UnitsLoadFinishTimestamp", offsetof(Manager, timestamps[MANAGER_TIMESTAMP_UNITS_LOAD_FINISH]), SD_BUS_VTABLE_PROPERTY_CONST),
+ BUS_PROPERTY_DUAL_TIMESTAMP("UnitsLoadTimestamp", offsetof(Manager, timestamps[MANAGER_TIMESTAMP_UNITS_LOAD]), SD_BUS_VTABLE_PROPERTY_CONST),
SD_BUS_WRITABLE_PROPERTY("LogLevel", "s", property_get_log_level, property_set_log_level, 0, 0),
SD_BUS_WRITABLE_PROPERTY("LogTarget", "s", property_get_log_target, property_set_log_target, 0, 0),
SD_BUS_PROPERTY("NNames", "u", property_get_hashmap_size, offsetof(Manager, units), 0),
diff --git a/src/core/manager.c b/src/core/manager.c
index f923cbce37..8aa398cac8 100644
--- a/src/core/manager.c
+++ b/src/core/manager.c
@@ -1580,6 +1580,7 @@ static void manager_preset_all(Manager *m) {
void manager_reloading_start(Manager *m) {
m->n_reloading++;
+ dual_timestamp_get(m->timestamps + MANAGER_TIMESTAMP_UNITS_LOAD);
}
int manager_startup(Manager *m, FILE *serialization, FDSet *fds) {
@@ -4622,6 +4623,7 @@ static const char *const manager_timestamp_table[_MANAGER_TIMESTAMP_MAX] = {
[MANAGER_TIMESTAMP_GENERATORS_FINISH] = "generators-finish",
[MANAGER_TIMESTAMP_UNITS_LOAD_START] = "units-load-start",
[MANAGER_TIMESTAMP_UNITS_LOAD_FINISH] = "units-load-finish",
+ [MANAGER_TIMESTAMP_UNITS_LOAD] = "units-load",
};
DEFINE_STRING_TABLE_LOOKUP(manager_timestamp, ManagerTimestamp);
diff --git a/src/core/manager.h b/src/core/manager.h
index adbbb518cb..98d381bc5b 100644
--- a/src/core/manager.h
+++ b/src/core/manager.h
@@ -67,6 +67,7 @@ typedef enum ManagerTimestamp {
MANAGER_TIMESTAMP_GENERATORS_FINISH,
MANAGER_TIMESTAMP_UNITS_LOAD_START,
MANAGER_TIMESTAMP_UNITS_LOAD_FINISH,
+ MANAGER_TIMESTAMP_UNITS_LOAD,
_MANAGER_TIMESTAMP_MAX,
_MANAGER_TIMESTAMP_INVALID = -1,
} ManagerTimestamp;

29
SOURCES/0843-core-Indicate-the-time-when-the-manager-started-load.patch Normal file
View File

@ -0,0 +1,29 @@
From f2de5398b0a1ebb3e6390506368c11329b843524 Mon Sep 17 00:00:00 2001
From: Jan Macku <jamacku@redhat.com>
Date: Thu, 2 Sep 2021 16:50:50 +0200
Subject: [PATCH] core: Indicate the time when the manager started loading
units the last time
(cherry picked from commit 15b9243c0d7f6d1531fa65dbc01bd11e8e6c12ca)
Resolves: #2136869
---
src/core/manager.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/src/core/manager.c b/src/core/manager.c
index 8aa398cac8..a9cd51b624 100644
--- a/src/core/manager.c
+++ b/src/core/manager.c
@@ -3554,6 +3554,11 @@ int manager_reload(Manager *m) {
/* Let's finally catch up with any changes that took place while we were reloading/reexecing */
manager_catchup(m);
+ /* Create a file which will indicate when the manager started loading units the last time. */
+ (void) touch_file("/run/systemd/systemd-units-load", false,
+ m->timestamps[MANAGER_TIMESTAMP_UNITS_LOAD].realtime ?: now(CLOCK_REALTIME),
+ UID_INVALID, GID_INVALID, 0444);
+
/* Sync current state of bus names with our set of listening units */
q = manager_enqueue_sync_bus_names(m);
if (q < 0 && r >= 0)

34
SOURCES/0844-core-do-not-touch-run-systemd-systemd-units-load-fro.patch Normal file
View File

@ -0,0 +1,34 @@
From 3c2d2345814935cea8525e802e764fb2949eb3df Mon Sep 17 00:00:00 2001
From: Luca Boccassi <luca.boccassi@microsoft.com>
Date: Mon, 27 Dec 2021 18:22:43 +0000
Subject: [PATCH] core: do not touch /run/systemd/systemd-units-load from user
session instances
Follow-up for: https://github.com/systemd/systemd/commit/15b9243c0d7f6d1531fa65dbc01bd11e8e6c12ca
Fixes: https://github.com/systemd/systemd/issues/21911
(cherry picked from commit 4b3ad81bfafcd97acb06db463495e348d159d8e6)
Related: #2136869
---
src/core/manager.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/src/core/manager.c b/src/core/manager.c
index a9cd51b624..e083596e58 100644
--- a/src/core/manager.c
+++ b/src/core/manager.c
@@ -3555,9 +3555,10 @@ int manager_reload(Manager *m) {
manager_catchup(m);
/* Create a file which will indicate when the manager started loading units the last time. */
- (void) touch_file("/run/systemd/systemd-units-load", false,
- m->timestamps[MANAGER_TIMESTAMP_UNITS_LOAD].realtime ?: now(CLOCK_REALTIME),
- UID_INVALID, GID_INVALID, 0444);
+ if (MANAGER_IS_SYSTEM(m))
+ (void) touch_file("/run/systemd/systemd-units-load", false,
+ m->timestamps[MANAGER_TIMESTAMP_UNITS_LOAD].realtime ?: now(CLOCK_REALTIME),
+ UID_INVALID, GID_INVALID, 0444);
/* Sync current state of bus names with our set of listening units */
q = manager_enqueue_sync_bus_names(m);

49
SOURCES/0845-sysctl-downgrade-message-when-we-have-no-permission.patch Normal file
View File

@ -0,0 +1,49 @@
From ffe4233155085b479c69abe844a34de212b8e5e1 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Zbigniew=20J=C4=99drzejewski-Szmek?= <zbyszek@in.waw.pl>
Date: Thu, 16 Jan 2020 14:45:28 +0100
Subject: [PATCH] sysctl: downgrade message when we have no permission
We need to run sysctl also in containers, because the network
subtree is namespaces and may legitimately be writable. But logging
all "errors" at notice level creates unwanted noise.
Also downgrade message about missing sysctls to log_info. This might also be
relatively common when configuration is targeted at different kernel
versions. With log_debug it'll still end up in the logs, but isn't really worth
of "notice" most of the time.
https://bugzilla.redhat.com/show_bug.cgi?id=1609806
(cherry picked from commit 32458cc9687c1b60ff0f22c0e71da93ce78b1534)
Resolves: #2158160
---
src/sysctl/sysctl.c | 16 +++++++++-------
1 file changed, 9 insertions(+), 7 deletions(-)
diff --git a/src/sysctl/sysctl.c b/src/sysctl/sysctl.c
index 4c85d6887f..dc14e1aaf1 100644
--- a/src/sysctl/sysctl.c
+++ b/src/sysctl/sysctl.c
@@ -82,13 +82,15 @@ static int apply_all(OrderedHashmap *sysctl_options) {
k = sysctl_write(option->key, option->value);
if (k < 0) {
/* If the sysctl is not available in the kernel or we are running with reduced
- * privileges and cannot write it, then log about the issue at LOG_NOTICE level, and
- * proceed without failing. (EROFS is treated as a permission problem here, since
- * that's how container managers usually protected their sysctls.) In all other cases
- * log an error and make the tool fail. */
-
- if (IN_SET(k, -EPERM, -EACCES, -EROFS, -ENOENT) || option->ignore_failure)
- log_notice_errno(k, "Couldn't write '%s' to '%s', ignoring: %m", option->value, option->key);
+ * privileges and cannot write it, then log about the issue, and proceed without
+ * failing. (EROFS is treated as a permission problem here, since that's how
+ * container managers usually protected their sysctls.) In all other cases log an
+ * error and make the tool fail. */
+
+ if (option->ignore_failure || k == -EROFS || ERRNO_IS_PRIVILEGE(k))
+ log_debug_errno(k, "Couldn't write '%s' to '%s', ignoring: %m", option->value, option->key);
+ else if (k == -ENOENT)
+ log_info_errno(k, "Couldn't write '%s' to '%s', ignoring: %m", option->value, option->key);
else {
log_error_errno(k, "Couldn't write '%s' to '%s': %m", option->value, option->key);
if (r == 0)

53
SOURCES/0846-core-respect-SELinuxContext-for-socket-creation.patch Normal file
View File

@ -0,0 +1,53 @@
From 1f408c8d9739b1038012eeec7bf0f918c8095bc4 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Christian=20G=C3=B6ttsche?= <cgzones@googlemail.com>
Date: Fri, 23 Sep 2022 19:00:22 +0200
Subject: [PATCH] core: respect SELinuxContext= for socket creation
On socket creation respect the SELinuxContext= setting of the associated
service, such that the initial created socket has the same label as the
future process accepting the connection (since w.r.t SELinux sockets
normally have the same label as the owning process).
Triggered by #24702
(cherry picked from commit 599b384924bbef9f8f7fa5700c6fa35a404d9a98)
Related: #2136738
---
src/core/socket.c | 15 ++++++++++++++-
1 file changed, 14 insertions(+), 1 deletion(-)
diff --git a/src/core/socket.c b/src/core/socket.c
index 9d47ca2616..d1ca0a07c5 100644
--- a/src/core/socket.c
+++ b/src/core/socket.c
@@ -1427,6 +1427,7 @@ fail:
static int socket_determine_selinux_label(Socket *s, char **ret) {
Service *service;
ExecCommand *c;
+ const char *exec_context;
_cleanup_free_ char *path = NULL;
int r;
@@ -1448,8 +1449,20 @@ static int socket_determine_selinux_label(Socket *s, char **ret) {
if (!UNIT_ISSET(s->service))
goto no_label;
-
service = SERVICE(UNIT_DEREF(s->service));
+
+ exec_context = service->exec_context.selinux_context;
+ if (exec_context) {
+ char *con;
+
+ con = strdup(exec_context);
+ if (!con)
+ return -ENOMEM;
+
+ *ret = TAKE_PTR(con);
+ return 0;
+ }
+
c = service->exec_command[SERVICE_EXEC_START];
if (!c)
goto no_label;

128
SOURCES/0847-manager-use-target-process-context-to-set-socket-con.patch Normal file
View File

@ -0,0 +1,128 @@
From 3f90090e70a5fa81bced17792fe08d9c46324da9 Mon Sep 17 00:00:00 2001
From: "Ted X. Toth" <txtoth@flycast.org>
Date: Thu, 13 Oct 2022 12:58:26 -0700
Subject: [PATCH] manager: use target process context to set socket context
Use target process context to set socket context when using SELinuxContextFromNet
not systemd's context. Currently when using the SELinuxContextFromNet option for
a socket activated services, systemd calls getcon_raw which returns init_t and
uses the resulting context to compute the context to be passed to the
setsockcreatecon call. A socket of type init_t is created and listened on and
this means that SELinux policy cannot be written to control which processes
(SELinux types) can connect to the socket since the ref policy allows all
'types' to connect to sockets of the type init_t. When security accessors see
that any process can connect to a socket this raises serious concerns. I have
spoken with SELinux contributors in person and on the mailing list and the
consensus is that the best solution is to use the target executables context
when computing the sockets context in all cases.
[zjs review/comment:
This removes the branch that was added in 16115b0a7b7cdf08fb38084d857d572d8a9088dc.
16115b0a7b7cdf08fb38084d857d572d8a9088dc did two things: it had the branch here
in 'socket_determine_selinux_label()' and a code in 'exec_child()' to call
'label_get_child_mls_label(socket_fd, command->path, &label)'.
Before this patch, the flow was:
'''
mac_selinux_get_child_mls_label:
peercon = getpeercon_raw(socket_fd);
if (!exec_label)
exec_label = getfilecon_raw(exe);
socket_open_fds:
if (params->selinux_context_net) #
label = mac_selinux_get_our_label(); # this part is removed
else #
label = mac_selinux_get_create_label_from_exe(path);
socket_address_listen_in_cgroup(s, &p->address, label);
exec_child():
exec_context = mac_selinux_get_child_mls_label(fd, executable, context->selinux_context);
setexeccon(exec_context);
'''
]
(cherry picked from commit 29dbc62d74f7b7881dc3136e68e03a03ea055b36)
Resolves: #2136738
---
src/core/socket.c | 58 ++++++++++++++++++++---------------------------
1 file changed, 24 insertions(+), 34 deletions(-)
diff --git a/src/core/socket.c b/src/core/socket.c
index d1ca0a07c5..8aa5463b25 100644
--- a/src/core/socket.c
+++ b/src/core/socket.c
@@ -1434,47 +1434,37 @@ static int socket_determine_selinux_label(Socket *s, char **ret) {
assert(s);
assert(ret);
- if (s->selinux_context_from_net) {
- /* If this is requested, get label from the network label */
-
- r = mac_selinux_get_our_label(ret);
- if (r == -EOPNOTSUPP)
- goto no_label;
-
- } else {
- /* Otherwise, get it from the executable we are about to start */
- r = socket_instantiate_service(s);
- if (r < 0)
- return r;
+ r = socket_instantiate_service(s);
+ if (r < 0)
+ return r;
- if (!UNIT_ISSET(s->service))
- goto no_label;
- service = SERVICE(UNIT_DEREF(s->service));
+ if (!UNIT_ISSET(s->service))
+ goto no_label;
+ service = SERVICE(UNIT_DEREF(s->service));
- exec_context = service->exec_context.selinux_context;
- if (exec_context) {
- char *con;
+ exec_context = service->exec_context.selinux_context;
+ if (exec_context) {
+ char *con;
- con = strdup(exec_context);
- if (!con)
- return -ENOMEM;
+ con = strdup(exec_context);
+ if (!con)
+ return -ENOMEM;
- *ret = TAKE_PTR(con);
- return 0;
- }
+ *ret = TAKE_PTR(con);
+ return 0;
+ }
- c = service->exec_command[SERVICE_EXEC_START];
- if (!c)
- goto no_label;
+ c = service->exec_command[SERVICE_EXEC_START];
+ if (!c)
+ goto no_label;
- r = chase_symlinks(c->path, service->exec_context.root_directory, CHASE_PREFIX_ROOT, &path);
- if (r < 0)
- goto no_label;
+ r = chase_symlinks(c->path, service->exec_context.root_directory, CHASE_PREFIX_ROOT, &path);
+ if (r < 0)
+ goto no_label;
- r = mac_selinux_get_create_label_from_exe(path, ret);
- if (IN_SET(r, -EPERM, -EOPNOTSUPP))
- goto no_label;
- }
+ r = mac_selinux_get_create_label_from_exe(path, ret);
+ if (IN_SET(r, -EPERM, -EOPNOTSUPP))
+ goto no_label;
return r;

142
SOURCES/0848-virt-detect-Amazon-EC2-Nitro-instance.patch Normal file
View File

@ -0,0 +1,142 @@
From 6ffd3de2ccc5901974f292c9694829e25441060d Mon Sep 17 00:00:00 2001
From: Bertrand Jacquin <bertrand@jacquin.bzh>
Date: Sun, 11 Oct 2020 21:25:00 +0100
Subject: [PATCH] virt: detect Amazon EC2 Nitro instance
Amazon EC2 Nitro hypervisor is technically based on KVM[1], which
systemd-detect-virt identify propely from CPUID. However the lack of
CPUID on aarch64 (A1, T4 instance type) prevents a correct
identification, impacting hostnamectl and systemd-random-seed. Instead
it's possible to identify virtualization from DMI vendor ID.
Prior to this commit:
# hostnamectl
Static hostname: n/a
Transient hostname: ip-10-97-8-12
Icon name: computer
Machine ID: 8e3772fbcfa3dd6f330a12ff5df5a63b
Boot ID: b7b7e2fe0079448db664839df59f9817
Operating System: Gentoo/Linux
Kernel: Linux 5.4.69-longterm
Architecture: arm64
After this commit:
# hostnamectl
Static hostname: n/a
Transient hostname: ip-10-97-8-12
Icon name: computer-vm
Chassis: vm
Machine ID: 8e3772fbcfa3dd6f330a12ff5df5a63b
Boot ID: bd04da57084e41078f20541101867113
Virtualization: amazon
Operating System: Gentoo/Linux
Kernel: Linux 5.4.69-longterm
Architecture: arm64
[1] https://aws.amazon.com/ec2/faqs/
(cherry picked from commit b6eca3731dd92b009b182f188936e1c2544574da)
Resolves: #2117948
---
man/systemd-detect-virt.xml | 7 ++++++-
man/systemd.unit.xml | 1 +
src/basic/virt.c | 8 +++++---
src/basic/virt.h | 1 +
src/test/test-condition.c | 1 +
5 files changed, 14 insertions(+), 4 deletions(-)
diff --git a/man/systemd-detect-virt.xml b/man/systemd-detect-virt.xml
index 6beb2c2aa1..61c210e24d 100644
--- a/man/systemd-detect-virt.xml
+++ b/man/systemd-detect-virt.xml
@@ -72,7 +72,12 @@
<row>
<entry><varname>kvm</varname></entry>
- <entry>Linux KVM kernel virtual machine, with whatever software, except Oracle Virtualbox</entry>
+ <entry>Linux KVM kernel virtual machine, in combination with QEMU. Not used for other virtualizers using the KVM interfaces, such as Oracle VirtualBox or Amazon EC2 Nitro, see below.</entry>
+ </row>
+
+ <row>
+ <entry><varname>amazon</varname></entry>
+ <entry>Amazon EC2 Nitro using Linux KVM</entry>
</row>
<row>
diff --git a/man/systemd.unit.xml b/man/systemd.unit.xml
index 6f213ccd56..5207a5bb3c 100644
--- a/man/systemd.unit.xml
+++ b/man/systemd.unit.xml
@@ -1068,6 +1068,7 @@
virtualization solution, or one of
<varname>qemu</varname>,
<varname>kvm</varname>,
+ <literal>amazon</literal>,
<varname>zvm</varname>,
<varname>vmware</varname>,
<varname>microsoft</varname>,
diff --git a/src/basic/virt.c b/src/basic/virt.c
index 8d862b6d67..78c68d66e0 100644
--- a/src/basic/virt.c
+++ b/src/basic/virt.c
@@ -147,6 +147,7 @@ static int detect_vm_dmi(void) {
int id;
} dmi_vendor_table[] = {
{ "KVM", VIRTUALIZATION_KVM },
+ { "Amazon EC2", VIRTUALIZATION_AMAZON },
{ "QEMU", VIRTUALIZATION_QEMU },
/* http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1009458 */
{ "VMware", VIRTUALIZATION_VMWARE },
@@ -339,8 +340,8 @@ int detect_vm(void) {
/* We have to use the correct order here:
*
- * → First, try to detect Oracle Virtualbox, even if it uses KVM, as well as Xen even if it cloaks as Microsoft
- * Hyper-V.
+ * → First, try to detect Oracle Virtualbox and Amazon EC2 Nitro, even if they use KVM, as well as Xen even if
+ * it cloaks as Microsoft Hyper-V.
*
* → Second, try to detect from CPUID, this will report KVM for whatever software is used even if info in DMI is
* overwritten.
@@ -348,7 +349,7 @@ int detect_vm(void) {
* → Third, try to detect from DMI. */
dmi = detect_vm_dmi();
- if (IN_SET(dmi, VIRTUALIZATION_ORACLE, VIRTUALIZATION_XEN)) {
+ if (IN_SET(dmi, VIRTUALIZATION_ORACLE, VIRTUALIZATION_XEN, VIRTUALIZATION_AMAZON)) {
r = dmi;
goto finish;
}
@@ -631,6 +632,7 @@ int running_in_chroot(void) {
static const char *const virtualization_table[_VIRTUALIZATION_MAX] = {
[VIRTUALIZATION_NONE] = "none",
[VIRTUALIZATION_KVM] = "kvm",
+ [VIRTUALIZATION_AMAZON] = "amazon",
[VIRTUALIZATION_QEMU] = "qemu",
[VIRTUALIZATION_BOCHS] = "bochs",
[VIRTUALIZATION_XEN] = "xen",
diff --git a/src/basic/virt.h b/src/basic/virt.h
index 640b3ed779..ed4ff063e0 100644
--- a/src/basic/virt.h
+++ b/src/basic/virt.h
@@ -10,6 +10,7 @@ enum {
VIRTUALIZATION_VM_FIRST,
VIRTUALIZATION_KVM = VIRTUALIZATION_VM_FIRST,
+ VIRTUALIZATION_AMAZON,
VIRTUALIZATION_QEMU,
VIRTUALIZATION_BOCHS,
VIRTUALIZATION_XEN,
diff --git a/src/test/test-condition.c b/src/test/test-condition.c
index 24395dafc6..29ea63c4ff 100644
--- a/src/test/test-condition.c
+++ b/src/test/test-condition.c
@@ -510,6 +510,7 @@ static void test_condition_test_virtualization(void) {
NULSTR_FOREACH(virt,
"kvm\0"
+ "amazon\0"
"qemu\0"
"bochs\0"
"xen\0"

37
SOURCES/0849-machine-id-setup-generate-machine-id-from-DMI-produc.patch Normal file
View File

@ -0,0 +1,37 @@
From e320f72150829228f10ec24f3fba34d5377c5120 Mon Sep 17 00:00:00 2001
From: Bertrand Jacquin <bertrand@jacquin.bzh>
Date: Sun, 11 Oct 2020 22:25:56 +0100
Subject: [PATCH] machine-id-setup: generate machine-id from DMI product ID on
Amazon EC2
Amazon EC2 Nitro hypervisor is technically based on KVM[1].
[1] https://aws.amazon.com/ec2/faqs/
(cherry picked from commit 382a46d129899ca9027b07c325102cab173dd563)
Related: #2117948
---
src/core/machine-id-setup.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/src/core/machine-id-setup.c b/src/core/machine-id-setup.c
index 11528f83c4..fe2abc4e68 100644
--- a/src/core/machine-id-setup.c
+++ b/src/core/machine-id-setup.c
@@ -57,11 +57,11 @@ static int generate_machine_id(const char *root, sd_id128_t *ret) {
return 0;
}
- } else if (detect_vm() == VIRTUALIZATION_KVM) {
+ } else if (IN_SET(detect_vm(), VIRTUALIZATION_KVM, VIRTUALIZATION_AMAZON, VIRTUALIZATION_QEMU)) {
- /* If we are not running in a container, see if we are
- * running in qemu/kvm and a machine ID was passed in
- * via -uuid on the qemu/kvm command line */
+ /* If we are not running in a container, see if we are running in a VM that provides
+ * a system UUID via the SMBIOS/DMI interfaces. Such environments include QEMU/KVM
+ * with the -uuid on the qemu command line or the Amazon EC2 Nitro hypervisor. */
if (id128_read("/sys/class/dmi/id/product_uuid", ID128_UUID, ret) >= 0) {
log_info("Initializing machine ID from KVM UUID.");

134
SOURCES/0850-virt-use-string-table-to-detect-VM-or-container.patch Normal file
View File

@ -0,0 +1,134 @@
From efa2cdb699df3e5d5d7180e50f3ebfff74788c5c Mon Sep 17 00:00:00 2001
From: Yu Watanabe <watanabe.yu+github@gmail.com>
Date: Tue, 7 Jan 2020 11:49:39 +0900
Subject: [PATCH] virt: use string table to detect VM or container
(cherry picked from commit 735ea55f5cd87a82757a8911edd80fba799b46ee)
Related: #2117948
---
src/basic/virt.c | 73 ++++++++++++++++++++++--------------------------
1 file changed, 33 insertions(+), 40 deletions(-)
diff --git a/src/basic/virt.c b/src/basic/virt.c
index 78c68d66e0..6e4c702051 100644
--- a/src/basic/virt.c
+++ b/src/basic/virt.c
@@ -22,27 +22,26 @@
#include "string-util.h"
#include "virt.h"
+static const char *const vm_table[_VIRTUALIZATION_MAX] = {
+ [VIRTUALIZATION_XEN] = "XenVMMXenVMM",
+ [VIRTUALIZATION_KVM] = "KVMKVMKVM",
+ [VIRTUALIZATION_QEMU] = "TCGTCGTCGTCG",
+ /* http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1009458 */
+ [VIRTUALIZATION_VMWARE] = "VMwareVMware",
+ /* https://docs.microsoft.com/en-us/virtualization/hyper-v-on-windows/reference/tlfs */
+ [VIRTUALIZATION_MICROSOFT] = "Microsoft Hv",
+ /* https://wiki.freebsd.org/bhyve */
+ [VIRTUALIZATION_BHYVE] = "bhyve bhyve ",
+ [VIRTUALIZATION_QNX] = "QNXQVMBSQG",
+};
+
+DEFINE_PRIVATE_STRING_TABLE_LOOKUP_FROM_STRING(vm, int);
+
static int detect_vm_cpuid(void) {
/* CPUID is an x86 specific interface. */
#if defined(__i386__) || defined(__x86_64__)
- static const struct {
- const char *cpuid;
- int id;
- } cpuid_vendor_table[] = {
- { "XenVMMXenVMM", VIRTUALIZATION_XEN },
- { "KVMKVMKVM", VIRTUALIZATION_KVM },
- { "TCGTCGTCGTCG", VIRTUALIZATION_QEMU },
- /* http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1009458 */
- { "VMwareVMware", VIRTUALIZATION_VMWARE },
- /* https://docs.microsoft.com/en-us/virtualization/hyper-v-on-windows/reference/tlfs */
- { "Microsoft Hv", VIRTUALIZATION_MICROSOFT },
- /* https://wiki.freebsd.org/bhyve */
- { "bhyve bhyve ", VIRTUALIZATION_BHYVE },
- { "QNXQVMBSQG", VIRTUALIZATION_QNX },
- };
-
uint32_t eax, ebx, ecx, edx;
bool hypervisor;
@@ -59,7 +58,7 @@ static int detect_vm_cpuid(void) {
uint32_t sig32[3];
char text[13];
} sig = {};
- unsigned j;
+ int v;
/* There is a hypervisor, see what it is */
__cpuid(0x40000000U, eax, ebx, ecx, edx);
@@ -70,11 +69,11 @@ static int detect_vm_cpuid(void) {
log_debug("Virtualization found, CPUID=%s", sig.text);
- for (j = 0; j < ELEMENTSOF(cpuid_vendor_table); j ++)
- if (streq(sig.text, cpuid_vendor_table[j].cpuid))
- return cpuid_vendor_table[j].id;
+ v = vm_from_string(sig.text);
+ if (v < 0)
+ return VIRTUALIZATION_VM_OTHER;
- return VIRTUALIZATION_VM_OTHER;
+ return v;
}
#endif
log_debug("No virtualization found in CPUID");
@@ -434,22 +433,20 @@ finish:
return r;
}
-int detect_container(void) {
- static const struct {
- const char *value;
- int id;
- } value_table[] = {
- { "lxc", VIRTUALIZATION_LXC },
- { "lxc-libvirt", VIRTUALIZATION_LXC_LIBVIRT },
- { "systemd-nspawn", VIRTUALIZATION_SYSTEMD_NSPAWN },
- { "docker", VIRTUALIZATION_DOCKER },
- { "rkt", VIRTUALIZATION_RKT },
- };
+static const char *const container_table[_VIRTUALIZATION_MAX] = {
+ [VIRTUALIZATION_LXC] = "lxc",
+ [VIRTUALIZATION_LXC_LIBVIRT] = "lxc-libvirt",
+ [VIRTUALIZATION_SYSTEMD_NSPAWN] = "systemd-nspawn",
+ [VIRTUALIZATION_DOCKER] = "docker",
+ [VIRTUALIZATION_RKT] = "rkt",
+};
+
+DEFINE_PRIVATE_STRING_TABLE_LOOKUP_FROM_STRING(container, int);
+int detect_container(void) {
static thread_local int cached_found = _VIRTUALIZATION_INVALID;
_cleanup_free_ char *m = NULL;
const char *e = NULL;
- unsigned j;
int r;
if (cached_found >= 0)
@@ -522,13 +519,9 @@ int detect_container(void) {
goto finish;
translate_name:
- for (j = 0; j < ELEMENTSOF(value_table); j++)
- if (streq(e, value_table[j].value)) {
- r = value_table[j].id;
- goto finish;
- }
-
- r = VIRTUALIZATION_CONTAINER_OTHER;
+ r = container_from_string(e);
+ if (r < 0)
+ r = VIRTUALIZATION_CONTAINER_OTHER;
finish:
log_debug("Found container virtualization %s.", virtualization_to_string(r));

194
SOURCES/0851-fileio-introduce-read_full_virtual_file-for-reading-.patch Normal file
View File

@ -0,0 +1,194 @@
From 7a3843972ea290daf1bec5e1133db654749b8c02 Mon Sep 17 00:00:00 2001
From: Franck Bui <fbui@suse.com>
Date: Tue, 22 Oct 2019 16:09:21 +0200
Subject: [PATCH] fileio: introduce read_full_virtual_file() for reading
virtual files in sysfs, procfs
Virtual filesystems such as sysfs or procfs use kernfs, and kernfs can work
with two sorts of virtual files.
One sort uses "seq_file", and the results of the first read are buffered for
the second read. The other sort uses "raw" reads which always go direct to the
device.
In the later case, the content of the virtual file must be retrieved with a
single read otherwise subsequent read might get the new value instead of
finding EOF immediately. That's the reason why the usage of fread(3) is
prohibited in this case as it always performs a second call to read(2) looking
for EOF which is subject to the race described previously.
Fixes: #13585.
(cherry picked from commit 21b40f16622f171a9969dc334d74fb5eb2f575c2)
Related: #2117948
---
src/basic/fileio.c | 115 ++++++++++++++++++++++++++-
src/basic/fileio.h | 1 +
src/libsystemd/sd-device/sd-device.c | 2 +-
3 files changed, 113 insertions(+), 5 deletions(-)
diff --git a/src/basic/fileio.c b/src/basic/fileio.c
index 6b0bad5b71..733fb42463 100644
--- a/src/basic/fileio.c
+++ b/src/basic/fileio.c
@@ -276,6 +276,113 @@ int verify_file(const char *fn, const char *blob, bool accept_extra_nl) {
return 1;
}
+int read_full_virtual_file(const char *filename, char **ret_contents, size_t *ret_size) {
+ _cleanup_free_ char *buf = NULL;
+ _cleanup_close_ int fd = -1;
+ struct stat st;
+ size_t n, size;
+ int n_retries;
+ char *p;
+
+ assert(ret_contents);
+
+ /* Virtual filesystems such as sysfs or procfs use kernfs, and kernfs can work
+ * with two sorts of virtual files. One sort uses "seq_file", and the results of
+ * the first read are buffered for the second read. The other sort uses "raw"
+ * reads which always go direct to the device. In the latter case, the content of
+ * the virtual file must be retrieved with a single read otherwise a second read
+ * might get the new value instead of finding EOF immediately. That's the reason
+ * why the usage of fread(3) is prohibited in this case as it always performs a
+ * second call to read(2) looking for EOF. See issue 13585. */
+
+ fd = open(filename, O_RDONLY|O_CLOEXEC);
+ if (fd < 0)
+ return -errno;
+
+ /* Start size for files in /proc which usually report a file size of 0. */
+ size = LINE_MAX / 2;
+
+ /* Limit the number of attempts to read the number of bytes returned by fstat(). */
+ n_retries = 3;
+
+ for (;;) {
+ if (n_retries <= 0)
+ return -EIO;
+
+ if (fstat(fd, &st) < 0)
+ return -errno;
+
+ if (!S_ISREG(st.st_mode))
+ return -EBADF;
+
+ /* Be prepared for files from /proc which generally report a file size of 0. */
+ if (st.st_size > 0) {
+ size = st.st_size;
+ n_retries--;
+ } else
+ size = size * 2;
+
+ if (size > READ_FULL_BYTES_MAX)
+ return -E2BIG;
+
+ p = realloc(buf, size + 1);
+ if (!p)
+ return -ENOMEM;
+ buf = TAKE_PTR(p);
+
+ for (;;) {
+ ssize_t k;
+
+ /* Read one more byte so we can detect whether the content of the
+ * file has already changed or the guessed size for files from /proc
+ * wasn't large enough . */
+ k = read(fd, buf, size + 1);
+ if (k >= 0) {
+ n = k;
+ break;
+ }
+
+ if (errno != -EINTR)
+ return -errno;
+ }
+
+ /* Consider a short read as EOF */
+ if (n <= size)
+ break;
+
+ /* Hmm... either we read too few bytes from /proc or less likely the content
+ * of the file might have been changed (and is now bigger) while we were
+ * processing, let's try again either with a bigger guessed size or the new
+ * file size. */
+
+ if (lseek(fd, 0, SEEK_SET) < 0)
+ return -errno;
+ }
+
+ if (n < size) {
+ p = realloc(buf, n + 1);
+ if (!p)
+ return -ENOMEM;
+ buf = TAKE_PTR(p);
+ }
+
+ if (!ret_size) {
+ /* Safety check: if the caller doesn't want to know the size of what we
+ * just read it will rely on the trailing NUL byte. But if there's an
+ * embedded NUL byte, then we should refuse operation as otherwise
+ * there'd be ambiguity about what we just read. */
+
+ if (memchr(buf, 0, n))
+ return -EBADMSG;
+ } else
+ *ret_size = n;
+
+ buf[n] = 0;
+ *ret_contents = TAKE_PTR(buf);
+
+ return 0;
+}
+
int read_full_stream(FILE *f, char **contents, size_t *size) {
_cleanup_free_ char *buf = NULL;
struct stat st;
@@ -300,9 +407,9 @@ int read_full_stream(FILE *f, char **contents, size_t *size) {
if (st.st_size > READ_FULL_BYTES_MAX)
return -E2BIG;
- /* Start with the right file size, but be prepared for files from /proc which generally report a file
- * size of 0. Note that we increase the size to read here by one, so that the first read attempt
- * already makes us notice the EOF. */
+ /* Start with the right file size. Note that we increase the size
+ * to read here by one, so that the first read attempt already
+ * makes us notice the EOF. */
if (st.st_size > 0)
n = st.st_size + 1;
}
@@ -986,7 +1093,7 @@ int get_proc_field(const char *filename, const char *pattern, const char *termin
assert(pattern);
assert(field);
- r = read_full_file(filename, &status, NULL);
+ r = read_full_virtual_file(filename, &status, NULL);
if (r < 0)
return r;
diff --git a/src/basic/fileio.h b/src/basic/fileio.h
index 77e6206e95..c6ad375b8d 100644
--- a/src/basic/fileio.h
+++ b/src/basic/fileio.h
@@ -38,6 +38,7 @@ int write_string_filef(const char *fn, WriteStringFileFlags flags, const char *f
int read_one_line_file(const char *fn, char **line);
int read_full_file(const char *fn, char **contents, size_t *size);
int read_full_stream(FILE *f, char **contents, size_t *size);
+int read_full_virtual_file(const char *filename, char **ret_contents, size_t *ret_size);
int verify_file(const char *fn, const char *blob, bool accept_extra_nl);
diff --git a/src/libsystemd/sd-device/sd-device.c b/src/libsystemd/sd-device/sd-device.c
index be29053f8c..49750ba9d7 100644
--- a/src/libsystemd/sd-device/sd-device.c
+++ b/src/libsystemd/sd-device/sd-device.c
@@ -1798,7 +1798,7 @@ _public_ int sd_device_get_sysattr_value(sd_device *device, const char *sysattr,
size_t size;
/* read attribute value */
- r = read_full_file(path, &value, &size);
+ r = read_full_virtual_file(path, &value, &size);
if (r < 0)
return r;

118
SOURCES/0852-Use-BIOS-characteristics-to-distinguish-EC2-bare-met.patch Normal file
View File

@ -0,0 +1,118 @@
From 44cbd79562ed55a8b0f2e5b5dc708265568ed9f8 Mon Sep 17 00:00:00 2001
From: Noah Meyerhans <nmeyerha@amazon.com>
Date: Fri, 30 Apr 2021 09:30:52 -0700
Subject: [PATCH] Use BIOS characteristics to distinguish EC2 bare-metal from
VMs
DMI vendor information fields do not provide enough information for us to
distinguish between Amazon EC2 virtual machines and bare-metal instances.
SMBIOS provides a BIOS Information
table (https://www.dmtf.org/sites/default/files/standards/documents/DSP0134_3.4.0.pdf
Ch. 7) that provides a field to indicate that the current machine is a virtual
machine. On EC2 virtual machine instances, this field is set, while bare-metal
instances leave this unset, so we inspect the field via the kernel's
/sys/firemware/dmi/entries interface.
Fixes #18929
(cherry picked from commit ce35037928f4c4c931088256853f07804ec7d235)
Related: #2117948
---
src/basic/virt.c | 65 +++++++++++++++++++++++++++++++++++++++++++++---
1 file changed, 61 insertions(+), 4 deletions(-)
diff --git a/src/basic/virt.c b/src/basic/virt.c
index 6e4c702051..00d1c894e6 100644
--- a/src/basic/virt.c
+++ b/src/basic/virt.c
@@ -22,6 +22,12 @@
#include "string-util.h"
#include "virt.h"
+enum {
+ SMBIOS_VM_BIT_SET,
+ SMBIOS_VM_BIT_UNSET,
+ SMBIOS_VM_BIT_UNKNOWN,
+};
+
static const char *const vm_table[_VIRTUALIZATION_MAX] = {
[VIRTUALIZATION_XEN] = "XenVMMXenVMM",
[VIRTUALIZATION_KVM] = "KVMKVMKVM",
@@ -131,9 +137,8 @@ static int detect_vm_device_tree(void) {
#endif
}
-static int detect_vm_dmi(void) {
#if defined(__i386__) || defined(__x86_64__) || defined(__arm__) || defined(__aarch64__)
-
+static int detect_vm_dmi_vendor(void) {
static const char *const dmi_vendors[] = {
"/sys/class/dmi/id/product_name", /* Test this before sys_vendor to detect KVM over QEMU */
"/sys/class/dmi/id/sys_vendor",
@@ -179,11 +184,63 @@ static int detect_vm_dmi(void) {
return dmi_vendor_table[j].id;
}
}
-#endif
+ return VIRTUALIZATION_NONE;
+}
+
+static int detect_vm_smbios(void) {
+ /* The SMBIOS BIOS Charateristics Extension Byte 2 (Section 2.1.2.2 of
+ * https://www.dmtf.org/sites/default/files/standards/documents/DSP0134_3.4.0.pdf), specifies that
+ * the 4th bit being set indicates a VM. The BIOS Characteristics table is exposed via the kernel in
+ * /sys/firmware/dmi/entries/0-0. Note that in the general case, this bit being unset should not
+ * imply that the system is running on bare-metal. For example, QEMU 3.1.0 (with or without KVM)
+ * with SeaBIOS does not set this bit. */
+ _cleanup_free_ char *s = NULL;
+ size_t readsize;
+ int r;
+
+ r = read_full_virtual_file("/sys/firmware/dmi/entries/0-0/raw", &s, &readsize);
+ if (r < 0) {
+ log_debug_errno(r, "Unable to read /sys/firmware/dmi/entries/0-0/raw, ignoring: %m");
+ return SMBIOS_VM_BIT_UNKNOWN;
+ }
+ if (readsize < 20 || s[1] < 20) {
+ /* The spec indicates that byte 1 contains the size of the table, 0x12 + the number of
+ * extension bytes. The data we're interested in is in extension byte 2, which would be at
+ * 0x13. If we didn't read that much data, or if the BIOS indicates that we don't have that
+ * much data, we don't infer anything from the SMBIOS. */
+ log_debug("Only read %zu bytes from /sys/firmware/dmi/entries/0-0/raw (expected 20)", readsize);
+ return SMBIOS_VM_BIT_UNKNOWN;
+ }
- log_debug("No virtualization found in DMI");
+ uint8_t byte = (uint8_t) s[19];
+ if (byte & (1U<<4)) {
+ log_debug("DMI BIOS Extension table indicates virtualization");
+ return SMBIOS_VM_BIT_SET;
+ }
+ log_debug("DMI BIOS Extension table does not indicate virtualization");
+ return SMBIOS_VM_BIT_UNSET;
+}
+#endif /* defined(__i386__) || defined(__x86_64__) || defined(__arm__) || defined(__aarch64__) */
+
+static int detect_vm_dmi(void) {
+#if defined(__i386__) || defined(__x86_64__) || defined(__arm__) || defined(__aarch64__)
+
+ int r;
+ r = detect_vm_dmi_vendor();
+ /* The DMI vendor tables in /sys/class/dmi/id don't help us distinguish between Amazon EC2
+ * virtual machines and bare-metal instances, so we need to look at SMBIOS. */
+ if (r == VIRTUALIZATION_AMAZON && detect_vm_smbios() == SMBIOS_VM_BIT_UNSET)
+ return VIRTUALIZATION_NONE;
+
+ /* If we haven't identified a VM, but the firmware indicates that there is one, indicate as much. We
+ * have no further information about what it is. */
+ if (r == VIRTUALIZATION_NONE && detect_vm_smbios() == SMBIOS_VM_BIT_SET)
+ return VIRTUALIZATION_VM_OTHER;
+ return r;
+#else
return VIRTUALIZATION_NONE;
+#endif
}
static int detect_vm_xen(void) {

37
SOURCES/0853-device-drop-refuse_after.patch Normal file
View File

@ -0,0 +1,37 @@
From bb9d00035c00b8590c389e66b5d94334bbb7379d Mon Sep 17 00:00:00 2001
From: Franck Bui <fbui@suse.com>
Date: Mon, 30 Mar 2020 10:49:29 +0200
Subject: [PATCH] device: drop refuse_after
Scheduling devices after a given unit can be useful to start device *jobs* at a
specific time in the transaction, see commit 4195077ab4c823c.
This (hidden) change was introduced by commit eef85c4a3f8054d2.
(cherry picked from commit b862c25716520d9381d5a841dba0f0c14e9c970a)
[dtardon: This picks just the minimal relevant change from
c80a9a33d04fb4381327a69ce929c94a9f1d0e6c and
b862c25716520d9381d5a841dba0f0c14e9c970a]
Resolves: #2043524
---
src/core/unit.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/src/core/unit.c b/src/core/unit.c
index dfe0c243ef..9be2a0c326 100644
--- a/src/core/unit.c
+++ b/src/core/unit.c
@@ -2841,8 +2841,9 @@ int unit_add_dependency(
return 0;
}
- if ((d == UNIT_BEFORE && other->type == UNIT_DEVICE) ||
- (d == UNIT_AFTER && u->type == UNIT_DEVICE)) {
+ /* Note that ordering a device unit after a unit is permitted since it
+ * allows to start its job running timeout at a specific time. */
+ if (d == UNIT_BEFORE && other->type == UNIT_DEVICE) {
log_unit_warning(u, "Dependency Before=%s ignored (.device units cannot be delayed)", other->id);
return 0;
}

44
SOURCES/0854-manager-limit-access-to-private-dbus-socket.patch Normal file
View File

@ -0,0 +1,44 @@
From 004130ae74688eb321aadc05192bab69fe5cbcbf Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Zbigniew=20J=C4=99drzejewski-Szmek?= <zbyszek@in.waw.pl>
Date: Fri, 22 Jul 2022 11:45:12 +0200
Subject: [PATCH] manager: limit access to private dbus socket
For the system manager, /run/systemd/private is publicly accessible, because
/run/systemd is 0755, and /run/systemd/private is 0777. For the user manager,
/run/user/<uid> is 0700, and /run/user/<uid>/systemd/private is 0777. This
does not directly cause any security issue because we check the sender in
bus_check_peercred (ucred.uid != 0 && ucred.uid != geteuid()).
But it makes sense to limit access to the socket to avoid wasting time in PID1.
Somebody could send messages there that'd we'd reject anyway. It also makes
things more explicit.
(cherry picked from commit df1cbd1adf26071aab41d96e054452a3d66103a4)
Resolves: #2119405
---
src/core/dbus.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/src/core/dbus.c b/src/core/dbus.c
index 66d838cdb4..ec6c52cb85 100644
--- a/src/core/dbus.c
+++ b/src/core/dbus.c
@@ -42,6 +42,7 @@
#include "string-util.h"
#include "strv.h"
#include "strxcpyx.h"
+#include "umask-util.h"
#include "user-util.h"
#define CONNECTIONS_MAX 4096
@@ -1019,7 +1020,8 @@ int bus_init_private(Manager *m) {
if (fd < 0)
return log_error_errno(errno, "Failed to allocate private socket: %m");
- r = bind(fd, &sa.sa, salen);
+ RUN_WITH_UMASK(0077)
+ r = bind(fd, &sa.sa, salen);
if (r < 0)
return log_error_errno(errno, "Failed to bind private socket: %m");

35
SOURCES/0855-journalctl-do-not-treat-EINTR-as-an-error-when-waiti.patch Normal file
View File

@ -0,0 +1,35 @@
From b0574acc0bddceb0af47f6cce327a87041ab4b52 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Zbigniew=20J=C4=99drzejewski-Szmek?= <zbyszek@in.waw.pl>
Date: Sun, 11 Nov 2018 12:33:06 +0100
Subject: [PATCH] journalctl: do not treat EINTR as an error when waiting for
events
Fixup for 2a1e0f2228bbdfbc18635e959f47df7da50b62fe. Fixes #10724.
Reproducer: start 'journalctl -f' in a terminal window, change window size.
(cherry picked from commit 8e143a123276a9636987b08f555603927ca9e186)
Resolves: #2161683
---
src/journal/journalctl.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/src/journal/journalctl.c b/src/journal/journalctl.c
index fa83dce562..228cfe7e49 100644
--- a/src/journal/journalctl.c
+++ b/src/journal/journalctl.c
@@ -2084,8 +2084,13 @@ static int wait_for_change(sd_journal *j, int poll_fd) {
if (r < 0)
return log_error_errno(r, "Failed to determine journal waiting time: %m");
- if (ppoll(pollfds, ELEMENTSOF(pollfds), timeout == USEC_INFINITY ? NULL : timespec_store(&ts, timeout), NULL) < 0)
+ if (ppoll(pollfds, ELEMENTSOF(pollfds),
+ timeout == USEC_INFINITY ? NULL : timespec_store(&ts, timeout), NULL) < 0) {
+ if (errno == EINTR)
+ return 0;
+
return log_error_errno(errno, "Couldn't wait for journal event: %m");
+ }
if (pollfds[1].revents & (POLLHUP|POLLERR)) { /* STDOUT has been closed? */
log_debug("Standard output has been closed.");

12
SOURCES/0835-core-bring-manager_startup-and-manager_reload-more-i.patch → SOURCES/0856-core-bring-manager_startup-and-manager_reload-more-i.patch
View File

@ -1,4 +1,4 @@
From 2385f40cc80d723b105add1496de292c3241595c Mon Sep 17 00:00:00 2001 From b9dd7ee5f4d0f6d51899d7e14ac7ef2fd2840b8f Mon Sep 17 00:00:00 2001
From: Lennart Poettering <lennart@poettering.net> From: Lennart Poettering <lennart@poettering.net>
Date: Tue, 9 Oct 2018 17:37:57 +0200 Date: Tue, 9 Oct 2018 17:37:57 +0200
Subject: [PATCH] core: bring manager_startup() and manager_reload() more Subject: [PATCH] core: bring manager_startup() and manager_reload() more
@ -26,16 +26,16 @@ This makes a number of changes:
(cherry picked from commit 3ad2afb6a204513c7834c64ab864e40169874390) (cherry picked from commit 3ad2afb6a204513c7834c64ab864e40169874390)
Resolves: #2164049 Resolves: #2059633
--- ---
src/core/manager.c | 14 +++++++++++--- src/core/manager.c | 14 +++++++++++---
1 file changed, 11 insertions(+), 3 deletions(-) 1 file changed, 11 insertions(+), 3 deletions(-)
diff --git a/src/core/manager.c b/src/core/manager.c diff --git a/src/core/manager.c b/src/core/manager.c
index f4611e6f8f..49d6d250ef 100644 index e083596e58..4a9f9bfcf9 100644
--- a/src/core/manager.c --- a/src/core/manager.c
+++ b/src/core/manager.c +++ b/src/core/manager.c
@@ -1660,12 +1660,12 @@ int manager_startup(Manager *m, FILE *serialization, FDSet *fds) { @@ -1665,12 +1665,12 @@ int manager_startup(Manager *m, FILE *serialization, FDSet *fds) {
/* Release any dynamic users no longer referenced */ /* Release any dynamic users no longer referenced */
dynamic_user_vacuum(m, true); dynamic_user_vacuum(m, true);
@ -50,7 +50,7 @@ index f4611e6f8f..49d6d250ef 100644
if (serialization) { if (serialization) {
assert(m->n_reloading > 0); assert(m->n_reloading > 0);
m->n_reloading--; m->n_reloading--;
@@ -1676,6 +1676,13 @@ int manager_startup(Manager *m, FILE *serialization, FDSet *fds) { @@ -1681,6 +1681,13 @@ int manager_startup(Manager *m, FILE *serialization, FDSet *fds) {
m->send_reloading_done = true; m->send_reloading_done = true;
} }
@ -64,7 +64,7 @@ index f4611e6f8f..49d6d250ef 100644
/* Let's finally catch up with any changes that took place while we were reloading/reexecing */ /* Let's finally catch up with any changes that took place while we were reloading/reexecing */
manager_catchup(m); manager_catchup(m);
@@ -3500,7 +3507,8 @@ int manager_reload(Manager *m) { @@ -3505,7 +3512,8 @@ int manager_reload(Manager *m) {
lookup_paths_reduce(&m->lookup_paths); lookup_paths_reduce(&m->lookup_paths);
manager_build_unit_path_cache(m); manager_build_unit_path_cache(m);

35
SOURCES/0857-pam-add-a-call-to-pam_namespace.patch Normal file
View File

@ -0,0 +1,35 @@
From b1b7aaf83414c5b0bed6e61d38aefe29a21fdbcf Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Zbigniew=20J=C4=99drzejewski-Szmek?= <zbyszek@in.waw.pl>
Date: Wed, 23 Nov 2022 16:09:56 +0100
Subject: [PATCH] pam: add a call to pam_namespace
A call to pam_namespace is required so that children of user@.service end up in
a namespace as expected. pam_namespace gets called as part of the stack that
creates a session (login, sshd, gdm, etc.) and those processes end up in a
namespace, but it also needs to be called from our stack which is parallel and
descends from pid1 itself.
The call to pam_namespace is similar to the call to pam_keyinit that was added
in ab79099d1684457d040ee7c28b2012e8c1ea9a4f. The pam stack for user@.service
creates a new session which is disconnected from the parent environment. Both
calls are not suitable for inclusion in the shared part of the stack (e.g.
@system-auth on Fedora/RHEL systems), because for example su/sudo/runuser
should not include them.
(cherry picked from commit 0ef48896d9f23b9fd547a532a4e6e6b8f8b12901)
Resolves: #1861836
---
src/login/systemd-user.m4 | 1 +
1 file changed, 1 insertion(+)
diff --git a/src/login/systemd-user.m4 b/src/login/systemd-user.m4
index 20c8999331..eb291beaed 100644
--- a/src/login/systemd-user.m4
+++ b/src/login/systemd-user.m4
@@ -9,4 +9,5 @@ session required pam_selinux.so nottys open
)m4_dnl
session required pam_loginuid.so
session optional pam_keyinit.so force revoke
+session required pam_namespace.so
session optional pam_systemd.so

31
SOURCES/0858-virt-Support-detection-for-ARM64-Hyper-V-guests.patch Normal file
View File

@ -0,0 +1,31 @@
From 4b573adbcc040fa50f1130cb8cf1bdb9559565cf Mon Sep 17 00:00:00 2001
From: Boqun Feng <boqun.feng@gmail.com>
Date: Wed, 13 Oct 2021 11:32:09 +0800
Subject: [PATCH] virt: Support detection for ARM64 Hyper-V guests
The detection of Microsoft Hyper-V VMs is done by cpuid currently,
however there is no cpuid on ARM64. And since ARM64 is now a supported
architecture for Microsoft Hyper-V guests[1], then use DMI tables to
detect a Hyper-V guest, which is more generic and works for ARM64.
[1]: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=7aff79e297ee1aa0126924921fd87a4ae59d2467
(cherry picked from commit 506bbc8569014253ea8614b680ccbc4fc2513a87)
Resolves: #2158307
---
src/basic/virt.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/src/basic/virt.c b/src/basic/virt.c
index 00d1c894e6..cc95097101 100644
--- a/src/basic/virt.c
+++ b/src/basic/virt.c
@@ -162,6 +162,7 @@ static int detect_vm_dmi_vendor(void) {
{ "Parallels", VIRTUALIZATION_PARALLELS },
/* https://wiki.freebsd.org/bhyve */
{ "BHYVE", VIRTUALIZATION_BHYVE },
+ { "Microsoft", VIRTUALIZATION_MICROSOFT },
};
unsigned i;
int r;

41
SOURCES/0859-virt-Fix-the-detection-for-Hyper-V-VMs.patch Normal file
View File

@ -0,0 +1,41 @@
From e732bc987f2f779e89f30193bf694e0456ab7ce0 Mon Sep 17 00:00:00 2001
From: Boqun Feng <boqun.feng@gmail.com>
Date: Tue, 23 Nov 2021 15:09:26 +0800
Subject: [PATCH] virt: Fix the detection for Hyper-V VMs
Use product_version instead of product_name in DMI table and the string
"Hyper-V" to avoid misdetection.
Fixes: #21468
Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
(cherry picked from commit 76eec0649936d9ae2f9087769f463feaf0cf5cb4)
Related: #2158307
---
src/basic/virt.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/src/basic/virt.c b/src/basic/virt.c
index cc95097101..f750a0463f 100644
--- a/src/basic/virt.c
+++ b/src/basic/virt.c
@@ -143,7 +143,8 @@ static int detect_vm_dmi_vendor(void) {
"/sys/class/dmi/id/product_name", /* Test this before sys_vendor to detect KVM over QEMU */
"/sys/class/dmi/id/sys_vendor",
"/sys/class/dmi/id/board_vendor",
- "/sys/class/dmi/id/bios_vendor"
+ "/sys/class/dmi/id/bios_vendor",
+ "/sys/class/dmi/id/product_version" /* For Hyper-V VMs test */
};
static const struct {
@@ -162,7 +163,7 @@ static int detect_vm_dmi_vendor(void) {
{ "Parallels", VIRTUALIZATION_PARALLELS },
/* https://wiki.freebsd.org/bhyve */
{ "BHYVE", VIRTUALIZATION_BHYVE },
- { "Microsoft", VIRTUALIZATION_MICROSOFT },
+ { "Hyper-V", VIRTUALIZATION_MICROSOFT },
};
unsigned i;
int r;

4
SOURCES/0836-basic-add-STRERROR-wrapper-for-strerror_r.patch → SOURCES/0860-basic-add-STRERROR-wrapper-for-strerror_r.patch
View File

@ -1,11 +1,11 @@
From 84a8245cf58fa7edea61eab5f1ebd86a3944c388 Mon Sep 17 00:00:00 2001 From 9c95d8dda42de288a57638a44dd5ea967469063d Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Zbigniew=20J=C4=99drzejewski-Szmek?= <zbyszek@in.waw.pl> From: =?UTF-8?q?Zbigniew=20J=C4=99drzejewski-Szmek?= <zbyszek@in.waw.pl>
Date: Fri, 7 Oct 2022 12:28:31 +0200 Date: Fri, 7 Oct 2022 12:28:31 +0200
Subject: [PATCH] basic: add STRERROR() wrapper for strerror_r() Subject: [PATCH] basic: add STRERROR() wrapper for strerror_r()
(cherry picked from commit 2c5d05b3cd986568105d67891e4010b868dea24f) (cherry picked from commit 2c5d05b3cd986568105d67891e4010b868dea24f)
Related: #2155519 Related: #2155520
--- ---
src/basic/util.h | 10 ++++++++++ src/basic/util.h | 10 ++++++++++
src/test/test-util.c | 40 ++++++++++++++++++++++++++++++++++++++++ src/test/test-util.c | 40 ++++++++++++++++++++++++++++++++++++++++

4
SOURCES/0837-coredump-put-context-array-into-a-struct.patch → SOURCES/0861-coredump-put-context-array-into-a-struct.patch
View File

@ -1,4 +1,4 @@
From e256e03710abfd96e74d0b278c2b7186e3c8888b Mon Sep 17 00:00:00 2001 From f53c6620c55488e2a3bd92957b21b6b95a7a3d35 Mon Sep 17 00:00:00 2001
From: David Tardon <dtardon@redhat.com> From: David Tardon <dtardon@redhat.com>
Date: Thu, 12 Jan 2023 15:47:09 +0100 Date: Thu, 12 Jan 2023 15:47:09 +0100
Subject: [PATCH] coredump: put context array into a struct Subject: [PATCH] coredump: put context array into a struct
@ -7,7 +7,7 @@ Subject: [PATCH] coredump: put context array into a struct
but does just the minimal change to introduce the Context struct that is but does just the minimal change to introduce the Context struct that is
needed by the following commit.] needed by the following commit.]
Related: #2155519 Related: #2155520
--- ---
src/coredump/coredump.c | 208 +++++++++++++++++++++------------------- src/coredump/coredump.c | 208 +++++++++++++++++++++-------------------
1 file changed, 108 insertions(+), 100 deletions(-) 1 file changed, 108 insertions(+), 100 deletions(-)

4
SOURCES/0838-coredump-do-not-allow-user-to-access-coredumps-with-.patch → SOURCES/0862-coredump-do-not-allow-user-to-access-coredumps-with-.patch
View File

@ -1,4 +1,4 @@
From 9fa7f77c5a3d92fe03d6eec8384043555ea16dc6 Mon Sep 17 00:00:00 2001 From d178865d3d9940423f4d99360e3dc2fcaf0b2c96 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Zbigniew=20J=C4=99drzejewski-Szmek?= <zbyszek@in.waw.pl> From: =?UTF-8?q?Zbigniew=20J=C4=99drzejewski-Szmek?= <zbyszek@in.waw.pl>
Date: Mon, 28 Nov 2022 12:12:55 +0100 Date: Mon, 28 Nov 2022 12:12:55 +0100
Subject: [PATCH] coredump: do not allow user to access coredumps with changed Subject: [PATCH] coredump: do not allow user to access coredumps with changed
@ -36,7 +36,7 @@ ppc64el, but not the whole coredump handling.
(cherry picked from commit 3e4d0f6cf99f8677edd6a237382a65bfe758de03) (cherry picked from commit 3e4d0f6cf99f8677edd6a237382a65bfe758de03)
Resolves: #2155519 Resolves: #2155520
--- ---
src/coredump/coredump.c | 190 ++++++++++++++++++++++++++++++++++++++-- src/coredump/coredump.c | 190 ++++++++++++++++++++++++++++++++++++++--
1 file changed, 182 insertions(+), 8 deletions(-) 1 file changed, 182 insertions(+), 8 deletions(-)

70
SOURCES/0863-logind-remember-our-idle-state-and-use-it-to-detect-.patch Normal file
View File

@ -0,0 +1,70 @@
From a55bdb007ca24900a704f72359305b04125d9e05 Mon Sep 17 00:00:00 2001
From: Michal Sekletar <msekleta@redhat.com>
Date: Wed, 20 Apr 2022 10:13:43 +0200
Subject: [PATCH] logind: remember our idle state and use it to detect idle
level transitions
Fixes #16391
(cherry picked from commit 4e2cfb778b9ed7f22ee98f48f28cf8678d25ad32)
Resolved: #1866955
---
src/login/logind.c | 23 +++++++++++++++++++----
src/login/logind.h | 1 +
2 files changed, 20 insertions(+), 4 deletions(-)
diff --git a/src/login/logind.c b/src/login/logind.c
index 6b576dad0d..bb1d3f3523 100644
--- a/src/login/logind.c
+++ b/src/login/logind.c
@@ -1027,18 +1027,33 @@ static int manager_dispatch_idle_action(sd_event_source *s, uint64_t t, void *us
n = now(CLOCK_MONOTONIC);
r = manager_get_idle_hint(m, &since);
- if (r <= 0)
+ if (r <= 0) {
/* Not idle. Let's check if after a timeout it might be idle then. */
elapse = n + m->idle_action_usec;
- else {
+ m->was_idle = false;
+ } else {
+
/* Idle! Let's see if it's time to do something, or if
* we shall sleep for longer. */
if (n >= since.monotonic + m->idle_action_usec &&
(m->idle_action_not_before_usec <= 0 || n >= m->idle_action_not_before_usec + m->idle_action_usec)) {
- log_info("System idle. Taking action.");
+ bool is_edge = false;
+
+ /* We weren't idle previously or some activity happened while we were sleeping, and now we are
+ * idle. Let's remember that for the next time and make this an edge transition. */
+ if (!m->was_idle || since.monotonic >= m->idle_action_not_before_usec) {
+ is_edge = true;
+ m->was_idle = true;
+ }
+
+ if (m->idle_action == HANDLE_LOCK && !is_edge)
+ /* We are idle and we were before so we are actually not taking any action. */
+ log_debug("System idle.");
+ else
+ log_info("System idle. Doing %s operation.", handle_action_to_string(m->idle_action));
- manager_handle_action(m, 0, m->idle_action, false, false);
+ manager_handle_action(m, 0, m->idle_action, false, is_edge);
m->idle_action_not_before_usec = n;
}
diff --git a/src/login/logind.h b/src/login/logind.h
index 606adf4fe6..b9b4a5113f 100644
--- a/src/login/logind.h
+++ b/src/login/logind.h
@@ -101,6 +101,7 @@ struct Manager {
usec_t idle_action_usec;
usec_t idle_action_not_before_usec;
HandleAction idle_action;
+ bool was_idle;
usec_t stop_idle_session_usec;

103
SOURCES/0864-test-import-logind-test-from-debian-ubuntu-test-suit.patch Normal file
View File

@ -0,0 +1,103 @@
From 1925845dc10330e4b48fec68333fac6ef2b7bf5c Mon Sep 17 00:00:00 2001
From: Yu Watanabe <watanabe.yu+github@gmail.com>
Date: Wed, 1 Jun 2022 08:56:08 +0900
Subject: [PATCH] test: import logind test from debian/ubuntu test suite
[dtardon: Picked just the scaffolding, not the tests themselves.]
(cherry picked from commit 9c94ab0f6ff22da4278a6e9a93ddc480607c55ac)
Related: #1866955
---
test/TEST-35-LOGIN/Makefile | 1 +
test/TEST-35-LOGIN/test.sh | 55 +++++++++++++++++++++++++++++++++
test/TEST-35-LOGIN/testsuite.sh | 9 ++++++
3 files changed, 65 insertions(+)
create mode 120000 test/TEST-35-LOGIN/Makefile
create mode 100755 test/TEST-35-LOGIN/test.sh
create mode 100755 test/TEST-35-LOGIN/testsuite.sh
diff --git a/test/TEST-35-LOGIN/Makefile b/test/TEST-35-LOGIN/Makefile
new file mode 120000
index 0000000000..e9f93b1104
--- /dev/null
+++ b/test/TEST-35-LOGIN/Makefile
@@ -0,0 +1 @@
+../TEST-01-BASIC/Makefile
\ No newline at end of file
diff --git a/test/TEST-35-LOGIN/test.sh b/test/TEST-35-LOGIN/test.sh
new file mode 100755
index 0000000000..32410c8149
--- /dev/null
+++ b/test/TEST-35-LOGIN/test.sh
@@ -0,0 +1,55 @@
+#!/usr/bin/env bash
+# SPDX-License-Identifier: LGPL-2.1-or-later
+set -e
+
+TEST_DESCRIPTION="LOGIN"
+
+# shellcheck source=test/test-functions
+. "${TEST_BASE_DIR:?}/test-functions"
+
+test_setup() {
+ create_empty_image
+ mkdir -p "$TESTDIR"/root
+ mount "${LOOPDEV}p1" "$TESTDIR"/root
+
+ (
+ LOG_LEVEL=5
+ eval $(udevadm info --export --query=env --name="${LOOPDEV}p2")
+
+ setup_basic_environment
+
+ inst_binary pkill
+ inst_binary useradd
+ inst_binary userdel
+
+ # mask some services that we do not want to run in these tests
+ ln -fs /dev/null "$initdir"/etc/systemd/system/systemd-hwdb-update.service
+ ln -fs /dev/null "$initdir"/etc/systemd/system/systemd-journal-catalog-update.service
+ ln -fs /dev/null "$initdir"/etc/systemd/system/systemd-networkd.service
+ ln -fs /dev/null "$initdir"/etc/systemd/system/systemd-networkd.socket
+ ln -fs /dev/null "$initdir"/etc/systemd/system/systemd-resolved.service
+ ln -fs /dev/null "$initdir"/etc/systemd/system/systemd-machined.service
+
+ # setup the testsuite service
+ cat >"$initdir"/etc/systemd/system/testsuite.service <<EOF
+[Unit]
+Description=Testsuite service
+
+[Service]
+ExecStart=/bin/bash -x /testsuite.sh
+Type=oneshot
+StandardOutput=tty
+StandardError=tty
+NotifyAccess=all
+EOF
+ cp testsuite.sh "$initdir"/
+
+ setup_testsuite
+ ) || return 1
+ setup_nspawn_root
+
+ ddebug "umount $TESTDIR/root"
+ umount "$TESTDIR"/root
+}
+
+do_test "$@"
diff --git a/test/TEST-35-LOGIN/testsuite.sh b/test/TEST-35-LOGIN/testsuite.sh
new file mode 100755
index 0000000000..9855b4bc80
--- /dev/null
+++ b/test/TEST-35-LOGIN/testsuite.sh
@@ -0,0 +1,9 @@
+#!/usr/bin/env bash
+# SPDX-License-Identifier: LGPL-2.1-or-later
+set -eux
+set -o pipefail
+
+: >/failed
+
+touch /testok
+rm /failed

37
SOURCES/0865-test-introduce-inst_recursive-helper-function.patch Normal file
View File

@ -0,0 +1,37 @@
From 21c281585983fe2efc2b74d712d98ff7e6c013b1 Mon Sep 17 00:00:00 2001
From: Yu Watanabe <watanabe.yu+github@gmail.com>
Date: Tue, 14 Jun 2022 09:11:00 +0900
Subject: [PATCH] test: introduce inst_recursive() helper function
(cherry picked from commit da0465dc95388afc15598357452afef85035c639)
Related: #1866955
---
test/test-functions | 14 ++++++++++++++
1 file changed, 14 insertions(+)
diff --git a/test/test-functions b/test/test-functions
index 19363be858..9606a1b085 100644
--- a/test/test-functions
+++ b/test/test-functions
@@ -1224,6 +1224,20 @@ inst_any() {
return 1
}
+inst_recursive() {
+ local p item
+
+ for p in "$@"; do
+ while read -r item; do
+ if [[ -d "$item" ]]; then
+ inst_dir "$item"
+ elif [[ -f "$item" ]]; then
+ inst_simple "$item"
+ fi
+ done < <(find "$p" 2>/dev/null)
+ done
+}
+
# dracut_install [-o ] <file> [<file> ... ]
# Install <file> to the initramfs image
# -o optionally install the <file> and don't fail, if it is not there

148
SOURCES/0866-tests-verify-that-Lock-D-Bus-signal-is-sent-when-Idl.patch Normal file
View File

@ -0,0 +1,148 @@
From 638c2418e705410344e07e77f944530df0f4608f Mon Sep 17 00:00:00 2001
From: Michal Sekletar <msekleta@redhat.com>
Date: Fri, 1 Jul 2022 12:59:57 +0200
Subject: [PATCH] tests: verify that Lock D-Bus signal is sent when
IdleAction=lock
(cherry picked from commit 181656fc0faa885d69bc34822b8e9b5de3fdf6bf)
Related: #1866955
---
test/TEST-35-LOGIN/test.sh | 11 +++-
test/TEST-35-LOGIN/testsuite.sh | 93 +++++++++++++++++++++++++++++++++
2 files changed, 103 insertions(+), 1 deletion(-)
diff --git a/test/TEST-35-LOGIN/test.sh b/test/TEST-35-LOGIN/test.sh
index 32410c8149..f83afcff49 100755
--- a/test/TEST-35-LOGIN/test.sh
+++ b/test/TEST-35-LOGIN/test.sh
@@ -14,14 +14,23 @@ test_setup() {
(
LOG_LEVEL=5
- eval $(udevadm info --export --query=env --name="${LOOPDEV}p2")
+ eval "$(udevadm info --export --query=env --name="${LOOPDEV}p2")"
setup_basic_environment
+ inst_binary awk
inst_binary pkill
inst_binary useradd
inst_binary userdel
+ if command -v expect >/dev/null && command -v tclsh >/dev/null ; then
+ # shellcheck disable=SC2016
+ version="$(tclsh <<< 'puts $tcl_version')"
+
+ dracut_install expect
+ inst_recursive /usr/lib64/tcl"$version" /usr/share/tcl"$version"
+ fi
+
# mask some services that we do not want to run in these tests
ln -fs /dev/null "$initdir"/etc/systemd/system/systemd-hwdb-update.service
ln -fs /dev/null "$initdir"/etc/systemd/system/systemd-journal-catalog-update.service
diff --git a/test/TEST-35-LOGIN/testsuite.sh b/test/TEST-35-LOGIN/testsuite.sh
index 9855b4bc80..e4d72beb74 100755
--- a/test/TEST-35-LOGIN/testsuite.sh
+++ b/test/TEST-35-LOGIN/testsuite.sh
@@ -3,7 +3,100 @@
set -eux
set -o pipefail
+setup_idle_action_lock() {
+ useradd testuser ||:
+
+ mkdir -p /run/systemd/logind.conf.d/
+ cat >/run/systemd/logind.conf.d/idle-action-lock.conf <<EOF
+[Login]
+IdleAction=lock
+IdleActionSec=1s
+EOF
+
+ mkdir -p /run/systemd/systemd-logind.service.d/
+ cat >/run/systemd/systemd-logind.service.d/debug.conf <<EOF
+[Service]
+Environment=SYSTEMD_LOG_LEVEL=debug
+EOF
+
+ systemctl restart systemd-logind.service
+}
+
+teardown_idle_action_lock() {(
+ set +ex
+ rm -f /run/systemd/logind.conf.d/idle-action-lock.conf
+ rm -f /run/systemd/systemd-logind.service.d/debug.conf
+ pkill -9 -u "$(id -u testuser)"
+ userdel -r testuser
+ systemctl restart systemd-logind.service
+)}
+
+test_lock_idle_action() {
+ if ! command -v expect >/dev/null ; then
+ echo >&2 "expect not installed, skiping test ${FUNCNAME[0]}"
+ return 0
+ fi
+
+ setup_idle_action_lock
+ trap teardown_idle_action_lock RETURN
+
+ if loginctl --no-legend | awk '{ print $3; }' | sort -u | grep -q testuser ; then
+ echo >&2 "Session of the \'testuser\' is already present."
+ return 1
+ fi
+
+ # IdleActionSec is set 1s but the accuracy of associated timer is 30s so we
+ # need to sleep in worst case for 31s to make sure timer elapsed. We sleep
+ # here for 35s to accomodate for any possible scheudling delays.
+ cat > /tmp/test.exp <<EOF
+spawn systemd-run -G -t -p PAMName=login -p User=testuser bash
+send "sleep 35\r"
+send "echo foobar\r"
+send "sleep 35\r"
+send "exit\r"
+interact
+wait
+EOF
+
+ ts="$(date '+%H:%M:%S')"
+ busctl --match "type='signal',sender='org.freedesktop.login1',interface='org.freedesktop.login1.Session',member='Lock'" monitor > dbus.log &
+
+ expect /tmp/test.exp &
+
+ # Sleep a bit to give expect time to spawn systemd-run before we check for
+ # the presence of resulting session.
+ sleep 2
+ if [ "$(loginctl --no-legend | awk '{ print $3; }' | sort -u | grep -c testuser)" != 1 ] ; then
+ echo >&2 "\'testuser\' is expected to have exactly one session running."
+ return 1
+ fi
+
+ wait %2
+ sleep 20
+ kill %1
+
+ # We slept for 35s , in that interval all sessions should have become idle
+ # and "Lock" signal should have been sent out. Then we wrote to tty to make
+ # session active again and next we slept for another 35s so sessions have
+ # become idle again. 'Lock' signal is sent out for each session, we have at
+ # least one session, so minimum of 2 "Lock" signals must have been sent.
+ if [ "$(grep -c Member=Lock dbus.log)" -lt 2 ]; then
+ echo >&2 "Too few 'Lock' D-Bus signal sent, expected at least 2."
+ return 1
+ fi
+
+ journalctl -b -u systemd-logind.service --since="$ts" > logind.log
+ if [ "$(grep -c 'System idle. Doing lock operation.' logind.log)" -lt 2 ]; then
+ echo >&2 "System haven't entered idle state at least 2 times."
+ return 1
+ fi
+
+ rm -f dbus.log logind.log
+}
+
: >/failed
+test_lock_idle_action
+
touch /testok
rm /failed

86
SOURCES/0867-systemctl-simplify-halt_main.patch Normal file
View File

@ -0,0 +1,86 @@
From 797b00e6a6f33d2b74beba02f678bf4d12e2146b Mon Sep 17 00:00:00 2001
From: Ludwig Nussel <ludwig.nussel@suse.de>
Date: Tue, 14 Dec 2021 17:27:05 +0100
Subject: [PATCH] systemctl: simplify halt_main()
The code at this point is not able to tell whether it was called as
halt/poweroff/reboot or shutdown with time "now".
The code also takes a shortcut to skip logind if called as root.
That however means asking shutdown for immediate action won't trigger a
wall message.
As per https://github.com/systemd/systemd/issues/8424#issuecomment-374677315
all commands should trigger a wall message.
That simplifies the code as we can try logind first always.
(cherry picked from commit adefc8789b63225662e50ceaa282f9553b5c64eb)
Resolves: #2053273
---
src/systemctl/systemctl.c | 44 ++++++++++++++++-----------------------
1 file changed, 18 insertions(+), 26 deletions(-)
diff --git a/src/systemctl/systemctl.c b/src/systemctl/systemctl.c
index b967550b97..4bedb52f2a 100644
--- a/src/systemctl/systemctl.c
+++ b/src/systemctl/systemctl.c
@@ -8658,34 +8658,23 @@ static int logind_schedule_shutdown(void) {
static int halt_main(void) {
int r;
- r = logind_check_inhibitors(arg_action);
- if (r < 0)
- return r;
-
+ /* always try logind first */
if (arg_when > 0)
- return logind_schedule_shutdown();
-
- if (geteuid() != 0) {
- if (arg_dry_run || arg_force > 0) {
- (void) must_be_root();
- return -EPERM;
- }
+ r = logind_schedule_shutdown();
+ else {
+ r = logind_check_inhibitors(arg_action);
+ if (r < 0)
+ return r;
- /* Try logind if we are a normal user and no special
- * mode applies. Maybe PolicyKit allows us to shutdown
- * the machine. */
- if (IN_SET(arg_action, ACTION_POWEROFF, ACTION_REBOOT, ACTION_HALT)) {
- r = logind_reboot(arg_action);
- if (r >= 0)
- return r;
- if (IN_SET(r, -EOPNOTSUPP, -EINPROGRESS))
- /* requested operation is not
- * supported on the local system or
- * already in progress */
- return r;
- /* on all other errors, try low-level operation */
- }
+ r = logind_reboot(arg_action);
}
+ if (r >= 0)
+ return r;
+ if (IN_SET(r, -EOPNOTSUPP, -EINPROGRESS))
+ /* Requested operation is not supported on the local system or already in
+ * progress */
+ return r;
+ /* on all other errors, try low-level operation */
/* In order to minimize the difference between operation with and
* without logind, we explicitly enable non-blocking mode for this,
@@ -8695,7 +8684,10 @@ static int halt_main(void) {
if (!arg_dry_run && !arg_force)
return start_with_fallback();
- assert(geteuid() == 0);
+ if (geteuid() != 0) {
+ (void) must_be_root();
+ return -EPERM;
+ }
if (!arg_no_wtmp) {
if (sd_booted() > 0)

51
SOURCES/0868-systemctl-shutdown-don-t-fallback-on-auth-fail.patch Normal file
View File

@ -0,0 +1,51 @@
From 1d63577410cde215c04921d62f435259a6b258d7 Mon Sep 17 00:00:00 2001
From: Ludwig Nussel <ludwig.nussel@suse.de>
Date: Mon, 20 Dec 2021 18:05:50 +0100
Subject: [PATCH] systemctl: shutdown don't fallback on auth fail
For shutdowns don't fall back to starting the target directly if talking
to logind failed with auth failure. That would just lead to another
polkit auth attempt.
(cherry picked from commit 38d55bf2641f345445cb4e6a5e5e808555591db2)
Related: #2053273
---
src/systemctl/systemctl.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/src/systemctl/systemctl.c b/src/systemctl/systemctl.c
index 4bedb52f2a..199f736f7f 100644
--- a/src/systemctl/systemctl.c
+++ b/src/systemctl/systemctl.c
@@ -3675,8 +3675,8 @@ static int start_special(int argc, char *argv[], void *userdata) {
r = logind_reboot(a);
if (r >= 0)
return r;
- if (IN_SET(r, -EOPNOTSUPP, -EINPROGRESS))
- /* requested operation is not supported or already in progress */
+ if (IN_SET(r, -EACCES, -EOPNOTSUPP, -EINPROGRESS))
+ /* Requested operation requires auth, is not supported or already in progress */
return r;
/* On all other errors, try low-level operation. In order to minimize the difference between
@@ -8644,7 +8644,7 @@ static int logind_schedule_shutdown(void) {
action,
arg_when);
if (r < 0)
- return log_warning_errno(r, "Failed to call ScheduleShutdown in logind, proceeding with immediate shutdown: %s", bus_error_message(&error, r));
+ return log_warning_errno(r, "Failed to schedule shutdown: %s", bus_error_message(&error, r));
if (!arg_quiet)
log_info("Shutdown scheduled for %s, use 'shutdown -c' to cancel.", format_timestamp(date, sizeof(date), arg_when));
@@ -8670,8 +8670,8 @@ static int halt_main(void) {
}
if (r >= 0)
return r;
- if (IN_SET(r, -EOPNOTSUPP, -EINPROGRESS))
- /* Requested operation is not supported on the local system or already in
+ if (IN_SET(r, -EACCES, -EOPNOTSUPP, -EINPROGRESS))
+ /* Requested operation requires auth, is not supported on the local system or already in
* progress */
return r;
/* on all other errors, try low-level operation */

82
SOURCES/0869-systemctl-reintroduce-the-original-halt_main.patch Normal file
View File

@ -0,0 +1,82 @@
From d36295d7c1b110d150b7af6e3354c28af4c4884d Mon Sep 17 00:00:00 2001
From: David Tardon <dtardon@redhat.com>
Date: Mon, 30 Jan 2023 14:27:24 +0100
Subject: [PATCH] systemctl: reintroduce the original halt_main()
RHEL-only
Related: #2053273
---
src/systemctl/systemctl.c | 59 +++++++++++++++++++++++++++++++++++++++
1 file changed, 59 insertions(+)
diff --git a/src/systemctl/systemctl.c b/src/systemctl/systemctl.c
index 199f736f7f..a26e4a913a 100644
--- a/src/systemctl/systemctl.c
+++ b/src/systemctl/systemctl.c
@@ -8655,6 +8655,65 @@ static int logind_schedule_shutdown(void) {
#endif
}
+static int halt_main_old(void) {
+ int r;
+
+ r = logind_check_inhibitors(arg_action);
+ if (r < 0)
+ return r;
+
+ if (arg_when > 0)
+ return logind_schedule_shutdown();
+
+ if (geteuid() != 0) {
+ if (arg_dry_run || arg_force > 0) {
+ (void) must_be_root();
+ return -EPERM;
+ }
+
+ /* Try logind if we are a normal user and no special
+ * mode applies. Maybe PolicyKit allows us to shutdown
+ * the machine. */
+ if (IN_SET(arg_action, ACTION_POWEROFF, ACTION_REBOOT, ACTION_HALT)) {
+ r = logind_reboot(arg_action);
+ if (r >= 0)
+ return r;
+ if (IN_SET(r, -EOPNOTSUPP, -EINPROGRESS))
+ /* requested operation is not
+ * supported on the local system or
+ * already in progress */
+ return r;
+ /* on all other errors, try low-level operation */
+ }
+ }
+
+ /* In order to minimize the difference between operation with and
+ * without logind, we explicitly enable non-blocking mode for this,
+ * as logind's shutdown operations are always non-blocking. */
+ arg_no_block = true;
+
+ if (!arg_dry_run && !arg_force)
+ return start_with_fallback();
+
+ assert(geteuid() == 0);
+
+ if (!arg_no_wtmp) {
+ if (sd_booted() > 0)
+ log_debug("Not writing utmp record, assuming that systemd-update-utmp is used.");
+ else {
+ r = utmp_put_shutdown();
+ if (r < 0)
+ log_warning_errno(r, "Failed to write utmp record: %m");
+ }
+ }
+
+ if (arg_dry_run)
+ return 0;
+
+ r = halt_now(arg_action);
+ return log_error_errno(r, "Failed to reboot: %m");
+}
+
static int halt_main(void) {
int r;

44
SOURCES/0870-systemctl-preserve-old-behavior-unless-requested.patch Normal file
View File

@ -0,0 +1,44 @@
From 74632586b46c7e88b09c57eec50f9c4aed254b98 Mon Sep 17 00:00:00 2001
From: David Tardon <dtardon@redhat.com>
Date: Mon, 30 Jan 2023 14:31:23 +0100
Subject: [PATCH] systemctl: preserve old behavior unless requested
Currently, the legacy shutdown commands ignore inhibitors and reboot
immediately if run by root. Let's preserve that behavior in RHEL-8 by
default. The new behavior can be turned on by those who want it by
exporting SYSTEMD_NEW_SHUTDOWN=1 .
RHEL-only
Related: #2053273
---
src/systemctl/systemctl.c | 9 ++++++++-
1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/src/systemctl/systemctl.c b/src/systemctl/systemctl.c
index a26e4a913a..1546119ce5 100644
--- a/src/systemctl/systemctl.c
+++ b/src/systemctl/systemctl.c
@@ -8714,7 +8714,7 @@ static int halt_main_old(void) {
return log_error_errno(r, "Failed to reboot: %m");
}
-static int halt_main(void) {
+static int halt_main_new(void) {
int r;
/* always try logind first */
@@ -8765,6 +8765,13 @@ static int halt_main(void) {
return log_error_errno(r, "Failed to reboot: %m");
}
+static int halt_main(void) {
+ if (getenv_bool("SYSTEMD_NEW_SHUTDOWN") > 0)
+ return halt_main_new();
+ else
+ return halt_main_old();
+}
+
static int runlevel_main(void) {
int r, runlevel, previous;

43
SOURCES/0871-pam_systemd-suppress-LOG_DEBUG-log-messages-if-debug.patch Normal file
View File

@ -0,0 +1,43 @@
From ea3910e561f043f5a131a846862955c77169da1b Mon Sep 17 00:00:00 2001
From: Lennart Poettering <lennart@poettering.net>
Date: Mon, 19 Nov 2018 11:39:45 +0100
Subject: [PATCH] pam_systemd: suppress LOG_DEBUG log messages if debugging is
off
In the PAM module we need to suppress LOG_DEBUG messages manually, if
debug logging is not on, as PAM won't do this for us. We did this
correctly for most log messages already, but two were missing. Let's fix
those too.
Fixes: #10822
(cherry picked from commit 2675747f3cdd6f1e6236bbb2f79abfa53fb307f1)
Resolves: #2170084
---
src/login/pam_systemd.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/src/login/pam_systemd.c b/src/login/pam_systemd.c
index 64e1b4d1bf..c87e980b18 100644
--- a/src/login/pam_systemd.c
+++ b/src/login/pam_systemd.c
@@ -476,7 +476,8 @@ _public_ PAM_EXTERN int pam_sm_open_session(
}
if (seat && !streq(seat, "seat0") && vtnr != 0) {
- pam_syslog(handle, LOG_DEBUG, "Ignoring vtnr %"PRIu32" for %s which is not seat0", vtnr, seat);
+ if (debug)
+ pam_syslog(handle, LOG_DEBUG, "Ignoring vtnr %"PRIu32" for %s which is not seat0", vtnr, seat);
vtnr = 0;
}
@@ -577,7 +578,8 @@ _public_ PAM_EXTERN int pam_sm_open_session(
r = sd_bus_call(bus, m, 0, &error, &reply);
if (r < 0) {
if (sd_bus_error_has_name(&error, BUS_ERROR_SESSION_BUSY)) {
- pam_syslog(handle, LOG_DEBUG, "Cannot create session: %s", bus_error_message(&error, r));
+ if (debug)
+ pam_syslog(handle, LOG_DEBUG, "Cannot create session: %s", bus_error_message(&error, r));
return PAM_SUCCESS;
} else {
pam_syslog(handle, LOG_ERR, "Failed to create session: %s", bus_error_message(&error, r));

50
SOURCES/0872-udev-net_id-introduce-naming-scheme-for-RHEL-8.8.patch Normal file
View File

@ -0,0 +1,50 @@
From 33351e103734188a4a30b88e7f2ea0613d628599 Mon Sep 17 00:00:00 2001
From: Jan Macku <jamacku@redhat.com>
Date: Thu, 16 Feb 2023 15:56:52 +0100
Subject: [PATCH] udev/net_id: introduce naming scheme for RHEL-8.8
RHEL-only
Resolves: #2170499
---
man/systemd.net-naming-scheme.xml | 6 ++++++
src/udev/udev-builtin-net_id.c | 2 ++
2 files changed, 8 insertions(+)
diff --git a/man/systemd.net-naming-scheme.xml b/man/systemd.net-naming-scheme.xml
index a567483995..3cc7719e99 100644
--- a/man/systemd.net-naming-scheme.xml
+++ b/man/systemd.net-naming-scheme.xml
@@ -328,6 +328,12 @@
for that, the limit is increased to now 65535.</para></listitem>
</varlistentry>
+ <varlistentry>
+ <term><constant>rhel-8.8</constant></term>
+
+ <para>Same as naming scheme <constant>rhel-8.7</constant>.</para>
+ </varlistentry>
+
<para>Note that <constant>latest</constant> may be used to denote the latest scheme known to this
particular version of systemd.</para>
</variablelist>
diff --git a/src/udev/udev-builtin-net_id.c b/src/udev/udev-builtin-net_id.c
index df84acf27c..ef2bb1b08e 100644
--- a/src/udev/udev-builtin-net_id.c
+++ b/src/udev/udev-builtin-net_id.c
@@ -141,6 +141,7 @@ typedef enum NamingSchemeFlags {
NAMING_RHEL_8_5 = NAMING_RHEL_8_4,
NAMING_RHEL_8_6 = NAMING_RHEL_8_4,
NAMING_RHEL_8_7 = NAMING_RHEL_8_4|NAMING_SLOT_FUNCTION_ID|NAMING_16BIT_INDEX,
+ NAMING_RHEL_8_8 = NAMING_RHEL_8_7,
_NAMING_SCHEME_FLAGS_INVALID = -1,
} NamingSchemeFlags;
@@ -161,6 +162,7 @@ static const NamingScheme naming_schemes[] = {
{ "rhel-8.5", NAMING_RHEL_8_5 },
{ "rhel-8.6", NAMING_RHEL_8_6 },
{ "rhel-8.7", NAMING_RHEL_8_7 },
+ { "rhel-8.8", NAMING_RHEL_8_8 },
/* … add more schemes here, as the logic to name devices is updated … */
};

176
SOURCES/0873-journald-add-API-to-move-logging-from-var-to-run-aga.patch Normal file
View File

@ -0,0 +1,176 @@
From 8f0a91b5192b72eb8b0f06e04ef3515d0397fcb8 Mon Sep 17 00:00:00 2001
From: Lennart Poettering <lennart@poettering.net>
Date: Fri, 5 Apr 2019 18:20:25 +0200
Subject: [PATCH] journald: add API to move logging from /var to /run again
We now have this nice little Varlink API, let's beef it up a bit.
[dtardon: This diverges from the upstream commit in two ways: One is
that the new action is bound to a signal, as varlink is not available.
The other is that this introduces a new "flag" file
/run/systemd/journal/relinquished, which is essentially the opposite of
/run/systemd/journal/flushed (hence at most one of them may exist at any
time).]
(cherry picked from commit b4e26d1d8e582d78e67ed766177f10e8e194404c)
Related: #1873540
---
src/journal/journald-server.c | 64 ++++++++++++++++++++++++++++-------
1 file changed, 51 insertions(+), 13 deletions(-)
diff --git a/src/journal/journald-server.c b/src/journal/journald-server.c
index 7632e2d9d0..279a32768c 100644
--- a/src/journal/journald-server.c
+++ b/src/journal/journald-server.c
@@ -283,13 +283,14 @@ static bool flushed_flag_is_set(void) {
return access("/run/systemd/journal/flushed", F_OK) >= 0;
}
-static int system_journal_open(Server *s, bool flush_requested) {
+static int system_journal_open(Server *s, bool flush_requested, bool relinquish_requested) {
const char *fn;
int r = 0;
if (!s->system_journal &&
IN_SET(s->storage, STORAGE_PERSISTENT, STORAGE_AUTO) &&
- (flush_requested || flushed_flag_is_set())) {
+ (flush_requested || flushed_flag_is_set()) &&
+ !relinquish_requested) {
/* If in auto mode: first try to create the machine
* path, but not the prefix.
@@ -331,7 +332,7 @@ static int system_journal_open(Server *s, bool flush_requested) {
fn = strjoina(s->runtime_storage.path, "/system.journal");
- if (s->system_journal) {
+ if (s->system_journal && !relinquish_requested) {
/* Try to open the runtime journal, but only
* if it already exists, so that we can flush
@@ -386,7 +387,7 @@ static JournalFile* find_journal(Server *s, uid_t uid) {
* else that's left the journals as NULL).
*
* Fixes https://github.com/systemd/systemd/issues/3968 */
- (void) system_journal_open(s, false);
+ (void) system_journal_open(s, false, false);
/* We split up user logs only on /var, not on /run. If the
* runtime file is open, we write to it exclusively, in order
@@ -964,7 +965,7 @@ int server_flush_to_var(Server *s, bool require_flag_file) {
char ts[FORMAT_TIMESPAN_MAX];
usec_t start;
unsigned n = 0;
- int r;
+ int r, k;
assert(s);
@@ -977,7 +978,7 @@ int server_flush_to_var(Server *s, bool require_flag_file) {
if (require_flag_file && !flushed_flag_is_set())
return 0;
- (void) system_journal_open(s, true);
+ (void) system_journal_open(s, true, false);
if (!s->system_journal)
return 0;
@@ -1056,6 +1057,13 @@ finish:
n),
NULL);
+ if (unlink("/run/systemd/journal/relinquished") < 0 && errno != ENOENT)
+ log_warning_errno(errno, "Failed to unlink /run/systemd/journal/relinquished, ignoring: %m");
+
+ k = touch("/run/systemd/journal/flushed");
+ if (k < 0)
+ log_warning_errno(k, "Failed to touch /run/systemd/journal/flushed, ignoring: %m");
+
return r;
}
@@ -1180,7 +1188,6 @@ int server_process_datagram(sd_event_source *es, int fd, uint32_t revents, void
static int dispatch_sigusr1(sd_event_source *es, const struct signalfd_siginfo *si, void *userdata) {
Server *s = userdata;
- int r;
assert(s);
@@ -1190,10 +1197,6 @@ static int dispatch_sigusr1(sd_event_source *es, const struct signalfd_siginfo *
server_sync(s);
server_vacuum(s, false);
- r = touch("/run/systemd/journal/flushed");
- if (r < 0)
- log_warning_errno(r, "Failed to touch /run/systemd/journal/flushed, ignoring: %m");
-
server_space_usage_message(s, NULL);
return 0;
}
@@ -1250,12 +1253,43 @@ static int dispatch_sigrtmin1(sd_event_source *es, const struct signalfd_siginfo
return 0;
}
+
+static int dispatch_sigrtmin2(sd_event_source *es, const struct signalfd_siginfo *si, void *userdata) {
+ Server *s = userdata;
+ int r;
+
+ assert(s);
+
+ if (s->storage == STORAGE_NONE)
+ return 0;
+
+ if (s->runtime_journal && !s->system_journal)
+ return 0;
+
+ log_debug("Received request to relinquish /var from PID " PID_FMT, si->ssi_pid);
+
+ (void) system_journal_open(s, false, true);
+
+ s->system_journal = journal_file_close(s->system_journal);
+ ordered_hashmap_clear_with_destructor(s->user_journals, journal_file_close);
+ set_clear_with_destructor(s->deferred_closes, journal_file_close);
+
+ if (unlink("/run/systemd/journal/flushed") < 0 && errno != ENOENT)
+ log_warning_errno(errno, "Failed to unlink /run/systemd/journal/flushed, ignoring: %m") ;
+
+ r = write_timestamp_file_atomic("/run/systemd/journal/relinquished", now(CLOCK_MONOTONIC));
+ if (r < 0)
+ log_warning_errno(r, "Failed to write /run/systemd/journal/relinquished, ignoring: %m");
+
+ return 0;
+}
+
static int setup_signals(Server *s) {
int r;
assert(s);
- assert_se(sigprocmask_many(SIG_SETMASK, NULL, SIGINT, SIGTERM, SIGUSR1, SIGUSR2, SIGRTMIN+1, -1) >= 0);
+ assert_se(sigprocmask_many(SIG_SETMASK, NULL, SIGINT, SIGTERM, SIGUSR1, SIGUSR2, SIGRTMIN+1, SIGRTMIN+2, -1) >= 0);
r = sd_event_add_signal(s->event, &s->sigusr1_event_source, SIGUSR1, dispatch_sigusr1, s);
if (r < 0)
@@ -1294,6 +1328,10 @@ static int setup_signals(Server *s) {
if (r < 0)
return r;
+ r = sd_event_add_signal(s->event, &s->sigrtmin1_event_source, SIGRTMIN+2, dispatch_sigrtmin2, s);
+ if (r < 0)
+ return r;
+
r = sd_event_source_set_priority(s->sigrtmin1_event_source, SD_EVENT_PRIORITY_NORMAL+15);
if (r < 0)
return r;
@@ -1876,7 +1914,7 @@ int server_init(Server *s) {
(void) client_context_acquire_default(s);
- return system_journal_open(s, false);
+ return system_journal_open(s, false, false);
}
void server_maybe_append_tags(Server *s) {

242
SOURCES/0874-journalctl-add-new-relinquish-and-smart-relinquish-o.patch Normal file
View File

@ -0,0 +1,242 @@
From fa93a97bdf18906d9517f4183802456986490c89 Mon Sep 17 00:00:00 2001
From: Lennart Poettering <lennart@poettering.net>
Date: Fri, 5 Apr 2019 18:21:02 +0200
Subject: [PATCH] journalctl: add new --relinquish and --smart-relinquish
options
The latter is identical to the former, but becomes a NOP if
/var/log/journal is on the same mount as /, and thus during shutdown
unmounting /var is not necessary and hence we can keep logging until the
very end.
[dtardon: The only divergence from the upstream commit is the impl. of
relinquish_var().]
(cherry picked from commit c0dfcb318c28d87e1176a8ad87ac7cc4ecc50305)
Related: #1873540
---
src/journal/journalctl.c | 161 +++++++++++++++++++++++++--------------
1 file changed, 103 insertions(+), 58 deletions(-)
diff --git a/src/journal/journalctl.c b/src/journal/journalctl.c
index 228cfe7e49..6928c79a06 100644
--- a/src/journal/journalctl.c
+++ b/src/journal/journalctl.c
@@ -47,6 +47,7 @@
#include "log.h"
#include "logs-show.h"
#include "mkdir.h"
+#include "mount-util.h"
#include "pager.h"
#include "parse-util.h"
#include "path-util.h"
@@ -162,6 +163,7 @@ static enum {
ACTION_UPDATE_CATALOG,
ACTION_LIST_BOOTS,
ACTION_FLUSH,
+ ACTION_RELINQUISH_VAR,
ACTION_SYNC,
ACTION_ROTATE,
ACTION_VACUUM,
@@ -358,6 +360,8 @@ static void help(void) {
" --vacuum-time=TIME Remove journal files older than specified time\n"
" --verify Verify journal file consistency\n"
" --sync Synchronize unwritten journal messages to disk\n"
+ " --relinquish-var Stop logging to disk, log to temporary file system\n"
+ " --smart-relinquish-var Similar, but NOP if log directory is on root mount\n"
" --flush Flush all journal data from /run into /var\n"
" --rotate Request immediate rotation of the journal files\n"
" --header Show journal header information\n"
@@ -402,6 +406,8 @@ static int parse_argv(int argc, char *argv[]) {
ARG_UTC,
ARG_SYNC,
ARG_FLUSH,
+ ARG_RELINQUISH_VAR,
+ ARG_SMART_RELINQUISH_VAR,
ARG_ROTATE,
ARG_VACUUM_SIZE,
ARG_VACUUM_FILES,
@@ -411,64 +417,66 @@ static int parse_argv(int argc, char *argv[]) {
};
static const struct option options[] = {
- { "help", no_argument, NULL, 'h' },
- { "version" , no_argument, NULL, ARG_VERSION },
- { "no-pager", no_argument, NULL, ARG_NO_PAGER },
- { "pager-end", no_argument, NULL, 'e' },
- { "follow", no_argument, NULL, 'f' },
- { "force", no_argument, NULL, ARG_FORCE },
- { "output", required_argument, NULL, 'o' },
- { "all", no_argument, NULL, 'a' },
- { "full", no_argument, NULL, 'l' },
- { "no-full", no_argument, NULL, ARG_NO_FULL },
- { "lines", optional_argument, NULL, 'n' },
- { "no-tail", no_argument, NULL, ARG_NO_TAIL },
- { "new-id128", no_argument, NULL, ARG_NEW_ID128 },
- { "quiet", no_argument, NULL, 'q' },
- { "merge", no_argument, NULL, 'm' },
- { "this-boot", no_argument, NULL, ARG_THIS_BOOT }, /* deprecated */
- { "boot", optional_argument, NULL, 'b' },
- { "list-boots", no_argument, NULL, ARG_LIST_BOOTS },
- { "dmesg", no_argument, NULL, 'k' },
- { "system", no_argument, NULL, ARG_SYSTEM },
- { "user", no_argument, NULL, ARG_USER },
- { "directory", required_argument, NULL, 'D' },
- { "file", required_argument, NULL, ARG_FILE },
- { "root", required_argument, NULL, ARG_ROOT },
- { "header", no_argument, NULL, ARG_HEADER },
- { "identifier", required_argument, NULL, 't' },
- { "priority", required_argument, NULL, 'p' },
- { "grep", required_argument, NULL, 'g' },
- { "case-sensitive", optional_argument, NULL, ARG_CASE_SENSITIVE },
- { "setup-keys", no_argument, NULL, ARG_SETUP_KEYS },
- { "interval", required_argument, NULL, ARG_INTERVAL },
- { "verify", no_argument, NULL, ARG_VERIFY },
- { "verify-key", required_argument, NULL, ARG_VERIFY_KEY },
- { "disk-usage", no_argument, NULL, ARG_DISK_USAGE },
- { "cursor", required_argument, NULL, 'c' },
- { "after-cursor", required_argument, NULL, ARG_AFTER_CURSOR },
- { "show-cursor", no_argument, NULL, ARG_SHOW_CURSOR },
- { "since", required_argument, NULL, 'S' },
- { "until", required_argument, NULL, 'U' },
- { "unit", required_argument, NULL, 'u' },
- { "user-unit", required_argument, NULL, ARG_USER_UNIT },
- { "field", required_argument, NULL, 'F' },
- { "fields", no_argument, NULL, 'N' },
- { "catalog", no_argument, NULL, 'x' },
- { "list-catalog", no_argument, NULL, ARG_LIST_CATALOG },
- { "dump-catalog", no_argument, NULL, ARG_DUMP_CATALOG },
- { "update-catalog", no_argument, NULL, ARG_UPDATE_CATALOG },
- { "reverse", no_argument, NULL, 'r' },
- { "machine", required_argument, NULL, 'M' },
- { "utc", no_argument, NULL, ARG_UTC },
- { "flush", no_argument, NULL, ARG_FLUSH },
- { "sync", no_argument, NULL, ARG_SYNC },
- { "rotate", no_argument, NULL, ARG_ROTATE },
- { "vacuum-size", required_argument, NULL, ARG_VACUUM_SIZE },
- { "vacuum-files", required_argument, NULL, ARG_VACUUM_FILES },
- { "vacuum-time", required_argument, NULL, ARG_VACUUM_TIME },
- { "no-hostname", no_argument, NULL, ARG_NO_HOSTNAME },
- { "output-fields", required_argument, NULL, ARG_OUTPUT_FIELDS },
+ { "help", no_argument, NULL, 'h' },
+ { "version" , no_argument, NULL, ARG_VERSION },
+ { "no-pager", no_argument, NULL, ARG_NO_PAGER },
+ { "pager-end", no_argument, NULL, 'e' },
+ { "follow", no_argument, NULL, 'f' },
+ { "force", no_argument, NULL, ARG_FORCE },
+ { "output", required_argument, NULL, 'o' },
+ { "all", no_argument, NULL, 'a' },
+ { "full", no_argument, NULL, 'l' },
+ { "no-full", no_argument, NULL, ARG_NO_FULL },
+ { "lines", optional_argument, NULL, 'n' },
+ { "no-tail", no_argument, NULL, ARG_NO_TAIL },
+ { "new-id128", no_argument, NULL, ARG_NEW_ID128 }, /* deprecated */
+ { "quiet", no_argument, NULL, 'q' },
+ { "merge", no_argument, NULL, 'm' },
+ { "this-boot", no_argument, NULL, ARG_THIS_BOOT }, /* deprecated */
+ { "boot", optional_argument, NULL, 'b' },
+ { "list-boots", no_argument, NULL, ARG_LIST_BOOTS },
+ { "dmesg", no_argument, NULL, 'k' },
+ { "system", no_argument, NULL, ARG_SYSTEM },
+ { "user", no_argument, NULL, ARG_USER },
+ { "directory", required_argument, NULL, 'D' },
+ { "file", required_argument, NULL, ARG_FILE },
+ { "root", required_argument, NULL, ARG_ROOT },
+ { "header", no_argument, NULL, ARG_HEADER },
+ { "identifier", required_argument, NULL, 't' },
+ { "priority", required_argument, NULL, 'p' },
+ { "grep", required_argument, NULL, 'g' },
+ { "case-sensitive", optional_argument, NULL, ARG_CASE_SENSITIVE },
+ { "setup-keys", no_argument, NULL, ARG_SETUP_KEYS },
+ { "interval", required_argument, NULL, ARG_INTERVAL },
+ { "verify", no_argument, NULL, ARG_VERIFY },
+ { "verify-key", required_argument, NULL, ARG_VERIFY_KEY },
+ { "disk-usage", no_argument, NULL, ARG_DISK_USAGE },
+ { "cursor", required_argument, NULL, 'c' },
+ { "after-cursor", required_argument, NULL, ARG_AFTER_CURSOR },
+ { "show-cursor", no_argument, NULL, ARG_SHOW_CURSOR },
+ { "since", required_argument, NULL, 'S' },
+ { "until", required_argument, NULL, 'U' },
+ { "unit", required_argument, NULL, 'u' },
+ { "user-unit", required_argument, NULL, ARG_USER_UNIT },
+ { "field", required_argument, NULL, 'F' },
+ { "fields", no_argument, NULL, 'N' },
+ { "catalog", no_argument, NULL, 'x' },
+ { "list-catalog", no_argument, NULL, ARG_LIST_CATALOG },
+ { "dump-catalog", no_argument, NULL, ARG_DUMP_CATALOG },
+ { "update-catalog", no_argument, NULL, ARG_UPDATE_CATALOG },
+ { "reverse", no_argument, NULL, 'r' },
+ { "machine", required_argument, NULL, 'M' },
+ { "utc", no_argument, NULL, ARG_UTC },
+ { "flush", no_argument, NULL, ARG_FLUSH },
+ { "relinquish-var", no_argument, NULL, ARG_RELINQUISH_VAR },
+ { "smart-relinquish-var", no_argument, NULL, ARG_SMART_RELINQUISH_VAR },
+ { "sync", no_argument, NULL, ARG_SYNC },
+ { "rotate", no_argument, NULL, ARG_ROTATE },
+ { "vacuum-size", required_argument, NULL, ARG_VACUUM_SIZE },
+ { "vacuum-files", required_argument, NULL, ARG_VACUUM_FILES },
+ { "vacuum-time", required_argument, NULL, ARG_VACUUM_TIME },
+ { "no-hostname", no_argument, NULL, ARG_NO_HOSTNAME },
+ { "output-fields", required_argument, NULL, ARG_OUTPUT_FIELDS },
{}
};
@@ -893,6 +901,35 @@ static int parse_argv(int argc, char *argv[]) {
arg_action = ACTION_FLUSH;
break;
+ case ARG_SMART_RELINQUISH_VAR: {
+ int root_mnt_id, log_mnt_id;
+
+ /* Try to be smart about relinquishing access to /var/log/journal/ during shutdown:
+ * if it's on the same mount as the root file system there's no point in
+ * relinquishing access and we can leave journald write to it until the very last
+ * moment. */
+
+ r = path_get_mnt_id("/", &root_mnt_id);
+ if (r < 0)
+ log_debug_errno(r, "Failed to get root mount ID, ignoring: %m");
+ else {
+ r = path_get_mnt_id("/var/log/journal/", &log_mnt_id);
+ if (r < 0)
+ log_debug_errno(r, "Failed to get journal directory mount ID, ignoring: %m");
+ else if (root_mnt_id == log_mnt_id) {
+ log_debug("/var/log/journal/ is on root file system, not relinquishing access to /var.");
+ return 0;
+ } else
+ log_debug("/var/log/journal/ is not on the root file system, relinquishing access to it.");
+ }
+
+ _fallthrough_;
+ }
+
+ case ARG_RELINQUISH_VAR:
+ arg_action = ACTION_RELINQUISH_VAR;
+ break;
+
case ARG_ROTATE:
arg_action = ACTION_ROTATE;
break;
@@ -2056,6 +2093,10 @@ static int send_signal_and_wait(int sig, const char *watch_path) {
return 0;
}
+static int relinquish_var(void) {
+ return send_signal_and_wait(SIGRTMIN+2, "/run/systemd/journal/relinquished");
+}
+
static int rotate(void) {
return send_signal_and_wait(SIGUSR2, "/run/systemd/journal/rotated");
}
@@ -2171,6 +2212,10 @@ int main(int argc, char *argv[]) {
r = flush_to_var();
goto finish;
+ case ACTION_RELINQUISH_VAR:
+ r = relinquish_var();
+ goto finish;
+
case ACTION_SYNC:
r = sync_journal();
goto finish;

26
SOURCES/0875-units-automatically-revert-to-run-logging-on-shutdow.patch Normal file
View File

@ -0,0 +1,26 @@
From ce6531045b337c3f793d1d74f1e5641e658805bb Mon Sep 17 00:00:00 2001
From: Lennart Poettering <lennart@poettering.net>
Date: Fri, 5 Apr 2019 18:22:31 +0200
Subject: [PATCH] units: automatically revert to /run logging on shutdown if
necessary
Fixes: #867
(cherry picked from commit 1e187d2dd52cbb4f0bb30e4d96acf7f72a145b91)
Resolves: #1873540
---
units/systemd-journal-flush.service.in | 1 +
1 file changed, 1 insertion(+)
diff --git a/units/systemd-journal-flush.service.in b/units/systemd-journal-flush.service.in
index 439f5f3f76..653dcebe0e 100644
--- a/units/systemd-journal-flush.service.in
+++ b/units/systemd-journal-flush.service.in
@@ -19,6 +19,7 @@ RequiresMountsFor=/var/log/journal
[Service]
ExecStart=@rootbindir@/journalctl --flush
+ExecStop=@rootbindir@/journalctl --smart-relinquish-var
Type=oneshot
RemainAfterExit=yes
TimeoutSec=90s

929
SOURCES/0876-pstore-Tool-to-archive-contents-of-pstore.patch Normal file
View File

@ -0,0 +1,929 @@
From 2f75df5cd6dcd56775fec9e89fc79672e702d826 Mon Sep 17 00:00:00 2001
From: Eric DeVolder <eric.devolder@oracle.com>
Date: Thu, 16 May 2019 08:59:01 -0500
Subject: [PATCH] pstore: Tool to archive contents of pstore
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
This patch introduces the systemd pstore service which will archive the
contents of the Linux persistent storage filesystem, pstore, to other storage,
thus preserving the existing information contained in the pstore, and clearing
pstore storage for future error events.
Linux provides a persistent storage file system, pstore[1], that can store
error records when the kernel dies (or reboots or powers-off). These records in
turn can be referenced to debug kernel problems (currently the kernel stuffs
the tail of the dmesg, which also contains a stack backtrace, into pstore).
The pstore file system supports a variety of backends that map onto persistent
storage, such as the ACPI ERST[2, Section 18.5 Error Serialization] and UEFI
variables[3 Appendix N Common Platform Error Record]. The pstore backends
typically offer a relatively small amount of persistent storage, e.g. 64KiB,
which can quickly fill up and thus prevent subsequent kernel crashes from
recording errors. Thus there is a need to monitor and extract the pstore
contents so that future kernel problems can also record information in the
pstore.
The pstore service is independent of the kdump service. In cloud environments
specifically, host and guest filesystems are on remote filesystems (eg. iSCSI
or NFS), thus kdump relies [implicitly and/or explicitly] upon proper operation
of networking software *and* hardware *and* infrastructure. Thus it may not be
possible to capture a kernel coredump to a file since writes over the network
may not be possible.
The pstore backend, on the other hand, is completely local and provides a path
to store error records which will survive a reboot and aid in post-mortem
debugging.
Usage Notes:
This tool moves files from /sys/fs/pstore into /var/lib/systemd/pstore.
To enable kernel recording of error records into pstore, one must either pass
crash_kexec_post_notifiers[4] to the kernel command line or enable via 'echo Y
> /sys/module/kernel/parameters/crash_kexec_post_notifiers'. This option
invokes the recording of errors into pstore *before* an attempt to kexec/kdump
on a kernel crash.
Optionally, to record reboots and shutdowns in the pstore, one can either pass
the printk.always_kmsg_dump[4] to the kernel command line or enable via 'echo Y >
/sys/module/printk/parameters/always_kmsg_dump'. This option enables code on the
shutdown path to record information via pstore.
This pstore service is a oneshot service. When run, the service invokes
systemd-pstore which is a tool that performs the following:
- reads the pstore.conf configuration file
- collects the lists of files in the pstore (eg. /sys/fs/pstore)
- for certain file types (eg. dmesg) a handler is invoked
- for all other files, the file is moved from pstore
- In the case of dmesg handler, final processing occurs as such:
- files processed in reverse lexigraphical order to faciliate
reconstruction of original dmesg
- the filename is examined to determine which dmesg it is a part
- the file is appended to the reconstructed dmesg
For example, the following pstore contents:
root@vm356:~# ls -al /sys/fs/pstore
total 0
drwxr-x--- 2 root root 0 May 9 09:50 .
drwxr-xr-x 7 root root 0 May 9 09:50 ..
-r--r--r-- 1 root root 1610 May 9 09:49 dmesg-efi-155741337601001
-r--r--r-- 1 root root 1778 May 9 09:49 dmesg-efi-155741337602001
-r--r--r-- 1 root root 1726 May 9 09:49 dmesg-efi-155741337603001
-r--r--r-- 1 root root 1746 May 9 09:49 dmesg-efi-155741337604001
-r--r--r-- 1 root root 1686 May 9 09:49 dmesg-efi-155741337605001
-r--r--r-- 1 root root 1690 May 9 09:49 dmesg-efi-155741337606001
-r--r--r-- 1 root root 1775 May 9 09:49 dmesg-efi-155741337607001
-r--r--r-- 1 root root 1811 May 9 09:49 dmesg-efi-155741337608001
-r--r--r-- 1 root root 1817 May 9 09:49 dmesg-efi-155741337609001
-r--r--r-- 1 root root 1795 May 9 09:49 dmesg-efi-155741337710001
-r--r--r-- 1 root root 1770 May 9 09:49 dmesg-efi-155741337711001
-r--r--r-- 1 root root 1796 May 9 09:49 dmesg-efi-155741337712001
-r--r--r-- 1 root root 1787 May 9 09:49 dmesg-efi-155741337713001
-r--r--r-- 1 root root 1808 May 9 09:49 dmesg-efi-155741337714001
-r--r--r-- 1 root root 1754 May 9 09:49 dmesg-efi-155741337715001
results in the following:
root@vm356:~# ls -al /var/lib/systemd/pstore/155741337/
total 92
drwxr-xr-x 2 root root 4096 May 9 09:50 .
drwxr-xr-x 4 root root 40 May 9 09:50 ..
-rw-r--r-- 1 root root 1610 May 9 09:50 dmesg-efi-155741337601001
-rw-r--r-- 1 root root 1778 May 9 09:50 dmesg-efi-155741337602001
-rw-r--r-- 1 root root 1726 May 9 09:50 dmesg-efi-155741337603001
-rw-r--r-- 1 root root 1746 May 9 09:50 dmesg-efi-155741337604001
-rw-r--r-- 1 root root 1686 May 9 09:50 dmesg-efi-155741337605001
-rw-r--r-- 1 root root 1690 May 9 09:50 dmesg-efi-155741337606001
-rw-r--r-- 1 root root 1775 May 9 09:50 dmesg-efi-155741337607001
-rw-r--r-- 1 root root 1811 May 9 09:50 dmesg-efi-155741337608001
-rw-r--r-- 1 root root 1817 May 9 09:50 dmesg-efi-155741337609001
-rw-r--r-- 1 root root 1795 May 9 09:50 dmesg-efi-155741337710001
-rw-r--r-- 1 root root 1770 May 9 09:50 dmesg-efi-155741337711001
-rw-r--r-- 1 root root 1796 May 9 09:50 dmesg-efi-155741337712001
-rw-r--r-- 1 root root 1787 May 9 09:50 dmesg-efi-155741337713001
-rw-r--r-- 1 root root 1808 May 9 09:50 dmesg-efi-155741337714001
-rw-r--r-- 1 root root 1754 May 9 09:50 dmesg-efi-155741337715001
-rw-r--r-- 1 root root 26754 May 9 09:50 dmesg.txt
where dmesg.txt is reconstructed from the group of related
dmesg-efi-155741337* files.
Configuration file:
The pstore.conf configuration file has four settings, described below.
- Storage : one of "none", "external", or "journal". With "none", this
tool leaves the contents of pstore untouched. With "external", the
contents of the pstore are moved into the /var/lib/systemd/pstore,
as well as logged into the journal. With "journal", the contents of
the pstore are recorded only in the systemd journal. The default is
"external".
- Unlink : is a boolean. When "true", the default, then files in the
pstore are removed once processed. When "false", processing of the
pstore occurs normally, but the pstore files remain.
References:
[1] "Persistent storage for a kernel's dying breath",
March 23, 2011.
https://lwn.net/Articles/434821/
[2] "Advanced Configuration and Power Interface Specification",
version 6.2, May 2017.
https://www.uefi.org/sites/default/files/resources/ACPI_6_2.pdf
[3] "Unified Extensible Firmware Interface Specification",
version 2.8, March 2019.
https://uefi.org/sites/default/files/resources/UEFI_Spec_2_8_final.pdf
[4] "The kernel’s command-line parameters",
https://static.lwn.net/kerneldoc/admin-guide/kernel-parameters.html
(cherry picked from commit 9b4abc69b201e5d7295e1b0762883659f053e747)
Resolves: #2158832
---
man/pstore.conf.xml | 89 +++++++
man/rules/meson.build | 2 +
man/systemd-pstore.xml | 99 ++++++++
meson.build | 20 ++
meson_options.txt | 2 +
src/pstore/meson.build | 10 +
src/pstore/pstore.c | 395 ++++++++++++++++++++++++++++++++
src/pstore/pstore.conf | 16 ++
units/meson.build | 1 +
units/systemd-pstore.service.in | 24 ++
10 files changed, 658 insertions(+)
create mode 100644 man/pstore.conf.xml
create mode 100644 man/systemd-pstore.xml
create mode 100644 src/pstore/meson.build
create mode 100644 src/pstore/pstore.c
create mode 100644 src/pstore/pstore.conf
create mode 100644 units/systemd-pstore.service.in
diff --git a/man/pstore.conf.xml b/man/pstore.conf.xml
new file mode 100644
index 0000000000..b5cda47d02
--- /dev/null
+++ b/man/pstore.conf.xml
@@ -0,0 +1,89 @@
+<?xml version='1.0'?>
+<!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook XML V4.5//EN"
+ "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd">
+<!-- SPDX-License-Identifier: LGPL-2.1+ -->
+
+<refentry id="pstore.conf" conditional="ENABLE_PSTORE"
+ xmlns:xi="http://www.w3.org/2001/XInclude">
+ <refentryinfo>
+ <title>pstore.conf</title>
+ <productname>systemd</productname>
+ </refentryinfo>
+
+ <refmeta>
+ <refentrytitle>pstore.conf</refentrytitle>
+ <manvolnum>5</manvolnum>
+ </refmeta>
+
+ <refnamediv>
+ <refname>pstore.conf</refname>
+ <refname>pstore.conf.d</refname>
+ <refpurpose>PStore configuration file</refpurpose>
+ </refnamediv>
+
+ <refsynopsisdiv>
+ <para>
+ <filename>/etc/systemd/pstore.conf</filename>
+ <filename>/etc/systemd/pstore.conf.d/*</filename>
+ </para>
+ </refsynopsisdiv>
+
+ <refsect1>
+ <title>Description</title>
+
+ <para>This file configures the behavior of
+ <citerefentry><refentrytitle>systemd-pstore</refentrytitle><manvolnum>8</manvolnum></citerefentry>,
+ a tool for archiving the contents of the persistent storage filesystem,
+ <ulink url="https://www.kernel.org/doc/Documentation/ABI/testing/pstore">pstore</ulink>.
+ </para>
+ </refsect1>
+
+ <xi:include href="standard-conf.xml" xpointer="main-conf" />
+
+ <refsect1>
+ <title>Options</title>
+
+ <para>All options are configured in the
+ <literal>[PStore]</literal> section:</para>
+
+ <variablelist>
+
+ <varlistentry>
+ <term><varname>Storage=</varname></term>
+
+ <listitem><para>Controls where to archive (i.e. copy) files from the pstore filesystem. One of <literal>none</literal>,
+ <literal>external</literal>, and <literal>journal</literal>. When
+ <literal>none</literal>, the tool exits without processing files in the pstore filesystem.
+ When <literal>external</literal> (the default), files are archived into <filename>/var/lib/systemd/pstore/</filename>,
+ and logged into the journal.
+ When <literal>journal</literal>, pstore file contents are logged only in the journal.</para>
+ </listitem>
+
+ </varlistentry>
+
+ <varlistentry>
+ <term><varname>Unlink=</varname></term>
+
+ <listitem><para>Controls whether or not files are removed from pstore after processing.
+ Takes a boolean value. When true, a pstore file is removed from the pstore once it has been
+ archived (either to disk or into the journal). When false, processing of pstore files occurs
+ normally, but the files remain in the pstore.
+ The default is true in order to maintain the pstore in a nearly empty state, so that the pstore
+ has storage available for the next kernel error event.
+ </para></listitem>
+ </varlistentry>
+ </variablelist>
+
+ <para>The defaults for all values are listed as comments in the
+ template <filename>/etc/systemd/pstore.conf</filename> file that
+ is installed by default.</para>
+ </refsect1>
+
+ <refsect1>
+ <title>See Also</title>
+ <para>
+ <citerefentry><refentrytitle>systemd-journald.service</refentrytitle><manvolnum>8</manvolnum></citerefentry>,
+ </para>
+ </refsect1>
+
+</refentry>
diff --git a/man/rules/meson.build b/man/rules/meson.build
index e6c0a99bbd..6295330c5e 100644
--- a/man/rules/meson.build
+++ b/man/rules/meson.build
@@ -44,6 +44,7 @@ manpages = [
['os-release', '5', [], ''],
['pam_systemd', '8', [], 'HAVE_PAM'],
['portablectl', '1', [], 'ENABLE_PORTABLED'],
+ ['pstore.conf', '5', ['pstore.conf.d'], 'ENABLE_PSTORE'],
['resolvectl', '1', ['resolvconf'], 'ENABLE_RESOLVE'],
['resolved.conf', '5', ['resolved.conf.d'], 'ENABLE_RESOLVE'],
['runlevel', '8', [], 'ENABLE_UTMP'],
@@ -633,6 +634,7 @@ manpages = [
['systemd-nspawn', '1', [], ''],
['systemd-path', '1', [], ''],
['systemd-portabled.service', '8', ['systemd-portabled'], 'ENABLE_PORTABLED'],
+ ['systemd-pstore', '8', ['systemd-pstore.service'], 'ENABLE_PSTORE'],
['systemd-quotacheck.service',
'8',
['systemd-quotacheck'],
diff --git a/man/systemd-pstore.xml b/man/systemd-pstore.xml
new file mode 100644
index 0000000000..dd1aa5e83b
--- /dev/null
+++ b/man/systemd-pstore.xml
@@ -0,0 +1,99 @@
+<?xml version='1.0'?>
+<!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook XML V4.5//EN"
+ "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd">
+<!-- SPDX-License-Identifier: LGPL-2.1+ -->
+
+<refentry id="systemd-pstore" conditional='ENABLE_PSTORE'
+ xmlns:xi="http://www.w3.org/2001/XInclude">
+
+ <refentryinfo>
+ <title>systemd-pstore</title>
+ <productname>systemd</productname>
+ </refentryinfo>
+
+ <refmeta>
+ <refentrytitle>systemd-pstore</refentrytitle>
+ <manvolnum>8</manvolnum>
+ </refmeta>
+
+ <refnamediv>
+ <refname>systemd-pstore</refname>
+ <refname>systemd-pstore.service</refname>
+ <refpurpose>Tool to archive contents of the persistent storage filesytem</refpurpose>
+ </refnamediv>
+
+ <refsynopsisdiv>
+ <para><filename>/usr/lib/systemd/systemd-pstore</filename></para>
+ <para><filename>systemd-pstore.service</filename></para>
+ </refsynopsisdiv>
+
+ <refsect1>
+ <title>Description</title>
+ <para><filename>systemd-pstore.service</filename> is a system service that archives the
+ contents of the Linux persistent storage filesystem, pstore, to other storage,
+ thus preserving the existing information contained in the pstore, and clearing
+ pstore storage for future error events.</para>
+
+ <para>Linux provides a persistent storage file system, pstore, that can store
+ error records when the kernel dies (or reboots or powers-off). These records in
+ turn can be referenced to debug kernel problems (currently the kernel stuffs
+ the tail of the dmesg, which also contains a stack backtrace, into pstore).</para>
+
+ <para>The pstore file system supports a variety of backends that map onto persistent
+ storage, such as the ACPI ERST and UEFI variables. The pstore backends
+ typically offer a relatively small amount of persistent storage, e.g. 64KiB,
+ which can quickly fill up and thus prevent subsequent kernel crashes from
+ recording errors. Thus there is a need to monitor and extract the pstore
+ contents so that future kernel problems can also record information in the
+ pstore.</para>
+
+ <para>The pstore service is independent of the kdump service. In cloud environments
+ specifically, host and guest filesystems are on remote filesystems (eg. iSCSI
+ or NFS), thus kdump relies [implicitly and/or explicitly] upon proper operation
+ of networking software *and* hardware *and* infrastructure. Thus it may not be
+ possible to capture a kernel coredump to a file since writes over the network
+ may not be possible.</para>
+
+ <para>The pstore backend, on the other hand, is completely local and provides a path
+ to store error records which will survive a reboot and aid in post-mortem
+ debugging.</para>
+
+ <para>The <command>systemd-pstore</command> executable does the actual work. Upon starting,
+ the <filename>pstore.conf</filename> is read to obtain options, then the /sys/fs/pstore
+ directory contents are processed according to the options. Pstore files are written to the
+ journal, and optionally saved into /var/lib/systemd/pstore.</para>
+ </refsect1>
+
+ <refsect1>
+ <title>Configuration</title>
+
+ <para>The behavior of <command>systemd-pstore</command> is configured through the configuration file
+ <filename>/etc/systemd/pstore.conf</filename> and corresponding snippets
+ <filename>/etc/systemd/pstore.conf.d/*.conf</filename>, see
+ <citerefentry><refentrytitle>pstore.conf</refentrytitle><manvolnum>5</manvolnum></citerefentry>.
+ </para>
+
+ <refsect2>
+ <title>Disabling pstore processing</title>
+
+ <para>To disable pstore processing by <command>systemd-pstore</command>,
+ set <programlisting>Storage=none</programlisting> in
+ <citerefentry><refentrytitle>pstore.conf</refentrytitle><manvolnum>5</manvolnum></citerefentry>.
+ </para>
+ </refsect2>
+ </refsect1>
+
+ <refsect1>
+ <title>Usage</title>
+ <para>Data stored in the journal can be viewed with
+ <citerefentry><refentrytitle>journalctl</refentrytitle><manvolnum>1</manvolnum></citerefentry>
+ as usual.</para>
+ </refsect1>
+
+ <refsect1>
+ <title>See Also</title>
+ <para>
+ <citerefentry><refentrytitle>pstore.conf</refentrytitle><manvolnum>5</manvolnum></citerefentry>
+ </para>
+ </refsect1>
+</refentry>
diff --git a/meson.build b/meson.build
index af4cf331da..972a8fb6f7 100644
--- a/meson.build
+++ b/meson.build
@@ -1224,6 +1224,7 @@ foreach term : ['utmp',
'environment-d',
'binfmt',
'coredump',
+ 'pstore',
'resolve',
'logind',
'hostnamed',
@@ -1439,6 +1440,7 @@ subdir('src/network')
subdir('src/analyze')
subdir('src/journal-remote')
subdir('src/coredump')
+subdir('src/pstore')
subdir('src/hostname')
subdir('src/import')
subdir('src/kernel-install')
@@ -2151,6 +2153,23 @@ if conf.get('ENABLE_COREDUMP') == 1
public_programs += [exe]
endif
+if conf.get('ENABLE_PSTORE') == 1
+ executable('systemd-pstore',
+ systemd_pstore_sources,
+ include_directories : includes,
+ link_with : [libshared],
+ dependencies : [threads,
+ libacl,
+ libdw,
+ libxz,
+ liblz4],
+ install_rpath : rootlibexecdir,
+ install : true,
+ install_dir : rootlibexecdir)
+
+ public_programs += exe
+endif
+
if conf.get('ENABLE_BINFMT') == 1
exe = executable('systemd-binfmt',
'src/binfmt/binfmt.c',
@@ -3014,6 +3033,7 @@ foreach tuple : [
['resolve'],
['DNS-over-TLS'],
['coredump'],
+ ['pstore'],
['polkit'],
['legacy pkla', install_polkit_pkla],
['efi'],
diff --git a/meson_options.txt b/meson_options.txt
index 213079ac15..5624304bf4 100644
--- a/meson_options.txt
+++ b/meson_options.txt
@@ -76,6 +76,8 @@ option('binfmt', type : 'boolean',
description : 'support for custom binary formats')
option('coredump', type : 'boolean',
description : 'install the coredump handler')
+option('pstore', type : 'boolean',
+ description : 'install the pstore archival tool')
option('logind', type : 'boolean',
description : 'install the systemd-logind stack')
option('hostnamed', type : 'boolean',
diff --git a/src/pstore/meson.build b/src/pstore/meson.build
new file mode 100644
index 0000000000..adbac24b54
--- /dev/null
+++ b/src/pstore/meson.build
@@ -0,0 +1,10 @@
+# SPDX-License-Identifier: LGPL-2.1+
+
+systemd_pstore_sources = files('''
+ pstore.c
+'''.split())
+
+if conf.get('ENABLE_PSTORE') == 1
+ install_data('pstore.conf',
+ install_dir : pkgsysconfdir)
+endif
diff --git a/src/pstore/pstore.c b/src/pstore/pstore.c
new file mode 100644
index 0000000000..f95e016eb6
--- /dev/null
+++ b/src/pstore/pstore.c
@@ -0,0 +1,395 @@
+/* SPDX-License-Identifier: LGPL-2.1+ */
+
+/* Copyright © 2019 Oracle and/or its affiliates. */
+
+/* Generally speaking, the pstore contains a small number of files
+ * that in turn contain a small amount of data. */
+#include <errno.h>
+#include <stdio.h>
+#include <stdio_ext.h>
+#include <sys/prctl.h>
+#include <sys/xattr.h>
+#include <unistd.h>
+
+#include "sd-daemon.h"
+#include "sd-journal.h"
+#include "sd-login.h"
+#include "sd-messages.h"
+
+#include "acl-util.h"
+#include "alloc-util.h"
+#include "capability-util.h"
+#include "cgroup-util.h"
+#include "compress.h"
+#include "conf-parser.h"
+#include "copy.h"
+#include "dirent-util.h"
+#include "escape.h"
+#include "fd-util.h"
+#include "fileio.h"
+#include "fs-util.h"
+#include "io-util.h"
+#include "journal-importer.h"
+#include "log.h"
+#include "macro.h"
+#include "missing.h"
+#include "mkdir.h"
+#include "parse-util.h"
+#include "process-util.h"
+#include "signal-util.h"
+#include "socket-util.h"
+#include "special.h"
+#include "string-table.h"
+#include "string-util.h"
+#include "strv.h"
+#include "user-util.h"
+#include "util.h"
+
+/* Command line argument handling */
+typedef enum PStoreStorage {
+ PSTORE_STORAGE_NONE,
+ PSTORE_STORAGE_EXTERNAL,
+ PSTORE_STORAGE_JOURNAL,
+ _PSTORE_STORAGE_MAX,
+ _PSTORE_STORAGE_INVALID = -1
+} PStoreStorage;
+
+static const char* const pstore_storage_table[_PSTORE_STORAGE_MAX] = {
+ [PSTORE_STORAGE_NONE] = "none",
+ [PSTORE_STORAGE_EXTERNAL] = "external",
+ [PSTORE_STORAGE_JOURNAL] = "journal",
+};
+
+DEFINE_PRIVATE_STRING_TABLE_LOOKUP(pstore_storage, PStoreStorage);
+static DEFINE_CONFIG_PARSE_ENUM(config_parse_pstore_storage, pstore_storage, PStoreStorage, "Failed to parse storage setting");
+
+static PStoreStorage arg_storage = PSTORE_STORAGE_EXTERNAL;
+
+static bool arg_unlink = true;
+static const char *arg_sourcedir = "/sys/fs/pstore";
+static const char *arg_archivedir = "/var/lib/systemd/pstore";
+
+static int parse_config(void) {
+ static const ConfigTableItem items[] = {
+ { "PStore", "Unlink", config_parse_bool, 0, &arg_unlink },
+ { "PStore", "Storage", config_parse_pstore_storage, 0, &arg_storage },
+ {}
+ };
+
+ return config_parse_many_nulstr(PKGSYSCONFDIR "/pstore.conf",
+ CONF_PATHS_NULSTR("systemd/pstore.conf.d"),
+ "PStore\0",
+ config_item_table_lookup, items,
+ CONFIG_PARSE_WARN, NULL);
+}
+
+/* File list handling - PStoreEntry is the struct and
+ * and PStoreEntry is the type that contains all info
+ * about a pstore entry. */
+typedef struct PStoreEntry {
+ struct dirent dirent;
+ bool is_binary;
+ bool handled;
+ char *content;
+ size_t content_size;
+} PStoreEntry;
+
+typedef struct PStoreList {
+ PStoreEntry *entries;
+ size_t n_entries;
+ size_t n_entries_allocated;
+} PStoreList;
+
+static void pstore_entries_reset(PStoreList *list) {
+ for (size_t i = 0; i < list->n_entries; i++)
+ free(list->entries[i].content);
+ free(list->entries);
+ list->n_entries = 0;
+}
+
+static int compare_pstore_entries(const void *_a, const void *_b) {
+ PStoreEntry *a = (PStoreEntry *)_a, *b = (PStoreEntry *)_b;
+ return strcmp(a->dirent.d_name, b->dirent.d_name);
+}
+
+static int move_file(PStoreEntry *pe, const char *subdir) {
+ _cleanup_free_ char *ifd_path = NULL;
+ _cleanup_free_ char *ofd_path = NULL;
+ int r = 0;
+ struct iovec iovec[2] = {};
+ int n_iovec = 0;
+ _cleanup_free_ void *field = NULL;
+ const char *suffix = NULL;
+ size_t field_size;
+
+ if (pe->handled)
+ return 0;
+
+ ifd_path = path_join(NULL, arg_sourcedir, pe->dirent.d_name);
+ if (!ifd_path)
+ return log_oom();
+
+ ofd_path = path_join(arg_archivedir, subdir, pe->dirent.d_name);
+ if (!ofd_path)
+ return log_oom();
+
+ /* Always log to the journal */
+ suffix = arg_storage == PSTORE_STORAGE_EXTERNAL ? strjoina(" moved to ", ofd_path) : (char *)".";
+ field = strjoina("MESSAGE=PStore ", pe->dirent.d_name, suffix);
+ iovec[n_iovec++] = IOVEC_MAKE_STRING(field);
+
+ field_size = strlen("FILE=") + pe->content_size;
+ field = malloc(field_size);
+ if (!field)
+ return log_oom();
+ memcpy(stpcpy(field, "FILE="), pe->content, pe->content_size);
+ iovec[n_iovec++] = IOVEC_MAKE(field, field_size);
+
+ r = sd_journal_sendv(iovec, n_iovec);
+ if (r < 0)
+ return log_error_errno(r, "Failed to log pstore entry: %m");
+
+ if (arg_storage == PSTORE_STORAGE_EXTERNAL) {
+ /* Move file from pstore to external storage */
+ r = mkdir_parents(ofd_path, 0755);
+ if (r < 0)
+ return log_error_errno(r, "Failed to create directoy %s: %m", ofd_path);
+ r = copy_file_atomic(ifd_path, ofd_path, 0600, 0, COPY_REPLACE);
+ if (r < 0)
+ return log_error_errno(r, "Failed to copy_file_atomic: %s to %s", ifd_path, ofd_path);
+ }
+
+ /* If file copied properly, remove it from pstore */
+ if (arg_unlink)
+ (void) unlink(ifd_path);
+
+ pe->handled = true;
+
+ return 0;
+}
+
+static int write_dmesg(const char *dmesg, size_t size, const char *id) {
+ _cleanup_(unlink_and_freep) char *ofd_path = NULL;
+ _cleanup_free_ char *tmp_path = NULL;
+ _cleanup_close_ int ofd = -1;
+ ssize_t wr;
+ int r;
+
+ if (isempty(dmesg) || size == 0)
+ return 0;
+
+ /* log_info("Record ID %s", id); */
+
+ ofd_path = path_join(arg_archivedir, id, "dmesg.txt");
+ if (!ofd_path)
+ return log_oom();
+
+ ofd = open_tmpfile_linkable(ofd_path, O_CLOEXEC|O_CREAT|O_TRUNC|O_WRONLY, &tmp_path);
+ if (ofd < 0)
+ return log_error_errno(ofd, "Failed to open temporary file %s: %m", ofd_path);
+ wr = write(ofd, dmesg, size);
+ if (wr < 0)
+ return log_error_errno(errno, "Failed to store dmesg to %s: %m", ofd_path);
+ if (wr != (ssize_t)size)
+ return log_error_errno(-EIO, "Failed to store dmesg to %s. %zu bytes are lost.", ofd_path, size - wr);
+ r = link_tmpfile(ofd, tmp_path, ofd_path);
+ if (r < 0)
+ return log_error_errno(r, "Failed to write temporary file %s: %m", ofd_path);
+ ofd_path = mfree(ofd_path);
+
+ return 0;
+}
+
+static void process_dmesg_files(PStoreList *list) {
+ /* Move files, reconstruct dmesg.txt */
+ PStoreEntry *pe;
+ _cleanup_free_ char *dmesg = NULL;
+ size_t dmesg_size = 0;
+ _cleanup_free_ char *dmesg_id = NULL;
+
+ /* Handle each dmesg file: files processed in reverse
+ * order so as to properly reconstruct original dmesg */
+ for (size_t n = list->n_entries; n > 0; n--) {
+ bool move_file_and_continue = false;
+ _cleanup_free_ char *pe_id = NULL;
+ char *p;
+ size_t plen;
+
+ pe = &list->entries[n-1];
+
+ if (pe->handled)
+ continue;
+ if (!startswith(pe->dirent.d_name, "dmesg-"))
+ continue;
+
+ if (endswith(pe->dirent.d_name, ".enc.z")) /* indicates a problem */
+ move_file_and_continue = true;
+ p = strrchr(pe->dirent.d_name, '-');
+ if (!p)
+ move_file_and_continue = true;
+
+ if (move_file_and_continue) {
+ /* A dmesg file on which we do NO additional processing */
+ (void) move_file(pe, NULL);
+ continue;
+ }
+
+ /* See if this file is one of a related group of files
+ * in order to reconstruct dmesg */
+
+ /* When dmesg is written into pstore, it is done so in
+ * small chunks, whatever the exchange buffer size is
+ * with the underlying pstore backend (ie. EFI may be
+ * ~2KiB), which means an example pstore with approximately
+ * 64KB of storage may have up to roughly 32 dmesg files
+ * that could be related, depending upon the size of the
+ * original dmesg.
+ *
+ * Here we look at the dmesg filename and try to discern
+ * if files are part of a related group, meaning the same
+ * original dmesg.
+ *
+ * The two known pstore backends are EFI and ERST. These
+ * backends store data in the Common Platform Error
+ * Record, CPER, format. The dmesg- filename contains the
+ * CPER record id, a 64bit number (in decimal notation).
+ * In Linux, the record id is encoded with two digits for
+ * the dmesg part (chunk) number and 3 digits for the
+ * count number. So allowing an additional digit to
+ * compensate for advancing time, this code ignores the
+ * last six digits of the filename in determining the
+ * record id.
+ *
+ * For the EFI backend, the record id encodes an id in the
+ * upper 32 bits, and a timestamp in the lower 32-bits.
+ * So ignoring the least significant 6 digits has proven
+ * to generally identify related dmesg entries. */
+#define PSTORE_FILENAME_IGNORE 6
+
+ /* determine common portion of record id */
+ ++p; /* move beyond dmesg- */
+ plen = strlen(p);
+ if (plen > PSTORE_FILENAME_IGNORE) {
+ pe_id = memdup_suffix0(p, plen - PSTORE_FILENAME_IGNORE);
+ if (!pe_id) {
+ log_oom();
+ return;
+ }
+ } else
+ pe_id = mfree(pe_id);
+
+ /* Now move file from pstore to archive storage */
+ move_file(pe, pe_id);
+
+ /* If the current record id is NOT the same as the
+ * previous record id, then start a new dmesg.txt file */
+ if (!pe_id || !dmesg_id || !streq(pe_id, dmesg_id)) {
+ /* Encountered a new dmesg group, close out old one, open new one */
+ if (dmesg) {
+ (void) write_dmesg(dmesg, dmesg_size, dmesg_id);
+ dmesg = mfree(dmesg);
+ dmesg_size = 0;
+ }
+
+ /* now point dmesg_id to storage of pe_id */
+ free_and_replace(dmesg_id, pe_id);
+ }
+
+ /* Reconstruction of dmesg is done as a useful courtesy, do not log errors */
+ dmesg = realloc(dmesg, dmesg_size + strlen(pe->dirent.d_name) + strlen(":\n") + pe->content_size + 1);
+ if (dmesg) {
+ dmesg_size += sprintf(&dmesg[dmesg_size], "%s:\n", pe->dirent.d_name);
+ if (pe->content) {
+ memcpy(&dmesg[dmesg_size], pe->content, pe->content_size);
+ dmesg_size += pe->content_size;
+ }
+ }
+
+ pe_id = mfree(pe_id);
+ }
+ if (dmesg)
+ (void) write_dmesg(dmesg, dmesg_size, dmesg_id);
+}
+
+static int list_files(PStoreList *list, const char *sourcepath) {
+ _cleanup_(closedirp) DIR *dirp = NULL;
+ struct dirent *de;
+ int r = 0;
+
+ dirp = opendir(sourcepath);
+ if (!dirp)
+ return log_error_errno(errno, "Failed to opendir %s: %m", sourcepath);
+
+ FOREACH_DIRENT(de, dirp, return log_error_errno(errno, "Failed to iterate through %s: %m", sourcepath)) {
+ _cleanup_free_ char *ifd_path = NULL;
+
+ ifd_path = path_join(NULL, sourcepath, de->d_name);
+ if (!ifd_path)
+ return log_oom();
+
+ _cleanup_free_ char *buf = NULL;
+ size_t buf_size;
+
+ /* Now read contents of pstore file */
+ r = read_full_file(ifd_path, &buf, &buf_size);
+ if (r < 0) {
+ log_warning_errno(r, "Failed to read file %s: %m", ifd_path);
+ continue;
+ }
+
+ if (!GREEDY_REALLOC(list->entries, list->n_entries_allocated, list->n_entries + 1))
+ return log_oom();
+
+ list->entries[list->n_entries++] = (PStoreEntry) {
+ .dirent = *de,
+ .content = TAKE_PTR(buf),
+ .content_size = buf_size,
+ .is_binary = true,
+ .handled = false,
+ };
+ }
+
+ return r;
+}
+
+static int run(int argc, char *argv[]) {
+ _cleanup_(pstore_entries_reset) PStoreList list = {};
+ int r;
+
+ log_open();
+
+ /* Ignore all parse errors */
+ (void) parse_config();
+
+ log_debug("Selected storage '%s'.", pstore_storage_to_string(arg_storage));
+ log_debug("Selected Unlink '%d'.", arg_unlink);
+
+ if (arg_storage == PSTORE_STORAGE_NONE)
+ /* Do nothing, intentionally, leaving pstore untouched */
+ return 0;
+
+ /* Obtain list of files in pstore */
+ r = list_files(&list, arg_sourcedir);
+ if (r < 0)
+ return r;
+
+ /* Handle each pstore file */
+ /* Sort files lexigraphically ascending, generally needed by all */
+ qsort_safe(list.entries, list.n_entries, sizeof(PStoreEntry), compare_pstore_entries);
+
+ /* Process known file types */
+ process_dmesg_files(&list);
+
+ /* Move left over files out of pstore */
+ for (size_t n = 0; n < list.n_entries; n++)
+ move_file(&list.entries[n], NULL);
+
+ return 0;
+}
+
+int main(int argc, char *argv[]) {
+ int r;
+
+ r = run(argc, argv);
+ return r < 0 ? EXIT_FAILURE : EXIT_SUCCESS;
+}
diff --git a/src/pstore/pstore.conf b/src/pstore/pstore.conf
new file mode 100644
index 0000000000..93a8b6707c
--- /dev/null
+++ b/src/pstore/pstore.conf
@@ -0,0 +1,16 @@
+# This file is part of systemd.
+#
+# systemd is free software; you can redistribute it and/or modify it
+# under the terms of the GNU Lesser General Public License as published by
+# the Free Software Foundation; either version 2.1 of the License, or
+# (at your option) any later version.
+#
+# Entries in this file show the compile time defaults.
+# You can change settings by editing this file.
+# Defaults can be restored by simply deleting this file.
+#
+# See pstore.conf(5) for details.
+
+[PStore]
+#Storage=external
+#Unlink=yes
diff --git a/units/meson.build b/units/meson.build
index a74fa95195..e8e64eb30a 100644
--- a/units/meson.build
+++ b/units/meson.build
@@ -136,6 +136,7 @@ in_units = [
['systemd-binfmt.service', 'ENABLE_BINFMT',
'sysinit.target.wants/'],
['systemd-coredump@.service', 'ENABLE_COREDUMP'],
+ ['systemd-pstore.service', 'ENABLE_PSTORE'],
['systemd-firstboot.service', 'ENABLE_FIRSTBOOT',
'sysinit.target.wants/'],
['systemd-fsck-root.service', ''],
diff --git a/units/systemd-pstore.service.in b/units/systemd-pstore.service.in
new file mode 100644
index 0000000000..fec2b1aebf
--- /dev/null
+++ b/units/systemd-pstore.service.in
@@ -0,0 +1,24 @@
+# SPDX-License-Identifier: LGPL-2.1+
+#
+# This file is part of systemd.
+#
+# systemd is free software; you can redistribute it and/or modify it
+# under the terms of the GNU Lesser General Public License as published by
+# the Free Software Foundation; either version 2.1 of the License, or
+# (at your option) any later version.
+
+[Unit]
+Description=Platform Persistent Storage Archival
+Documentation=man:systemd-pstore(8)
+DefaultDependencies=no
+Wants=systemd-remount-fs.service
+After=systemd-remount-fs.service
+
+[Service]
+Type=oneshot
+ExecStart=@rootlibexecdir@/systemd-pstore
+RemainAfterExit=yes
+StateDirectory=systemd/pstore
+
+[Install]
+WantedBy=systemd-remount-fs.service

27
SOURCES/0877-meson-drop-redundant-line.patch Normal file
View File

@ -0,0 +1,27 @@
From c95ba53ab720dfbd7f692e0a87d7f5d4f89ea36b Mon Sep 17 00:00:00 2001
From: Yu Watanabe <watanabe.yu+github@gmail.com>
Date: Mon, 22 Jul 2019 10:46:53 +0900
Subject: [PATCH] meson: drop redundant line
Found by @mattiasb.
(cherry picked from commit 3f708e7f6909faad307bdb60ed0f8d68e84f6584)
Related: #2158832
---
meson.build | 2 --
1 file changed, 2 deletions(-)
diff --git a/meson.build b/meson.build
index 972a8fb6f7..673800a1a7 100644
--- a/meson.build
+++ b/meson.build
@@ -2166,8 +2166,6 @@ if conf.get('ENABLE_PSTORE') == 1
install_rpath : rootlibexecdir,
install : true,
install_dir : rootlibexecdir)
-
- public_programs += exe
endif
if conf.get('ENABLE_BINFMT') == 1

48
SOURCES/0878-pstore-drop-unnecessary-initializations.patch Normal file
View File

@ -0,0 +1,48 @@
From 7e4b7cc35af0e3b3afbf32fa0fd9961cd01ad9a9 Mon Sep 17 00:00:00 2001
From: Yu Watanabe <watanabe.yu+github@gmail.com>
Date: Mon, 22 Jul 2019 10:52:12 +0900
Subject: [PATCH] pstore: drop unnecessary initializations
(cherry picked from commit 2e4effd129343d22bfed34e94810d3f87c8f0e85)
Related: #2158832
---
src/pstore/pstore.c | 15 ++++++---------
1 file changed, 6 insertions(+), 9 deletions(-)
diff --git a/src/pstore/pstore.c b/src/pstore/pstore.c
index f95e016eb6..e6a342fc50 100644
--- a/src/pstore/pstore.c
+++ b/src/pstore/pstore.c
@@ -113,14 +113,12 @@ static int compare_pstore_entries(const void *_a, const void *_b) {
}
static int move_file(PStoreEntry *pe, const char *subdir) {
- _cleanup_free_ char *ifd_path = NULL;
- _cleanup_free_ char *ofd_path = NULL;
- int r = 0;
- struct iovec iovec[2] = {};
- int n_iovec = 0;
+ _cleanup_free_ char *ifd_path = NULL, *ofd_path = NULL;
_cleanup_free_ void *field = NULL;
- const char *suffix = NULL;
+ struct iovec iovec[2];
+ const char *suffix;
size_t field_size;
+ int n_iovec = 0, r;
if (pe->handled)
return 0;
@@ -202,10 +200,9 @@ static int write_dmesg(const char *dmesg, size_t size, const char *id) {
static void process_dmesg_files(PStoreList *list) {
/* Move files, reconstruct dmesg.txt */
- PStoreEntry *pe;
- _cleanup_free_ char *dmesg = NULL;
+ _cleanup_free_ char *dmesg = NULL, *dmesg_id = NULL;
size_t dmesg_size = 0;
- _cleanup_free_ char *dmesg_id = NULL;
+ PStoreEntry *pe;
/* Handle each dmesg file: files processed in reverse
* order so as to properly reconstruct original dmesg */

46
SOURCES/0879-pstopre-fix-return-value-of-list_files.patch Normal file
View File

@ -0,0 +1,46 @@
From a0485b96118d3d2ac439f510e404ffb3db03e23f Mon Sep 17 00:00:00 2001
From: Yu Watanabe <watanabe.yu+github@gmail.com>
Date: Mon, 22 Jul 2019 10:55:10 +0900
Subject: [PATCH] pstopre: fix return value of list_files()
Previously, the return value of the last read_full_file() is returned.
This makes the error in read_full_file() is always ignored.
(cherry picked from commit 337874a45fff46a80e4974c681a5e651f3a0fac9)
Related: #2158832
---
src/pstore/pstore.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/src/pstore/pstore.c b/src/pstore/pstore.c
index e6a342fc50..2fbef48543 100644
--- a/src/pstore/pstore.c
+++ b/src/pstore/pstore.c
@@ -311,7 +311,7 @@ static void process_dmesg_files(PStoreList *list) {
static int list_files(PStoreList *list, const char *sourcepath) {
_cleanup_(closedirp) DIR *dirp = NULL;
struct dirent *de;
- int r = 0;
+ int r;
dirp = opendir(sourcepath);
if (!dirp)
@@ -330,7 +330,7 @@ static int list_files(PStoreList *list, const char *sourcepath) {
/* Now read contents of pstore file */
r = read_full_file(ifd_path, &buf, &buf_size);
if (r < 0) {
- log_warning_errno(r, "Failed to read file %s: %m", ifd_path);
+ log_warning_errno(r, "Failed to read file %s, skipping: %m", ifd_path);
continue;
}
@@ -346,7 +346,7 @@ static int list_files(PStoreList *list, const char *sourcepath) {
};
}
- return r;
+ return 0;
}
static int run(int argc, char *argv[]) {

36
SOURCES/0880-pstore-remove-temporary-file-on-failure.patch Normal file
View File

@ -0,0 +1,36 @@
From 58000dc7dd93ff6e8357de64154b0849d3c17c5d Mon Sep 17 00:00:00 2001
From: Yu Watanabe <watanabe.yu+github@gmail.com>
Date: Mon, 22 Jul 2019 11:01:43 +0900
Subject: [PATCH] pstore: remove temporary file on failure
(cherry picked from commit 03c5f6cc02648eeff3179b2b762d46b9e1889bb1)
Related: #2158832
---
src/pstore/pstore.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/src/pstore/pstore.c b/src/pstore/pstore.c
index 2fbef48543..ce8080ceed 100644
--- a/src/pstore/pstore.c
+++ b/src/pstore/pstore.c
@@ -167,8 +167,8 @@ static int move_file(PStoreEntry *pe, const char *subdir) {
}
static int write_dmesg(const char *dmesg, size_t size, const char *id) {
- _cleanup_(unlink_and_freep) char *ofd_path = NULL;
- _cleanup_free_ char *tmp_path = NULL;
+ _cleanup_(unlink_and_freep) char *tmp_path = NULL;
+ _cleanup_free_ char *ofd_path = NULL;
_cleanup_close_ int ofd = -1;
ssize_t wr;
int r;
@@ -193,7 +193,7 @@ static int write_dmesg(const char *dmesg, size_t size, const char *id) {
r = link_tmpfile(ofd, tmp_path, ofd_path);
if (r < 0)
return log_error_errno(r, "Failed to write temporary file %s: %m", ofd_path);
- ofd_path = mfree(ofd_path);
+ tmp_path = mfree(tmp_path);
return 0;
}

57
SOURCES/0881-pstore-do-not-add-FILE-journal-entry-if-content_size.patch Normal file
View File

@ -0,0 +1,57 @@
From 5006e4bd9aecea40ca3d907adc692c4c8001a6c1 Mon Sep 17 00:00:00 2001
From: Yu Watanabe <watanabe.yu+github@gmail.com>
Date: Mon, 22 Jul 2019 11:08:06 +0900
Subject: [PATCH] pstore: do not add FILE= journal entry if content_size == 0
(cherry picked from commit 6bf18debddbe1b231f783617e054cc194bb36d1e)
Related: #2158832
---
src/pstore/pstore.c | 25 ++++++++++++++-----------
1 file changed, 14 insertions(+), 11 deletions(-)
diff --git a/src/pstore/pstore.c b/src/pstore/pstore.c
index ce8080ceed..eb251d61c8 100644
--- a/src/pstore/pstore.c
+++ b/src/pstore/pstore.c
@@ -114,10 +114,8 @@ static int compare_pstore_entries(const void *_a, const void *_b) {
static int move_file(PStoreEntry *pe, const char *subdir) {
_cleanup_free_ char *ifd_path = NULL, *ofd_path = NULL;
- _cleanup_free_ void *field = NULL;
+ const char *suffix, *message;
struct iovec iovec[2];
- const char *suffix;
- size_t field_size;
int n_iovec = 0, r;
if (pe->handled)
@@ -133,15 +131,20 @@ static int move_file(PStoreEntry *pe, const char *subdir) {
/* Always log to the journal */
suffix = arg_storage == PSTORE_STORAGE_EXTERNAL ? strjoina(" moved to ", ofd_path) : (char *)".";
- field = strjoina("MESSAGE=PStore ", pe->dirent.d_name, suffix);
- iovec[n_iovec++] = IOVEC_MAKE_STRING(field);
+ message = strjoina("MESSAGE=PStore ", pe->dirent.d_name, suffix);
+ iovec[n_iovec++] = IOVEC_MAKE_STRING(message);
- field_size = strlen("FILE=") + pe->content_size;
- field = malloc(field_size);
- if (!field)
- return log_oom();
- memcpy(stpcpy(field, "FILE="), pe->content, pe->content_size);
- iovec[n_iovec++] = IOVEC_MAKE(field, field_size);
+ if (pe->content_size > 0) {
+ _cleanup_free_ void *field = NULL;
+ size_t field_size;
+
+ field_size = strlen("FILE=") + pe->content_size;
+ field = malloc(field_size);
+ if (!field)
+ return log_oom();
+ memcpy(stpcpy(field, "FILE="), pe->content, pe->content_size);
+ iovec[n_iovec++] = IOVEC_MAKE(field, field_size);
+ }
r = sd_journal_sendv(iovec, n_iovec);
if (r < 0)

24
SOURCES/0882-pstore-run-only-when-sys-fs-pstore-is-not-empty.patch Normal file
View File

@ -0,0 +1,24 @@
From a7247899f156761934bcb4b380861b3d3ec5449f Mon Sep 17 00:00:00 2001
From: Yu Watanabe <watanabe.yu+github@gmail.com>
Date: Mon, 22 Jul 2019 14:09:12 +0900
Subject: [PATCH] pstore: run only when /sys/fs/pstore is not empty
(cherry picked from commit 6d4f213b1f6afb2901f0d97cec0e28e20809b713)
Related: #2158832
---
units/systemd-pstore.service.in | 1 +
1 file changed, 1 insertion(+)
diff --git a/units/systemd-pstore.service.in b/units/systemd-pstore.service.in
index fec2b1aebf..dde21bc33e 100644
--- a/units/systemd-pstore.service.in
+++ b/units/systemd-pstore.service.in
@@ -10,6 +10,7 @@
[Unit]
Description=Platform Persistent Storage Archival
Documentation=man:systemd-pstore(8)
+ConditionDirectoryNotEmpty=/sys/fs/pstore
DefaultDependencies=no
Wants=systemd-remount-fs.service
After=systemd-remount-fs.service

34
SOURCES/0883-pstore-fix-use-after-free.patch Normal file
View File

@ -0,0 +1,34 @@
From 7f5bfbd5485e1cb779d7568cabb5783651fd9da3 Mon Sep 17 00:00:00 2001
From: Michael Olbrich <m.olbrich@pengutronix.de>
Date: Fri, 6 Sep 2019 15:04:01 +0200
Subject: [PATCH] pstore: fix use after free
The memory is still needed in the sd_journal_sendv() after the 'if' block.
(cherry picked from commit 1e19f5ac0d680a63eccae7ef1fc6ce225dca0bbf)
Related: #2158832
---
src/pstore/pstore.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/pstore/pstore.c b/src/pstore/pstore.c
index eb251d61c8..cafb1804c6 100644
--- a/src/pstore/pstore.c
+++ b/src/pstore/pstore.c
@@ -114,6 +114,7 @@ static int compare_pstore_entries(const void *_a, const void *_b) {
static int move_file(PStoreEntry *pe, const char *subdir) {
_cleanup_free_ char *ifd_path = NULL, *ofd_path = NULL;
+ _cleanup_free_ void *field = NULL;
const char *suffix, *message;
struct iovec iovec[2];
int n_iovec = 0, r;
@@ -135,7 +136,6 @@ static int move_file(PStoreEntry *pe, const char *subdir) {
iovec[n_iovec++] = IOVEC_MAKE_STRING(message);
if (pe->content_size > 0) {
- _cleanup_free_ void *field = NULL;
size_t field_size;
field_size = strlen("FILE=") + pe->content_size;

29
SOURCES/0884-pstore-refuse-to-run-if-arguments-are-specified.patch Normal file
View File

@ -0,0 +1,29 @@
From a35a90322f8587f650aeb72bfbe1ebcc93e503aa Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Zbigniew=20J=C4=99drzejewski-Szmek?= <zbyszek@in.waw.pl>
Date: Mon, 22 Jul 2019 10:43:19 +0200
Subject: [PATCH] pstore: refuse to run if arguments are specified
(This is why the --help chech passed.)
(cherry picked from commit 22d6bea8820612e6a1483d8b6cfd820f1417ae6b)
Related: #2158832
---
src/pstore/pstore.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/src/pstore/pstore.c b/src/pstore/pstore.c
index cafb1804c6..b0b21dedd4 100644
--- a/src/pstore/pstore.c
+++ b/src/pstore/pstore.c
@@ -358,6 +358,10 @@ static int run(int argc, char *argv[]) {
log_open();
+ if (argc > 1)
+ return log_error_errno(-EINVAL,
+ "This program takes no arguments.");
+
/* Ignore all parse errors */
(void) parse_config();

44
SOURCES/0885-pstore-allow-specifying-src-and-dst-dirs-are-argumen.patch Normal file
View File

@ -0,0 +1,44 @@
From 76fe0974f62f0a2beb2d3d8e224e80a57c0ebd09 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Zbigniew=20J=C4=99drzejewski-Szmek?= <zbyszek@in.waw.pl>
Date: Fri, 4 Oct 2019 16:14:47 +0200
Subject: [PATCH] pstore: allow specifying src and dst dirs are arguments
This makes it much easier to debug the program as a normal user, since we
don't need to set up fake input under /sys/fs/pstore/.
Also, let's make the debug output a bit nicer.
(cherry picked from commit e05a72d7587ff916a983588f8393af624d330dd0)
Related: #2158832
---
src/pstore/pstore.c | 11 +++++++----
1 file changed, 7 insertions(+), 4 deletions(-)
diff --git a/src/pstore/pstore.c b/src/pstore/pstore.c
index b0b21dedd4..7353e83a81 100644
--- a/src/pstore/pstore.c
+++ b/src/pstore/pstore.c
@@ -358,15 +358,18 @@ static int run(int argc, char *argv[]) {
log_open();
- if (argc > 1)
+ if (argc == 3) {
+ arg_sourcedir = argv[1];
+ arg_archivedir = argv[2];
+ } else if (argc > 1)
return log_error_errno(-EINVAL,
- "This program takes no arguments.");
+ "This program takes zero or two arguments.");
/* Ignore all parse errors */
(void) parse_config();
- log_debug("Selected storage '%s'.", pstore_storage_to_string(arg_storage));
- log_debug("Selected Unlink '%d'.", arg_unlink);
+ log_debug("Selected storage: %s.", pstore_storage_to_string(arg_storage));
+ log_debug("Selected unlink: %s.", yes_no(arg_unlink));
if (arg_storage == PSTORE_STORAGE_NONE)
/* Do nothing, intentionally, leaving pstore untouched */

133
SOURCES/0886-pstore-rework-memory-handling-for-dmesg.patch Normal file
View File

@ -0,0 +1,133 @@
From a2ba34a79de3748f51d57541c54dbe22e1d03a9e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Zbigniew=20J=C4=99drzejewski-Szmek?= <zbyszek@in.waw.pl>
Date: Fri, 4 Oct 2019 16:17:27 +0200
Subject: [PATCH] pstore: rework memory handling for dmesg
Semmle Security Reports report:
> The problem occurs on the way realloc is being used. When a size
> bigger than the chunk that wants to be reallocated is passed, realloc
> try to malloc a bigger size, however in the case that malloc fails
> (for example, by forcing a big allocation) realloc will return NULL.
>
> According to the man page:
> "The realloc() function returns a pointer to the newly allocated
> memory, which is suitably aligned for any built-in type and may be
> different from ptr, or NULL if the request fails. If size was
> equal to 0, either NULL or a pointer suitable to be passed to free()
> is returned. If realloc() fails, the original block is left
> untouched; it is not freed or moved."
>
> The problem occurs when the memory ptr passed to the first argument of
> realloc is the same as the one used for the result, for example in
> this case:
>
> dmesg = realloc(dmesg, dmesg_size + strlen(pe->dirent.d_name) +
> strlen(":\n") + pe->content_size + 1);
>
> https://lgtm.com/projects/g/systemd/systemd/snapshot/f8bcb81955f9e93a4787627e28f43fffb2a84836/files/src/pstore/pstore.c?sort=name&dir=A
> SC&mode=heatmap#L300
>
> If the malloc inside that realloc fails, then the original memory
> chunk will never be free but since realloc will return NULL, the
> pointer to that memory chunk will be lost and a memory leak will
> occur.
>
> In case you are curious, this is the query we used to find this problem:
> https://lgtm.com/query/8650323308193591473/
Let's use a more standard pattern: allocate memory using greedy_realloc, and
instead of freeing it when we wrote out a chunk, let's just move the cursor
back to the beginning and reuse the memory we allocated previously.
If we fail to allocate the memory for dmesg contents, don't write the dmesg
entry, but let's still process the files to move them out of pstore.
(cherry picked from commit 8198c3e42b0614b6bd1db6f38813b842c8577304)
Related: #2158832
---
src/pstore/pstore.c | 43 ++++++++++++++++++++++++++-----------------
1 file changed, 26 insertions(+), 17 deletions(-)
diff --git a/src/pstore/pstore.c b/src/pstore/pstore.c
index 7353e83a81..d70e142b4d 100644
--- a/src/pstore/pstore.c
+++ b/src/pstore/pstore.c
@@ -176,9 +176,11 @@ static int write_dmesg(const char *dmesg, size_t size, const char *id) {
ssize_t wr;
int r;
- if (isempty(dmesg) || size == 0)
+ if (size == 0)
return 0;
+ assert(dmesg);
+
/* log_info("Record ID %s", id); */
ofd_path = path_join(arg_archivedir, id, "dmesg.txt");
@@ -204,7 +206,8 @@ static int write_dmesg(const char *dmesg, size_t size, const char *id) {
static void process_dmesg_files(PStoreList *list) {
/* Move files, reconstruct dmesg.txt */
_cleanup_free_ char *dmesg = NULL, *dmesg_id = NULL;
- size_t dmesg_size = 0;
+ size_t dmesg_size = 0, dmesg_allocated = 0;
+ bool dmesg_bad = false;
PStoreEntry *pe;
/* Handle each dmesg file: files processed in reverse
@@ -281,33 +284,39 @@ static void process_dmesg_files(PStoreList *list) {
/* Now move file from pstore to archive storage */
move_file(pe, pe_id);
+ if (dmesg_bad)
+ continue;
+
/* If the current record id is NOT the same as the
* previous record id, then start a new dmesg.txt file */
- if (!pe_id || !dmesg_id || !streq(pe_id, dmesg_id)) {
+ if (!streq_ptr(pe_id, dmesg_id)) {
/* Encountered a new dmesg group, close out old one, open new one */
- if (dmesg) {
- (void) write_dmesg(dmesg, dmesg_size, dmesg_id);
- dmesg = mfree(dmesg);
- dmesg_size = 0;
- }
+ (void) write_dmesg(dmesg, dmesg_size, dmesg_id);
+ dmesg_size = 0;
/* now point dmesg_id to storage of pe_id */
free_and_replace(dmesg_id, pe_id);
}
- /* Reconstruction of dmesg is done as a useful courtesy, do not log errors */
- dmesg = realloc(dmesg, dmesg_size + strlen(pe->dirent.d_name) + strlen(":\n") + pe->content_size + 1);
- if (dmesg) {
- dmesg_size += sprintf(&dmesg[dmesg_size], "%s:\n", pe->dirent.d_name);
- if (pe->content) {
- memcpy(&dmesg[dmesg_size], pe->content, pe->content_size);
- dmesg_size += pe->content_size;
- }
+ /* Reconstruction of dmesg is done as a useful courtesy: do not fail, but don't write garbled
+ * output either. */
+ size_t needed = strlen(pe->dirent.d_name) + strlen(":\n") + pe->content_size + 1;
+ if (!GREEDY_REALLOC(dmesg, dmesg_allocated, dmesg_size + needed)) {
+ log_warning_errno(ENOMEM, "Failed to write dmesg file: %m");
+ dmesg_bad = true;
+ continue;
+ }
+
+ dmesg_size += sprintf(dmesg + dmesg_size, "%s:\n", pe->dirent.d_name);
+ if (pe->content) {
+ memcpy(dmesg + dmesg_size, pe->content, pe->content_size);
+ dmesg_size += pe->content_size;
}
pe_id = mfree(pe_id);
}
- if (dmesg)
+
+ if (!dmesg_bad)
(void) write_dmesg(dmesg, dmesg_size, dmesg_id);
}

504
SOURCES/0887-pstore-fixes-for-dmesg.txt-reconstruction.patch Normal file
View File

@ -0,0 +1,504 @@
From 5ac15c7dc49476e7cd7cc3a4b507282c9f78d528 Mon Sep 17 00:00:00 2001
From: Eric DeVolder <eric.devolder@oracle.com>
Date: Mon, 21 Nov 2022 11:27:27 -0500
Subject: [PATCH] pstore: fixes for dmesg.txt reconstruction
This patch fixes problems with the re-assembly of the dmesg
from the records stored in pstore.
The current code simply ignores the last 6 characters of the
file name to form a base record id, which then groups any
pstore files with this base id into the reconstructed dmesg.txt.
This approach fails when the following oops generated the
following in pstore:
-rw-------. 1 root root 1808 Oct 27 22:07 dmesg-efi-166692286101001
-rw-------. 1 root root 1341 Oct 27 22:07 dmesg-efi-166692286101002
-rw-------. 1 root root 1812 Oct 27 22:07 dmesg-efi-166692286102001
-rw-------. 1 root root 1820 Oct 27 22:07 dmesg-efi-166692286102002
-rw-------. 1 root root 1807 Oct 27 22:07 dmesg-efi-166692286103001
-rw-------. 1 root root 1791 Oct 27 22:07 dmesg-efi-166692286103002
-rw-------. 1 root root 1773 Oct 27 22:07 dmesg-efi-166692286104001
-rw-------. 1 root root 1801 Oct 27 22:07 dmesg-efi-166692286104002
-rw-------. 1 root root 1821 Oct 27 22:07 dmesg-efi-166692286105001
-rw-------. 1 root root 1809 Oct 27 22:07 dmesg-efi-166692286105002
-rw-------. 1 root root 1804 Oct 27 22:07 dmesg-efi-166692286106001
-rw-------. 1 root root 1817 Oct 27 22:07 dmesg-efi-166692286106002
-rw-------. 1 root root 1792 Oct 27 22:07 dmesg-efi-166692286107001
-rw-------. 1 root root 1810 Oct 27 22:07 dmesg-efi-166692286107002
-rw-------. 1 root root 1717 Oct 27 22:07 dmesg-efi-166692286108001
-rw-------. 1 root root 1808 Oct 27 22:07 dmesg-efi-166692286108002
-rw-------. 1 root root 1764 Oct 27 22:07 dmesg-efi-166692286109001
-rw-------. 1 root root 1765 Oct 27 22:07 dmesg-efi-166692286109002
-rw-------. 1 root root 1796 Oct 27 22:07 dmesg-efi-166692286110001
-rw-------. 1 root root 1816 Oct 27 22:07 dmesg-efi-166692286110002
-rw-------. 1 root root 1793 Oct 27 22:07 dmesg-efi-166692286111001
-rw-------. 1 root root 1751 Oct 27 22:07 dmesg-efi-166692286111002
-rw-------. 1 root root 1813 Oct 27 22:07 dmesg-efi-166692286112001
-rw-------. 1 root root 1786 Oct 27 22:07 dmesg-efi-166692286112002
-rw-------. 1 root root 1754 Oct 27 22:07 dmesg-efi-166692286113001
-rw-------. 1 root root 1752 Oct 27 22:07 dmesg-efi-166692286113002
-rw-------. 1 root root 1803 Oct 27 22:07 dmesg-efi-166692286114001
-rw-------. 1 root root 1759 Oct 27 22:07 dmesg-efi-166692286114002
-rw-------. 1 root root 1805 Oct 27 22:07 dmesg-efi-166692286115001
-rw-------. 1 root root 1787 Oct 27 22:07 dmesg-efi-166692286115002
-rw-------. 1 root root 1815 Oct 27 22:07 dmesg-efi-166692286116001
-rw-------. 1 root root 1771 Oct 27 22:07 dmesg-efi-166692286116002
-rw-------. 1 root root 1816 Oct 27 22:07 dmesg-efi-166692286117002
-rw-------. 1 root root 1388 Oct 27 22:07 dmesg-efi-166692286701003
-rw-------. 1 root root 1824 Oct 27 22:07 dmesg-efi-166692286702003
-rw-------. 1 root root 1795 Oct 27 22:07 dmesg-efi-166692286703003
-rw-------. 1 root root 1805 Oct 27 22:07 dmesg-efi-166692286704003
-rw-------. 1 root root 1813 Oct 27 22:07 dmesg-efi-166692286705003
-rw-------. 1 root root 1821 Oct 27 22:07 dmesg-efi-166692286706003
-rw-------. 1 root root 1814 Oct 27 22:07 dmesg-efi-166692286707003
-rw-------. 1 root root 1812 Oct 27 22:07 dmesg-efi-166692286708003
-rw-------. 1 root root 1769 Oct 27 22:07 dmesg-efi-166692286709003
-rw-------. 1 root root 1820 Oct 27 22:07 dmesg-efi-166692286710003
-rw-------. 1 root root 1755 Oct 27 22:07 dmesg-efi-166692286711003
-rw-------. 1 root root 1790 Oct 27 22:07 dmesg-efi-166692286712003
-rw-------. 1 root root 1756 Oct 27 22:07 dmesg-efi-166692286713003
-rw-------. 1 root root 1763 Oct 27 22:07 dmesg-efi-166692286714003
-rw-------. 1 root root 1791 Oct 27 22:07 dmesg-efi-166692286715003
-rw-------. 1 root root 1775 Oct 27 22:07 dmesg-efi-166692286716003
-rw-------. 1 root root 1820 Oct 27 22:07 dmesg-efi-166692286717003
The "reconstructed" dmesg.txt that resulted from the above contained
the following (ignoring actual contents, just providing the Part info):
Emergency#3 Part17
Emergency#3 Part16
Emergency#3 Part15
Emergency#3 Part14
Emergency#3 Part13
Emergency#3 Part12
Emergency#3 Part11
Emergency#3 Part10
Emergency#3 Part9
Emergency#3 Part8
Emergency#3 Part7
Emergency#3 Part6
Emergency#3 Part5
Emergency#3 Part4
Emergency#3 Part3
Emergency#3 Part2
Emergency#3 Part1
Panic#2 Part17
Panic#2 Part16
Oops#1 Part16
Panic#2 Part15
Oops#1 Part15
Panic#2 Part14
Oops#1 Part14
Panic#2 Part13
Oops#1 Part13
Panic#2 Part12
Oops#1 Part12
Panic#2 Part11
Oops#1 Part11
Panic#2 Part10
Oops#1 Part10
Panic#2 Part9
Oops#1 Part9
Panic#2 Part8
Oops#1 Part8
Panic#2 Part7
Oops#1 Part7
Panic#2 Part6
Oops#1 Part6
Panic#2 Part5
Oops#1 Part5
Panic#2 Part4
Oops#1 Part4
Panic#2 Part3
Oops#1 Part3
Panic#2 Part2
Oops#1 Part2
Panic#2 Part1
Oops#1 Part1
The above is a interleaved mess of three dmesg dumps.
This patch fixes the above problems, and simplifies the dmesg
reconstruction process. The code now distinguishes between
records on EFI vs ERST, which have differently formatted
record identifiers. Using knowledge of the format of the
record ids allows vastly improved reconstruction process.
With this change in place, the above pstore records now
result in the following:
# ls -alR /var/lib/systemd/pstore
1666922861:
total 8
drwxr-xr-x. 4 root root 28 Nov 18 14:58 .
drwxr-xr-x. 7 root root 144 Nov 18 14:58 ..
drwxr-xr-x. 2 root root 4096 Nov 18 14:58 001
drwxr-xr-x. 2 root root 4096 Nov 18 14:58 002
1666922861/001:
total 100
drwxr-xr-x. 2 root root 4096 Nov 18 14:58 .
drwxr-xr-x. 4 root root 28 Nov 18 14:58 ..
-rw-------. 1 root root 1808 Oct 27 22:07 dmesg-efi-166692286101001
-rw-------. 1 root root 1812 Oct 27 22:07 dmesg-efi-166692286102001
-rw-------. 1 root root 1807 Oct 27 22:07 dmesg-efi-166692286103001
-rw-------. 1 root root 1773 Oct 27 22:07 dmesg-efi-166692286104001
-rw-------. 1 root root 1821 Oct 27 22:07 dmesg-efi-166692286105001
-rw-------. 1 root root 1804 Oct 27 22:07 dmesg-efi-166692286106001
-rw-------. 1 root root 1792 Oct 27 22:07 dmesg-efi-166692286107001
-rw-------. 1 root root 1717 Oct 27 22:07 dmesg-efi-166692286108001
-rw-------. 1 root root 1764 Oct 27 22:07 dmesg-efi-166692286109001
-rw-------. 1 root root 1796 Oct 27 22:07 dmesg-efi-166692286110001
-rw-------. 1 root root 1793 Oct 27 22:07 dmesg-efi-166692286111001
-rw-------. 1 root root 1813 Oct 27 22:07 dmesg-efi-166692286112001
-rw-------. 1 root root 1754 Oct 27 22:07 dmesg-efi-166692286113001
-rw-------. 1 root root 1803 Oct 27 22:07 dmesg-efi-166692286114001
-rw-------. 1 root root 1805 Oct 27 22:07 dmesg-efi-166692286115001
-rw-------. 1 root root 1815 Oct 27 22:07 dmesg-efi-166692286116001
-rw-r-----. 1 root root 28677 Nov 18 14:58 dmesg.txt
1666922861/002:
total 104
drwxr-xr-x. 2 root root 4096 Nov 18 14:58 .
drwxr-xr-x. 4 root root 28 Nov 18 14:58 ..
-rw-------. 1 root root 1341 Oct 27 22:07 dmesg-efi-166692286101002
-rw-------. 1 root root 1820 Oct 27 22:07 dmesg-efi-166692286102002
-rw-------. 1 root root 1791 Oct 27 22:07 dmesg-efi-166692286103002
-rw-------. 1 root root 1801 Oct 27 22:07 dmesg-efi-166692286104002
-rw-------. 1 root root 1809 Oct 27 22:07 dmesg-efi-166692286105002
-rw-------. 1 root root 1817 Oct 27 22:07 dmesg-efi-166692286106002
-rw-------. 1 root root 1810 Oct 27 22:07 dmesg-efi-166692286107002
-rw-------. 1 root root 1808 Oct 27 22:07 dmesg-efi-166692286108002
-rw-------. 1 root root 1765 Oct 27 22:07 dmesg-efi-166692286109002
-rw-------. 1 root root 1816 Oct 27 22:07 dmesg-efi-166692286110002
-rw-------. 1 root root 1751 Oct 27 22:07 dmesg-efi-166692286111002
-rw-------. 1 root root 1786 Oct 27 22:07 dmesg-efi-166692286112002
-rw-------. 1 root root 1752 Oct 27 22:07 dmesg-efi-166692286113002
-rw-------. 1 root root 1759 Oct 27 22:07 dmesg-efi-166692286114002
-rw-------. 1 root root 1787 Oct 27 22:07 dmesg-efi-166692286115002
-rw-------. 1 root root 1771 Oct 27 22:07 dmesg-efi-166692286116002
-rw-------. 1 root root 1816 Oct 27 22:07 dmesg-efi-166692286117002
-rw-r-----. 1 root root 30000 Nov 18 14:58 dmesg.txt
1666922867:
total 4
drwxr-xr-x. 3 root root 17 Nov 18 14:58 .
drwxr-xr-x. 7 root root 144 Nov 18 14:58 ..
drwxr-xr-x. 2 root root 4096 Nov 18 14:58 003
1666922867/003:
total 104
drwxr-xr-x. 2 root root 4096 Nov 18 14:58 .
drwxr-xr-x. 3 root root 17 Nov 18 14:58 ..
-rw-------. 1 root root 1388 Oct 27 22:07 dmesg-efi-166692286701003
-rw-------. 1 root root 1824 Oct 27 22:07 dmesg-efi-166692286702003
-rw-------. 1 root root 1795 Oct 27 22:07 dmesg-efi-166692286703003
-rw-------. 1 root root 1805 Oct 27 22:07 dmesg-efi-166692286704003
-rw-------. 1 root root 1813 Oct 27 22:07 dmesg-efi-166692286705003
-rw-------. 1 root root 1821 Oct 27 22:07 dmesg-efi-166692286706003
-rw-------. 1 root root 1814 Oct 27 22:07 dmesg-efi-166692286707003
-rw-------. 1 root root 1812 Oct 27 22:07 dmesg-efi-166692286708003
-rw-------. 1 root root 1769 Oct 27 22:07 dmesg-efi-166692286709003
-rw-------. 1 root root 1820 Oct 27 22:07 dmesg-efi-166692286710003
-rw-------. 1 root root 1755 Oct 27 22:07 dmesg-efi-166692286711003
-rw-------. 1 root root 1790 Oct 27 22:07 dmesg-efi-166692286712003
-rw-------. 1 root root 1756 Oct 27 22:07 dmesg-efi-166692286713003
-rw-------. 1 root root 1763 Oct 27 22:07 dmesg-efi-166692286714003
-rw-------. 1 root root 1791 Oct 27 22:07 dmesg-efi-166692286715003
-rw-------. 1 root root 1775 Oct 27 22:07 dmesg-efi-166692286716003
-rw-------. 1 root root 1820 Oct 27 22:07 dmesg-efi-166692286717003
-rw-r-----. 1 root root 30111 Nov 18 14:58 dmesg.txt
Furthemore, pstore records on ERST are now able to accurately
identify the change in timestamp sequence in order to start a
new dmesg.txt, as needed.
(cherry picked from commit 5fbaa757077bde2db8d33b1c358518c41b990339)
Related: #2158832
---
src/pstore/pstore.c | 216 +++++++++++++++++++-------------------------
1 file changed, 92 insertions(+), 124 deletions(-)
diff --git a/src/pstore/pstore.c b/src/pstore/pstore.c
index d70e142b4d..9f61e8f7f8 100644
--- a/src/pstore/pstore.c
+++ b/src/pstore/pstore.c
@@ -112,8 +112,8 @@ static int compare_pstore_entries(const void *_a, const void *_b) {
return strcmp(a->dirent.d_name, b->dirent.d_name);
}
-static int move_file(PStoreEntry *pe, const char *subdir) {
- _cleanup_free_ char *ifd_path = NULL, *ofd_path = NULL;
+static int move_file(PStoreEntry *pe, const char *subdir1, const char *subdir2) {
+ _cleanup_free_ char *ifd_path = NULL, *ofd_path = NULL, *ofd_path_base = NULL;
_cleanup_free_ void *field = NULL;
const char *suffix, *message;
struct iovec iovec[2];
@@ -126,7 +126,11 @@ static int move_file(PStoreEntry *pe, const char *subdir) {
if (!ifd_path)
return log_oom();
- ofd_path = path_join(arg_archivedir, subdir, pe->dirent.d_name);
+ ofd_path_base = path_join(arg_archivedir, subdir1, subdir2);
+ if (!ofd_path_base)
+ return log_oom();
+
+ ofd_path = path_join(NULL, ofd_path_base, pe->dirent.d_name);
if (!ofd_path)
return log_oom();
@@ -169,155 +173,119 @@ static int move_file(PStoreEntry *pe, const char *subdir) {
return 0;
}
-static int write_dmesg(const char *dmesg, size_t size, const char *id) {
- _cleanup_(unlink_and_freep) char *tmp_path = NULL;
- _cleanup_free_ char *ofd_path = NULL;
+static int append_dmesg(PStoreEntry *pe, const char *subdir1, const char *subdir2) {
+ /* Append dmesg chunk to end, create if needed */
+ _cleanup_free_ char *ofd_path = NULL, *ofd_path_base = NULL;
_cleanup_close_ int ofd = -1;
ssize_t wr;
- int r;
- if (size == 0)
- return 0;
+ assert(pe);
- assert(dmesg);
+ if (pe->content_size == 0)
+ return 0;
- /* log_info("Record ID %s", id); */
+ ofd_path_base = path_join(arg_archivedir, subdir1, subdir2);
+ if (!ofd_path_base)
+ return log_oom();
- ofd_path = path_join(arg_archivedir, id, "dmesg.txt");
+ ofd_path = path_join(NULL, ofd_path_base, "dmesg.txt");
if (!ofd_path)
return log_oom();
- ofd = open_tmpfile_linkable(ofd_path, O_CLOEXEC|O_CREAT|O_TRUNC|O_WRONLY, &tmp_path);
+ ofd = open(ofd_path, O_CREAT|O_NOFOLLOW|O_NOCTTY|O_CLOEXEC|O_APPEND|O_WRONLY, 0640);
if (ofd < 0)
- return log_error_errno(ofd, "Failed to open temporary file %s: %m", ofd_path);
- wr = write(ofd, dmesg, size);
+ return log_error_errno(ofd, "Failed to open file %s: %m", ofd_path);
+ wr = write(ofd, pe->content, pe->content_size);
if (wr < 0)
return log_error_errno(errno, "Failed to store dmesg to %s: %m", ofd_path);
- if (wr != (ssize_t)size)
- return log_error_errno(-EIO, "Failed to store dmesg to %s. %zu bytes are lost.", ofd_path, size - wr);
- r = link_tmpfile(ofd, tmp_path, ofd_path);
- if (r < 0)
- return log_error_errno(r, "Failed to write temporary file %s: %m", ofd_path);
- tmp_path = mfree(tmp_path);
+ if ((size_t)wr != pe->content_size)
+ return log_error_errno(-EIO, "Failed to store dmesg to %s. %zu bytes are lost.", ofd_path, pe->content_size - wr);
return 0;
}
-static void process_dmesg_files(PStoreList *list) {
+static int process_dmesg_files(PStoreList *list) {
/* Move files, reconstruct dmesg.txt */
- _cleanup_free_ char *dmesg = NULL, *dmesg_id = NULL;
- size_t dmesg_size = 0, dmesg_allocated = 0;
- bool dmesg_bad = false;
- PStoreEntry *pe;
+ _cleanup_free_ char *erst_subdir = NULL;
+ uint64_t last_record_id = 0;
+
+ /* When dmesg is written into pstore, it is done so in small chunks, whatever the exchange buffer
+ * size is with the underlying pstore backend (ie. EFI may be ~2KiB), which means an example
+ * pstore with approximately 64KB of storage may have up to roughly 32 dmesg files, some likely
+ * related.
+ *
+ * Here we look at the dmesg filename and try to discern if files are part of a related group,
+ * meaning the same original dmesg.
+ *
+ * The dmesg- filename contains the backend-type and the Common Platform Error Record, CPER,
+ * record id, a 64-bit number.
+ *
+ * Files are processed in reverse lexigraphical order so as to properly reconstruct original dmesg.*/
- /* Handle each dmesg file: files processed in reverse
- * order so as to properly reconstruct original dmesg */
for (size_t n = list->n_entries; n > 0; n--) {
- bool move_file_and_continue = false;
- _cleanup_free_ char *pe_id = NULL;
+ PStoreEntry *pe;
char *p;
- size_t plen;
pe = &list->entries[n-1];
if (pe->handled)
continue;
- if (!startswith(pe->dirent.d_name, "dmesg-"))
- continue;
-
if (endswith(pe->dirent.d_name, ".enc.z")) /* indicates a problem */
- move_file_and_continue = true;
- p = strrchr(pe->dirent.d_name, '-');
- if (!p)
- move_file_and_continue = true;
-
- if (move_file_and_continue) {
- /* A dmesg file on which we do NO additional processing */
- (void) move_file(pe, NULL);
continue;
- }
-
- /* See if this file is one of a related group of files
- * in order to reconstruct dmesg */
-
- /* When dmesg is written into pstore, it is done so in
- * small chunks, whatever the exchange buffer size is
- * with the underlying pstore backend (ie. EFI may be
- * ~2KiB), which means an example pstore with approximately
- * 64KB of storage may have up to roughly 32 dmesg files
- * that could be related, depending upon the size of the
- * original dmesg.
- *
- * Here we look at the dmesg filename and try to discern
- * if files are part of a related group, meaning the same
- * original dmesg.
- *
- * The two known pstore backends are EFI and ERST. These
- * backends store data in the Common Platform Error
- * Record, CPER, format. The dmesg- filename contains the
- * CPER record id, a 64bit number (in decimal notation).
- * In Linux, the record id is encoded with two digits for
- * the dmesg part (chunk) number and 3 digits for the
- * count number. So allowing an additional digit to
- * compensate for advancing time, this code ignores the
- * last six digits of the filename in determining the
- * record id.
- *
- * For the EFI backend, the record id encodes an id in the
- * upper 32 bits, and a timestamp in the lower 32-bits.
- * So ignoring the least significant 6 digits has proven
- * to generally identify related dmesg entries. */
-#define PSTORE_FILENAME_IGNORE 6
-
- /* determine common portion of record id */
- ++p; /* move beyond dmesg- */
- plen = strlen(p);
- if (plen > PSTORE_FILENAME_IGNORE) {
- pe_id = memdup_suffix0(p, plen - PSTORE_FILENAME_IGNORE);
- if (!pe_id) {
- log_oom();
- return;
- }
- } else
- pe_id = mfree(pe_id);
-
- /* Now move file from pstore to archive storage */
- move_file(pe, pe_id);
-
- if (dmesg_bad)
+ if (!startswith(pe->dirent.d_name, "dmesg-"))
continue;
- /* If the current record id is NOT the same as the
- * previous record id, then start a new dmesg.txt file */
- if (!streq_ptr(pe_id, dmesg_id)) {
- /* Encountered a new dmesg group, close out old one, open new one */
- (void) write_dmesg(dmesg, dmesg_size, dmesg_id);
- dmesg_size = 0;
-
- /* now point dmesg_id to storage of pe_id */
- free_and_replace(dmesg_id, pe_id);
- }
-
- /* Reconstruction of dmesg is done as a useful courtesy: do not fail, but don't write garbled
- * output either. */
- size_t needed = strlen(pe->dirent.d_name) + strlen(":\n") + pe->content_size + 1;
- if (!GREEDY_REALLOC(dmesg, dmesg_allocated, dmesg_size + needed)) {
- log_warning_errno(ENOMEM, "Failed to write dmesg file: %m");
- dmesg_bad = true;
- continue;
- }
-
- dmesg_size += sprintf(dmesg + dmesg_size, "%s:\n", pe->dirent.d_name);
- if (pe->content) {
- memcpy(dmesg + dmesg_size, pe->content, pe->content_size);
- dmesg_size += pe->content_size;
- }
-
- pe_id = mfree(pe_id);
+ if ((p = startswith(pe->dirent.d_name, "dmesg-efi-"))) {
+ /* For the EFI backend, the 3 least significant digits of record id encodes a
+ * "count" number, the next 2 least significant digits for the dmesg part
+ * (chunk) number, and the remaining digits as the timestamp. See
+ * linux/drivers/firmware/efi/efi-pstore.c in efi_pstore_write(). */
+ _cleanup_free_ char *subdir1 = NULL, *subdir2 = NULL;
+ size_t plen = strlen(p);
+
+ if (plen < 6)
+ continue;
+
+ /* Extract base record id */
+ subdir1 = strndup(p, plen - 5);
+ if (!subdir1)
+ return log_oom();
+ /* Extract "count" field */
+ subdir2 = strndup(p + plen - 3, 3);
+ if (!subdir2)
+ return log_oom();
+
+ /* Now move file from pstore to archive storage */
+ (void) move_file(pe, subdir1, subdir2);
+
+ /* Append to the dmesg */
+ (void) append_dmesg(pe, subdir1, subdir2);
+ } else if ((p = startswith(pe->dirent.d_name, "dmesg-erst-"))) {
+ /* For the ERST backend, the record is a monotonically increasing number, seeded as
+ * a timestamp. See linux/drivers/acpi/apei/erst.c in erst_writer(). */
+ uint64_t record_id;
+
+ if (safe_atou64(p, &record_id) < 0)
+ continue;
+ if (last_record_id - 1 != record_id)
+ /* A discontinuity in the number has been detected, this current record id
+ * will become the directory name for all pieces of the dmesg in this
+ * series. */
+ if (free_and_strdup(&erst_subdir, p) < 0)
+ return log_oom();
+
+ /* Now move file from pstore to archive storage */
+ (void) move_file(pe, erst_subdir, NULL);
+
+ /* Append to the dmesg */
+ (void) append_dmesg(pe, erst_subdir, NULL);
+
+ /* Update, but keep erst_subdir for next file */
+ last_record_id = record_id;
+ } else
+ log_debug("Unknown backend, ignoring \"%s\".", pe->dirent.d_name);
}
-
- if (!dmesg_bad)
- (void) write_dmesg(dmesg, dmesg_size, dmesg_id);
+ return 0;
}
static int list_files(PStoreList *list, const char *sourcepath) {
@@ -394,11 +362,11 @@ static int run(int argc, char *argv[]) {
qsort_safe(list.entries, list.n_entries, sizeof(PStoreEntry), compare_pstore_entries);
/* Process known file types */
- process_dmesg_files(&list);
+ (void) process_dmesg_files(&list);
/* Move left over files out of pstore */
for (size_t n = 0; n < list.n_entries; n++)
- move_file(&list.entries[n], NULL);
+ (void) move_file(&list.entries[n], NULL, NULL);
return 0;
}

27
SOURCES/0888-pstore-Don-t-start-systemd-pstore.service-in-contain.patch Normal file
View File

@ -0,0 +1,27 @@
From 653a635086cfeaf0af12da3a722b0ebe2029b927 Mon Sep 17 00:00:00 2001
From: Balint Reczey <balint.reczey@canonical.com>
Date: Mon, 16 Dec 2019 19:03:19 +0100
Subject: [PATCH] pstore: Don't start systemd-pstore.service in containers
Usually it is not useful and can also fail making
boot-and-services autopkgtest fail.
(cherry picked from commit 287f506c32f3f4a48ba020408f964cb0f964d752)
Related: #2158832
---
units/systemd-pstore.service.in | 1 +
1 file changed, 1 insertion(+)
diff --git a/units/systemd-pstore.service.in b/units/systemd-pstore.service.in
index dde21bc33e..89f34afe34 100644
--- a/units/systemd-pstore.service.in
+++ b/units/systemd-pstore.service.in
@@ -11,6 +11,7 @@
Description=Platform Persistent Storage Archival
Documentation=man:systemd-pstore(8)
ConditionDirectoryNotEmpty=/sys/fs/pstore
+ConditionVirtualization=!container
DefaultDependencies=no
Wants=systemd-remount-fs.service
After=systemd-remount-fs.service

36
SOURCES/0889-units-pull-in-systemd-pstore.service-from-sysinit.ta.patch Normal file
View File

@ -0,0 +1,36 @@
From c7e65774a4ccc8a431f63c5a12ab776b24ee1190 Mon Sep 17 00:00:00 2001
From: Lennart Poettering <lennart@poettering.net>
Date: Wed, 8 Apr 2020 16:12:00 +0200
Subject: [PATCH] units: pull in systemd-pstore.service from sysinit.target
sysinit.target is the target our early boot services are generally
pulled in from, make systemd-pstore.service not an exception of that.
Effectively this doesn't mean much, either way our unit is part of the
initial transaction.
(cherry picked from commit 167241912f51fbc0d7d0869b9af34c15b5ecc4b6)
Related: #2158832
---
units/systemd-pstore.service.in | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/units/systemd-pstore.service.in b/units/systemd-pstore.service.in
index 89f34afe34..37fcf878f0 100644
--- a/units/systemd-pstore.service.in
+++ b/units/systemd-pstore.service.in
@@ -15,6 +15,7 @@ ConditionVirtualization=!container
DefaultDependencies=no
Wants=systemd-remount-fs.service
After=systemd-remount-fs.service
+Before=sysinit.target
[Service]
Type=oneshot
@@ -23,4 +24,4 @@ RemainAfterExit=yes
StateDirectory=systemd/pstore
[Install]
-WantedBy=systemd-remount-fs.service
+WantedBy=sysinit.target

36
SOURCES/0890-units-drop-dependency-on-systemd-remount-fs.service-.patch Normal file
View File

@ -0,0 +1,36 @@
From bc6f273a0475a1fa7ab56bc1e498ee62c96aa660 Mon Sep 17 00:00:00 2001
From: Lennart Poettering <lennart@poettering.net>
Date: Wed, 8 Apr 2020 16:10:38 +0200
Subject: [PATCH] units: drop dependency on systemd-remount-fs.service from
systemd-pstore.service
This dependency is now generated automatically given we use
StateDirectory=. Moreover the combination of Wants= and After= was too
strong anway, as whether remount-fs is pulled in or not should not be up
to systemd-pstore.service, and in fact is part of the initial
transaction anyway.
[dtardon: This only removes Wants=, not After=, because I haven't
backported the auto-generation code the description talks about. The
code is simple, but it's just an optimisation allowing for slightly
shorter unit files, hence I don't think we really need it.]
(cherry picked from commit 0c978faa16fa9ecf92f0bbb5c7cc709dc472d115)
Related: #2158832
---
units/systemd-pstore.service.in | 1 -
1 file changed, 1 deletion(-)
diff --git a/units/systemd-pstore.service.in b/units/systemd-pstore.service.in
index 37fcf878f0..9a86f3145c 100644
--- a/units/systemd-pstore.service.in
+++ b/units/systemd-pstore.service.in
@@ -13,7 +13,6 @@ Documentation=man:systemd-pstore(8)
ConditionDirectoryNotEmpty=/sys/fs/pstore
ConditionVirtualization=!container
DefaultDependencies=no
-Wants=systemd-remount-fs.service
After=systemd-remount-fs.service
Before=sysinit.target

29
SOURCES/0891-units-make-sure-systemd-pstore-stops-at-shutdown.patch Normal file
View File

@ -0,0 +1,29 @@
From 818ddd1efd751ef50f9960920284465befe9d704 Mon Sep 17 00:00:00 2001
From: Lennart Poettering <lennart@poettering.net>
Date: Wed, 8 Apr 2020 16:25:03 +0200
Subject: [PATCH] units: make sure systemd-pstore stops at shutdown
This doesn't matter too much given that the service doesn't do anything
on shutdown, but let's still stop it to make things cleaner.
(cherry picked from commit b0c1a07654c80d3cbbbcc52f860d4206707c0b08)
Related: #2158832
---
units/systemd-pstore.service.in | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/units/systemd-pstore.service.in b/units/systemd-pstore.service.in
index 9a86f3145c..8cbf264a99 100644
--- a/units/systemd-pstore.service.in
+++ b/units/systemd-pstore.service.in
@@ -14,7 +14,8 @@ ConditionDirectoryNotEmpty=/sys/fs/pstore
ConditionVirtualization=!container
DefaultDependencies=no
After=systemd-remount-fs.service
-Before=sysinit.target
+Conflicts=shutdown.target
+Before=sysinit.target shutdown.target
[Service]
Type=oneshot

45
SOURCES/0892-pstore-Run-after-modules-are-loaded.patch Normal file
View File

@ -0,0 +1,45 @@
From 9cc6ee46e0083bc36b53d19e14fb637f7a1542dd Mon Sep 17 00:00:00 2001
From: Alexander Graf <graf@amazon.com>
Date: Thu, 9 Jun 2022 16:20:43 +0200
Subject: [PATCH] pstore: Run after modules are loaded
The systemd-pstore service takes pstore files on boot and transfers them
to disk. It only does it once on boot and only if it finds any. The typical
location of the pstore on modern systems is the UEFI variable store.
Most distributions ship with CONFIG_EFI_VARS_PSTORE=m. That means, the
UEFI variable store is only available on boot after the respective module
is loaded.
In most situations, the pstore service gets loaded before the UEFI pstore,
so we don't get to transfer logs. Instead, they accumulate, filling up the
pstore over time, potentially breaking the UEFI variable store.
Let's add a service dependency on any kernel module that can provide a
pstore to ensure we only scan for pstate after we can actually see pstate.
I have seen live occurences of systems breaking because we did not erase
the pstates and ran out of UEFI nvram space.
Fixes https://github.com/systemd/systemd/issues/18540
(cherry picked from commit 70e74a5997ae2ce7ba72a74ac949c3b2dad1a1d6)
Related: #2158832
---
units/systemd-pstore.service.in | 2 ++
1 file changed, 2 insertions(+)
diff --git a/units/systemd-pstore.service.in b/units/systemd-pstore.service.in
index 8cbf264a99..1983a9738f 100644
--- a/units/systemd-pstore.service.in
+++ b/units/systemd-pstore.service.in
@@ -16,6 +16,8 @@ DefaultDependencies=no
After=systemd-remount-fs.service
Conflicts=shutdown.target
Before=sysinit.target shutdown.target
+After=modprobe@efi_pstore.service modprobe@mtdpstore.service modprobe@chromeos_pstore.service modprobe@ramoops.service modprobe@pstore_zone.service modprobe@pstore_blk.service
+Wants=modprobe@efi_pstore.service modprobe@mtdpstore.service modprobe@chromeos_pstore.service modprobe@ramoops.service modprobe@pstore_zone.service modprobe@pstore_blk.service
[Service]
Type=oneshot

48
SOURCES/0893-pstore-do-not-try-to-load-all-known-pstore-modules.patch Normal file
View File

@ -0,0 +1,48 @@
From 6a6f108b59e47581d93cbc6bdc604ee84f1bb791 Mon Sep 17 00:00:00 2001
From: Nick Rosbrook <nick.rosbrook@canonical.com>
Date: Wed, 7 Sep 2022 13:25:13 -0400
Subject: [PATCH] pstore: do not try to load all known pstore modules
Commit 70e74a5997 ("pstore: Run after modules are loaded") added After=
and Wants= entries for all known kernel modules providing a pstore.
While adding these dependencies on systems where one of the modules is
not present, or not configured, should not have a real affect on the
system, it can produce annoying error messages in the kernel log. E.g.
"mtd device must be supplied (device name is empty)" when the mtdpstore
module is not configured correctly.
Since dependencies cannot be removed with drop-ins, if a distro wants to
remove some of these modules from systemd-pstore.service, they need to
patch units/systemd-pstore.service.in. On the other hand, if they want
to append to the dependencies this can be done by shipping a drop-in.
Since the original intent of the previous commit was to fix [1], which
only requires the efi_pstore module, remove all other kernel module
dependencies from systemd-pstore.service, and let distros ship drop-ins
to add dependencies if needed.
[1] https://github.com/systemd/systemd/issues/18540
(cherry picked from commit 8b8bd621e1d16808678fc3afed257df1fa03a281)
Related: #2158832
---
units/systemd-pstore.service.in | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/units/systemd-pstore.service.in b/units/systemd-pstore.service.in
index 1983a9738f..19ffa8d4e8 100644
--- a/units/systemd-pstore.service.in
+++ b/units/systemd-pstore.service.in
@@ -16,8 +16,8 @@ DefaultDependencies=no
After=systemd-remount-fs.service
Conflicts=shutdown.target
Before=sysinit.target shutdown.target
-After=modprobe@efi_pstore.service modprobe@mtdpstore.service modprobe@chromeos_pstore.service modprobe@ramoops.service modprobe@pstore_zone.service modprobe@pstore_blk.service
-Wants=modprobe@efi_pstore.service modprobe@mtdpstore.service modprobe@chromeos_pstore.service modprobe@ramoops.service modprobe@pstore_zone.service modprobe@pstore_blk.service
+After=modprobe@efi_pstore.service
+Wants=modprobe@efi_pstore.service
[Service]
Type=oneshot

26
SOURCES/0894-logind-session-make-stopping-of-idle-session-visible.patch Normal file
View File

@ -0,0 +1,26 @@
From b18e19f2262e7ed95c25d53268d12427fe77102d Mon Sep 17 00:00:00 2001
From: David Tardon <dtardon@redhat.com>
Date: Tue, 21 Feb 2023 10:41:47 +0100
Subject: [PATCH] logind-session: make stopping of idle session visible to
admins
(cherry picked from commit 6269ffe7ee8a659df7336a2582054ecd9eecf4b1)
Resolves: #2156780
---
src/login/logind-session.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/login/logind-session.c b/src/login/logind-session.c
index 18a07efcdb..916202a65a 100644
--- a/src/login/logind-session.c
+++ b/src/login/logind-session.c
@@ -673,7 +673,7 @@ static int session_dispatch_stop_on_idle(sd_event_source *source, uint64_t t, vo
idle = session_get_idle_hint(s, &ts);
if (idle) {
- log_debug("Session \"%s\" of user \"%s\" is idle, stopping.", s->id, s->user->name);
+ log_info("Session \"%s\" of user \"%s\" is idle, stopping.", s->id, s->user->name);
return session_stop(s, /* force */ true);
}

32
SOURCES/0895-journald-Increase-stdout-buffer-size-sooner-when-alm.patch Normal file
View File

@ -0,0 +1,32 @@
From a0b52398692f3e4bda18520db9e2397f7b2c80dd Mon Sep 17 00:00:00 2001
From: Benjamin Robin <dev@benjarobin.fr>
Date: Sun, 3 May 2020 18:37:21 +0200
Subject: [PATCH] journald: Increase stdout buffer size sooner, when almost
full
If the previous received buffer length is almost equal to the allocated
buffer size, before this change the next read can only receive a couple
of bytes (in the worst case only 1 byte), which is not efficient.
(cherry picked from commit 034e9719ac1ba88a36b05da38c7aa98761d42c77)
Related: #2029426
---
src/journal/journald-stream.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/journal/journald-stream.c b/src/journal/journald-stream.c
index 302a82d3d7..c8de984335 100644
--- a/src/journal/journald-stream.c
+++ b/src/journal/journald-stream.c
@@ -507,8 +507,8 @@ static int stdout_stream_process(sd_event_source *es, int fd, uint32_t revents,
goto terminate;
}
- /* If the buffer is full already (discounting the extra NUL we need), add room for another 1K */
- if (s->length + 1 >= s->allocated) {
+ /* If the buffer is almost full, add room for another 1K */
+ if (s->length + 512 >= s->allocated) {
if (!GREEDY_REALLOC(s->buffer, s->allocated, s->length + 1 + 1024)) {
log_oom();
goto terminate;

77
SOURCES/0896-journald-rework-end-of-line-marker-handling-to-use-a.patch Normal file
View File

@ -0,0 +1,77 @@
From d8fabe7a6839eeb0d5d0504471f2d18b07545238 Mon Sep 17 00:00:00 2001
From: Lennart Poettering <lennart@poettering.net>
Date: Tue, 12 May 2020 18:53:35 +0200
Subject: [PATCH] journald: rework end of line marker handling to use a field
table
(cherry picked from commit 549b7379ba404c33fd448d2bca46a57f6529b00b)
Related: #2029426
---
src/journal/journald-stream.c | 29 ++++++++++++++++++++---------
1 file changed, 20 insertions(+), 9 deletions(-)
diff --git a/src/journal/journald-stream.c b/src/journal/journald-stream.c
index c8de984335..58752a5a24 100644
--- a/src/journal/journald-stream.c
+++ b/src/journal/journald-stream.c
@@ -54,6 +54,8 @@ typedef enum LineBreak {
LINE_BREAK_NUL,
LINE_BREAK_LINE_MAX,
LINE_BREAK_EOF,
+ _LINE_BREAK_MAX,
+ _LINE_BREAK_INVALID = -1,
} LineBreak;
struct StdoutStream {
@@ -233,7 +235,11 @@ fail:
return log_error_errno(r, "Failed to save stream data %s: %m", s->state_file);
}
-static int stdout_stream_log(StdoutStream *s, const char *p, LineBreak line_break) {
+static int stdout_stream_log(
+ StdoutStream *s,
+ const char *p,
+ LineBreak line_break) {
+
struct iovec *iovec;
int priority;
char syslog_priority[] = "PRIORITY=\0";
@@ -245,6 +251,9 @@ static int stdout_stream_log(StdoutStream *s, const char *p, LineBreak line_brea
assert(s);
assert(p);
+ assert(line_break >= 0);
+ assert(line_break < _LINE_BREAK_MAX);
+
if (s->context)
(void) client_context_maybe_refresh(s->server, s->context, NULL, NULL, 0, NULL, USEC_INFINITY);
else if (pid_is_valid(s->ucred.pid)) {
@@ -296,17 +305,19 @@ static int stdout_stream_log(StdoutStream *s, const char *p, LineBreak line_brea
iovec[n++] = IOVEC_MAKE_STRING(syslog_identifier);
}
- if (line_break != LINE_BREAK_NEWLINE) {
- const char *c;
+ static const char * const line_break_field_table[_LINE_BREAK_MAX] = {
+ [LINE_BREAK_NEWLINE] = NULL, /* Do not add field if traditional newline */
+ [LINE_BREAK_NUL] = "_LINE_BREAK=nul",
+ [LINE_BREAK_LINE_MAX] = "_LINE_BREAK=line-max",
+ [LINE_BREAK_EOF] = "_LINE_BREAK=eof",
+ };
- /* If this log message was generated due to an uncommon line break then mention this in the log
- * entry */
+ const char *c = line_break_field_table[line_break];
- c = line_break == LINE_BREAK_NUL ? "_LINE_BREAK=nul" :
- line_break == LINE_BREAK_LINE_MAX ? "_LINE_BREAK=line-max" :
- "_LINE_BREAK=eof";
+ /* If this log message was generated due to an uncommon line break then mention this in the log
+ * entry */
+ if (c)
iovec[n++] = IOVEC_MAKE_STRING(c);
- }
message = strappend("MESSAGE=", p);
if (message)

27
SOURCES/0897-journald-use-the-fact-that-client_context_release-re.patch Normal file
View File

@ -0,0 +1,27 @@
From cd85a657c932725ac7c1b506dc6dd4270d1dc068 Mon Sep 17 00:00:00 2001
From: Lennart Poettering <lennart@poettering.net>
Date: Tue, 12 May 2020 19:15:38 +0200
Subject: [PATCH] journald: use the fact that client_context_release() returns
NULL
(cherry picked from commit 020b4a023c2c6dda83afb9a82a62e640569c40c1)
Related: #2029426
---
src/journal/journald-stream.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/src/journal/journald-stream.c b/src/journal/journald-stream.c
index 58752a5a24..ab1a855943 100644
--- a/src/journal/journald-stream.c
+++ b/src/journal/journald-stream.c
@@ -570,8 +570,7 @@ static int stdout_stream_process(sd_event_source *es, int fd, uint32_t revents,
goto terminate;
s->ucred = *ucred;
- client_context_release(s->server, s->context);
- s->context = NULL;
+ s->context = client_context_release(s->server, s->context);
}
s->length += l;

219
SOURCES/0898-journald-rework-pid-change-handling.patch Normal file
View File

@ -0,0 +1,219 @@
From 538bd9b42dabf1145cf7ab77f32347d516b01e88 Mon Sep 17 00:00:00 2001
From: Lennart Poettering <lennart@poettering.net>
Date: Tue, 12 May 2020 18:56:34 +0200
Subject: [PATCH] journald: rework pid change handling
Let's introduce an explicit line ending marker for line endings due to
pid change.
Let's also make sure we don't get confused with buffer management.
Fixes: #15654
(cherry picked from commit 45ba1ea5e9264d385fa565328fe957ef1d78caa1)
Resolves: #2029426
---
src/journal/journald-stream.c | 99 +++++++++++++++++++++++------------
1 file changed, 66 insertions(+), 33 deletions(-)
diff --git a/src/journal/journald-stream.c b/src/journal/journald-stream.c
index ab1a855943..5be5b0939c 100644
--- a/src/journal/journald-stream.c
+++ b/src/journal/journald-stream.c
@@ -54,6 +54,7 @@ typedef enum LineBreak {
LINE_BREAK_NUL,
LINE_BREAK_LINE_MAX,
LINE_BREAK_EOF,
+ LINE_BREAK_PID_CHANGE,
_LINE_BREAK_MAX,
_LINE_BREAK_INVALID = -1,
} LineBreak;
@@ -310,6 +311,7 @@ static int stdout_stream_log(
[LINE_BREAK_NUL] = "_LINE_BREAK=nul",
[LINE_BREAK_LINE_MAX] = "_LINE_BREAK=line-max",
[LINE_BREAK_EOF] = "_LINE_BREAK=eof",
+ [LINE_BREAK_PID_CHANGE] = "_LINE_BREAK=pid-change",
};
const char *c = line_break_field_table[line_break];
@@ -431,21 +433,43 @@ static int stdout_stream_line(StdoutStream *s, char *p, LineBreak line_break) {
assert_not_reached("Unknown stream state");
}
-static int stdout_stream_scan(StdoutStream *s, bool force_flush) {
- char *p;
- size_t remaining;
+static int stdout_stream_found(
+ StdoutStream *s,
+ char *p,
+ size_t l,
+ LineBreak line_break) {
+
+ char saved;
int r;
assert(s);
+ assert(p);
+
+ /* Let's NUL terminate the specified buffer for this call, and revert back afterwards */
+ saved = p[l];
+ p[l] = 0;
+ r = stdout_stream_line(s, p, line_break);
+ p[l] = saved;
- p = s->buffer;
- remaining = s->length;
+ return r;
+}
+
+static int stdout_stream_scan(
+ StdoutStream *s,
+ char *p,
+ size_t remaining,
+ LineBreak force_flush,
+ size_t *ret_consumed) {
- /* XXX: This function does nothing if (s->length == 0) */
+ size_t consumed = 0;
+ int r;
+
+ assert(s);
+ assert(p);
for (;;) {
LineBreak line_break;
- size_t skip;
+ size_t skip, found;
char *end1, *end2;
end1 = memchr(p, '\n', remaining);
@@ -453,43 +477,40 @@ static int stdout_stream_scan(StdoutStream *s, bool force_flush) {
if (end2) {
/* We found a NUL terminator */
- skip = end2 - p + 1;
+ found = end2 - p;
+ skip = found + 1;
line_break = LINE_BREAK_NUL;
} else if (end1) {
/* We found a \n terminator */
- *end1 = 0;
- skip = end1 - p + 1;
+ found = end1 - p;
+ skip = found + 1;
line_break = LINE_BREAK_NEWLINE;
} else if (remaining >= s->server->line_max) {
/* Force a line break after the maximum line length */
- *(p + s->server->line_max) = 0;
- skip = remaining;
+ found = skip = s->server->line_max;
line_break = LINE_BREAK_LINE_MAX;
} else
break;
- r = stdout_stream_line(s, p, line_break);
+ r = stdout_stream_found(s, p, found, line_break);
if (r < 0)
return r;
- remaining -= skip;
p += skip;
+ consumed += skip;
+ remaining -= skip;
}
- if (force_flush && remaining > 0) {
- p[remaining] = 0;
- r = stdout_stream_line(s, p, LINE_BREAK_EOF);
+ if (force_flush >= 0 && remaining > 0) {
+ r = stdout_stream_found(s, p, remaining, force_flush);
if (r < 0)
return r;
- p += remaining;
- remaining = 0;
+ consumed += remaining;
}
- if (p > s->buffer) {
- memmove(s->buffer, p, remaining);
- s->length = remaining;
- }
+ if (ret_consumed)
+ *ret_consumed = consumed;
return 0;
}
@@ -497,11 +518,12 @@ static int stdout_stream_scan(StdoutStream *s, bool force_flush) {
static int stdout_stream_process(sd_event_source *es, int fd, uint32_t revents, void *userdata) {
uint8_t buf[CMSG_SPACE(sizeof(struct ucred))];
StdoutStream *s = userdata;
+ size_t limit, consumed;
struct ucred *ucred = NULL;
struct cmsghdr *cmsg;
struct iovec iovec;
- size_t limit;
ssize_t l;
+ char *p;
int r;
struct msghdr msghdr = {
@@ -529,7 +551,7 @@ static int stdout_stream_process(sd_event_source *es, int fd, uint32_t revents,
/* Try to make use of the allocated buffer in full, but never read more than the configured line size. Also,
* always leave room for a terminating NUL we might need to add. */
limit = MIN(s->allocated - 1, s->server->line_max);
-
+ assert(s->length <= limit);
iovec = IOVEC_MAKE(s->buffer + s->length, limit - s->length);
l = recvmsg(s->fd, &msghdr, MSG_DONTWAIT|MSG_CMSG_CLOEXEC);
@@ -543,7 +565,7 @@ static int stdout_stream_process(sd_event_source *es, int fd, uint32_t revents,
cmsg_close_all(&msghdr);
if (l == 0) {
- stdout_stream_scan(s, true);
+ (void) stdout_stream_scan(s, s->buffer, s->length, /* force_flush = */ LINE_BREAK_EOF, NULL);
goto terminate;
}
@@ -562,22 +584,33 @@ static int stdout_stream_process(sd_event_source *es, int fd, uint32_t revents,
* in the meantime.
*/
if (ucred && ucred->pid != s->ucred.pid) {
- /* force out any previously half-written lines from a
- * different process, before we switch to the new ucred
- * structure for everything we just added */
- r = stdout_stream_scan(s, true);
+ /* Force out any previously half-written lines from a different process, before we switch to
+ * the new ucred structure for everything we just added */
+ r = stdout_stream_scan(s, s->buffer, s->length, /* force_flush = */ LINE_BREAK_PID_CHANGE, NULL);
if (r < 0)
goto terminate;
- s->ucred = *ucred;
s->context = client_context_release(s->server, s->context);
+
+ p = s->buffer + s->length;
+ } else {
+ p = s->buffer;
+ l += s->length;
}
- s->length += l;
- r = stdout_stream_scan(s, false);
+ /* Always copy in the new credentials */
+ if (ucred)
+ s->ucred = *ucred;
+
+ r = stdout_stream_scan(s, p, l, _LINE_BREAK_INVALID, &consumed);
if (r < 0)
goto terminate;
+ /* Move what wasn't consumed to the front of the buffer */
+ assert(consumed <= (size_t) l);
+ s->length = l - consumed;
+ memmove(s->buffer, p + consumed, s->length);
+
return 1;
terminate:

34
SOURCES/0899-test-Add-a-test-case-for-15654.patch Normal file
View File

@ -0,0 +1,34 @@
From e019afeefb396ea42d03f4c3f9713e262aff6450 Mon Sep 17 00:00:00 2001
From: Benjamin Robin <dev@benjarobin.fr>
Date: Wed, 6 May 2020 23:28:02 +0200
Subject: [PATCH] test: Add a test case for #15654
(cherry picked from commit c11d8fd1dab3bc3f0abbc861ba5eb34518cec1da)
Related: #2029426
---
test/TEST-04-JOURNAL/test-journal.sh | 11 +++++++++++
1 file changed, 11 insertions(+)
diff --git a/test/TEST-04-JOURNAL/test-journal.sh b/test/TEST-04-JOURNAL/test-journal.sh
index a3db1a7472..bdf137cd69 100755
--- a/test/TEST-04-JOURNAL/test-journal.sh
+++ b/test/TEST-04-JOURNAL/test-journal.sh
@@ -76,6 +76,17 @@ journalctl -b -o export -t "$ID" --output-fields=_PID | grep '^_PID=' >/output
grep -q "^_PID=$PID" /output
grep -vq "^_PID=$PID" /output
+# https://github.com/systemd/systemd/issues/15654
+ID=$(journalctl --new-id128 | sed -n 2p)
+printf "This will\nusually fail\nand be truncated\n">/expected
+systemd-cat -t "$ID" /bin/sh -c 'env echo -n "This will";echo;env echo -n "usually fail";echo;env echo -n "and be truncated";echo;'
+journalctl --sync
+journalctl -b -o cat -t "$ID" >/output
+cmp /expected /output
+
+# Add new tests before here, the journald restarts below
+# may make tests flappy.
+
# Don't lose streams on restart
systemctl start forever-print-hola
sleep 3

32
SOURCES/0900-test-Stricter-test-case-for-15654-Add-more-checks.patch Normal file
View File

@ -0,0 +1,32 @@
From afcfb65ce7514bf32e59e0b9d212ae18d023a4b5 Mon Sep 17 00:00:00 2001
From: Benjamin Robin <dev@benjarobin.fr>
Date: Sat, 9 May 2020 12:01:07 +0200
Subject: [PATCH] test: Stricter test case for #15654 (Add more checks)
Check:
- There is only 3 messages logged with type stdout
- Check all messages logged does not have new line: LINE_BREAK=eof
- Check that the 3 messages are logged from a different PID
- Check the 3 MESSAGE= content
(cherry picked from commit d38b3b74dbde2d65b23e5963354c3db96acd3420)
Related: #2029426
---
test/TEST-04-JOURNAL/test-journal.sh | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/test/TEST-04-JOURNAL/test-journal.sh b/test/TEST-04-JOURNAL/test-journal.sh
index bdf137cd69..641259ce24 100755
--- a/test/TEST-04-JOURNAL/test-journal.sh
+++ b/test/TEST-04-JOURNAL/test-journal.sh
@@ -83,6 +83,10 @@ systemd-cat -t "$ID" /bin/sh -c 'env echo -n "This will";echo;env echo -n "usual
journalctl --sync
journalctl -b -o cat -t "$ID" >/output
cmp /expected /output
+[[ $(journalctl -b -o export -t "$ID" --output-fields=_TRANSPORT | grep -Pc "^_TRANSPORT=stdout$") -eq 3 ]]
+[[ $(journalctl -b -o export -t "$ID" --output-fields=_LINE_BREAK | grep -Pc "^_LINE_BREAK=pid-change$") -eq 3 ]]
+[[ $(journalctl -b -o export -t "$ID" --output-fields=_PID | sort -u | grep -c "^_PID=.*$") -eq 3 ]]
+[[ $(journalctl -b -o export -t "$ID" --output-fields=MESSAGE | grep -Pc "^MESSAGE=(This will|usually fail|and be truncated)$") -eq 3 ]]
# Add new tests before here, the journald restarts below
# may make tests flappy.

42
SOURCES/0901-man-document-the-new-_LINE_BREAK-type.patch Normal file
View File

@ -0,0 +1,42 @@
From 2f55aeadef3dcdf65c61f24a41178148c3544ac3 Mon Sep 17 00:00:00 2001
From: Lennart Poettering <lennart@poettering.net>
Date: Wed, 13 May 2020 00:09:43 +0200
Subject: [PATCH] man: document the new _LINE_BREAK= type
(cherry picked from commit a3d9aee14fa2f7df429dc401582877176206b7fd)
Related: #2029426
---
man/systemd.journal-fields.xml | 19 ++++++++++---------
1 file changed, 10 insertions(+), 9 deletions(-)
diff --git a/man/systemd.journal-fields.xml b/man/systemd.journal-fields.xml
index 0c95c4cd95..ad2b94dbd5 100644
--- a/man/systemd.journal-fields.xml
+++ b/man/systemd.journal-fields.xml
@@ -326,15 +326,16 @@
<varlistentry>
<term><varname>_LINE_BREAK=</varname></term>
<listitem>
- <para>Only applies to <literal>_TRANSPORT=stdout</literal> records: indicates that the log message in the
- standard output/error stream was not terminated with a normal newline character (<literal>\n</literal>,
- i.e. ASCII 10). Specifically, when set this field is one of <option>nul</option> (in case the line was
- terminated by a NUL byte), <option>line-max</option> (in case the maximum log line length was reached, as
- configured with <varname>LineMax=</varname> in
- <citerefentry><refentrytitle>journald.conf</refentrytitle><manvolnum>5</manvolnum></citerefentry>) or
- <option>eof</option> (if this was the last log record of a stream and the stream ended without a final
- newline character). Note that this record is not generated when a normal newline character was used for
- marking the log line end.</para>
+ <para>Only applies to <literal>_TRANSPORT=stdout</literal> records: indicates that the log message
+ in the standard output/error stream was not terminated with a normal newline character
+ (<literal>\n</literal>, i.e. ASCII 10). Specifically, when set this field is one of
+ <option>nul</option> (in case the line was terminated by a NUL byte), <option>line-max</option> (in
+ case the maximum log line length was reached, as configured with <varname>LineMax=</varname> in
+ <citerefentry><refentrytitle>journald.conf</refentrytitle><manvolnum>5</manvolnum></citerefentry>),
+ <option>eof</option> (if this was the last log record of a stream and the stream ended without a
+ final newline character), or <option>pid-change</option> (if the process which generated the log
+ output changed in the middle of a line). Note that this record is not generated when a normal
+ newline character was used for marking the log line end.</para>
</listitem>
</varlistentry>
</variablelist>

38
SOURCES/0902-journald-server-always-create-state-file-in-signal-h.patch Normal file
View File

@ -0,0 +1,38 @@
From 47cc8f7e1d153e576f146d309b4043739997a673 Mon Sep 17 00:00:00 2001
From: David Tardon <dtardon@redhat.com>
Date: Mon, 13 Mar 2023 14:22:28 +0100
Subject: [PATCH] journald-server: always create state file in signal handler
`journalctl --flush` waits on that file, so we must create if even if
nothing has really happened.
RHEL-only
Resolves: #2174645
---
src/journal/journald-server.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/src/journal/journald-server.c b/src/journal/journald-server.c
index 279a32768c..c72cb68095 100644
--- a/src/journal/journald-server.c
+++ b/src/journal/journald-server.c
@@ -1188,6 +1188,7 @@ int server_process_datagram(sd_event_source *es, int fd, uint32_t revents, void
static int dispatch_sigusr1(sd_event_source *es, const struct signalfd_siginfo *si, void *userdata) {
Server *s = userdata;
+ int r;
assert(s);
@@ -1197,6 +1198,10 @@ static int dispatch_sigusr1(sd_event_source *es, const struct signalfd_siginfo *
server_sync(s);
server_vacuum(s, false);
+ r = touch("/run/systemd/journal/flushed");
+ if (r < 0)
+ log_warning_errno(r, "Failed to touch /run/systemd/journal/flushed, ignoring: %m");
+
server_space_usage_message(s, NULL);
return 0;
}

62
SOURCES/0903-journald-server-move-relinquish-code-into-function.patch Normal file
View File

@ -0,0 +1,62 @@
From 7a7b0c4ec7a5595a44d9c70d8270b0724a8b8c45 Mon Sep 17 00:00:00 2001
From: David Tardon <dtardon@redhat.com>
Date: Mon, 13 Mar 2023 14:31:38 +0100
Subject: [PATCH] journald-server: move relinquish code into function
No functional change, just refactoring.
RHEL-only
Related: #2174645
---
src/journal/journald-server.c | 23 ++++++++++++++++-------
1 file changed, 16 insertions(+), 7 deletions(-)
diff --git a/src/journal/journald-server.c b/src/journal/journald-server.c
index c72cb68095..aa70db95cc 100644
--- a/src/journal/journald-server.c
+++ b/src/journal/journald-server.c
@@ -1258,20 +1258,16 @@ static int dispatch_sigrtmin1(sd_event_source *es, const struct signalfd_siginfo
return 0;
}
-
-static int dispatch_sigrtmin2(sd_event_source *es, const struct signalfd_siginfo *si, void *userdata) {
- Server *s = userdata;
+static void relinquish_var(Server *s) {
int r;
assert(s);
if (s->storage == STORAGE_NONE)
- return 0;
+ return;
if (s->runtime_journal && !s->system_journal)
- return 0;
-
- log_debug("Received request to relinquish /var from PID " PID_FMT, si->ssi_pid);
+ return;
(void) system_journal_open(s, false, true);
@@ -1286,6 +1282,19 @@ static int dispatch_sigrtmin2(sd_event_source *es, const struct signalfd_siginfo
if (r < 0)
log_warning_errno(r, "Failed to write /run/systemd/journal/relinquished, ignoring: %m");
+ return;
+}
+
+static int dispatch_sigrtmin2(sd_event_source *es, const struct signalfd_siginfo *si, void *userdata) {
+ Server *s = userdata;
+
+ assert(s);
+ assert(si);
+
+ log_debug("Received request to relinquish /var from PID " PID_FMT, si->ssi_pid);
+
+ relinquish_var(s);
+
return 0;
}

58
SOURCES/0904-journald-server-always-touch-state-file-in-signal-ha.patch Normal file
View File

@ -0,0 +1,58 @@
From 980add7d84084a474c6c604c0670743c2d1e624c Mon Sep 17 00:00:00 2001
From: David Tardon <dtardon@redhat.com>
Date: Mon, 13 Mar 2023 14:32:20 +0100
Subject: [PATCH] journald-server: always touch state file in signal handler
`journalctl --relinquish-var` waits on that file, so we must create if
even if nothing has really happened.
RHEL-only
Related: #2174645
---
src/journal/journald-server.c | 12 +++++++-----
1 file changed, 7 insertions(+), 5 deletions(-)
diff --git a/src/journal/journald-server.c b/src/journal/journald-server.c
index aa70db95cc..4788ff78bb 100644
--- a/src/journal/journald-server.c
+++ b/src/journal/journald-server.c
@@ -1259,8 +1259,6 @@ static int dispatch_sigrtmin1(sd_event_source *es, const struct signalfd_siginfo
}
static void relinquish_var(Server *s) {
- int r;
-
assert(s);
if (s->storage == STORAGE_NONE)
@@ -1278,15 +1276,15 @@ static void relinquish_var(Server *s) {
if (unlink("/run/systemd/journal/flushed") < 0 && errno != ENOENT)
log_warning_errno(errno, "Failed to unlink /run/systemd/journal/flushed, ignoring: %m") ;
- r = write_timestamp_file_atomic("/run/systemd/journal/relinquished", now(CLOCK_MONOTONIC));
- if (r < 0)
- log_warning_errno(r, "Failed to write /run/systemd/journal/relinquished, ignoring: %m");
+ /* NOTE: We don't create our own state file here, because dispatch_sigrtmin2() has to do it anyway.
+ * But if this function is ever called from another place, the creation must be done here too. */
return;
}
static int dispatch_sigrtmin2(sd_event_source *es, const struct signalfd_siginfo *si, void *userdata) {
Server *s = userdata;
+ int r;
assert(s);
assert(si);
@@ -1295,6 +1293,10 @@ static int dispatch_sigrtmin2(sd_event_source *es, const struct signalfd_siginfo
relinquish_var(s);
+ r = write_timestamp_file_atomic("/run/systemd/journal/relinquished", now(CLOCK_MONOTONIC));
+ if (r < 0)
+ log_warning_errno(r, "Failed to write /run/systemd/journal/relinquished, ignoring: %m");
+
return 0;
}

1
SOURCES/systemd-user
View File

@ -8,4 +8,5 @@ account include system-auth
session required pam_selinux.so close session required pam_selinux.so close
session required pam_selinux.so nottys open session required pam_selinux.so nottys open
session required pam_loginuid.so session required pam_loginuid.so
session required pam_namespace.so
session include system-auth session include system-auth

177
SPECS/systemd.spec
View File

@ -13,7 +13,7 @@
Name: systemd Name: systemd
Url: http://www.freedesktop.org/wiki/Software/systemd Url: http://www.freedesktop.org/wiki/Software/systemd
Version: 239 Version: 239
Release: 68%{?dist}.4 Release: 74%{?dist}
# For a breakdown of the licensing, see README # For a breakdown of the licensing, see README
License: LGPLv2+ and MIT and GPLv2+ License: LGPLv2+ and MIT and GPLv2+
Summary: System and Service Manager Summary: System and Service Manager
@ -881,13 +881,79 @@ Patch0828: 0828-core-load-fragment-move-config_parse_sec_fix_0-to-sr.patch
Patch0829: 0829-sd-event-add-relative-timer-calls.patch Patch0829: 0829-sd-event-add-relative-timer-calls.patch
Patch0830: 0830-logind-add-option-to-stop-idle-sessions-after-specif.patch Patch0830: 0830-logind-add-option-to-stop-idle-sessions-after-specif.patch
Patch0831: 0831-logind-schedule-idle-check-full-interval-from-now-if.patch Patch0831: 0831-logind-schedule-idle-check-full-interval-from-now-if.patch
Patch0832: 0832-time-util-fix-buffer-over-run.patch Patch0832: 0832-ci-lint-add-shell-linter-Differential-ShellCheck.patch
Patch0833: 0833-core-move-reset_arguments-to-the-end-of-main-s-finis.patch Patch0833: 0833-meson-do-not-compare-objects-of-different-types.patch
Patch0834: 0834-basic-recognize-pdfs-filesystem-as-a-network-filesys.patch Patch0834: 0834-journal-remote-use-MHD_HTTP_CONTENT_TOO_LARGE-as-MHD.patch
Patch0835: 0835-core-bring-manager_startup-and-manager_reload-more-i.patch Patch0835: 0835-Fix-build-with-httpd-0.9.71.patch
Patch0836: 0836-basic-add-STRERROR-wrapper-for-strerror_r.patch Patch0836: 0836-ci-replace-LGTM-with-CodeQL.patch
Patch0837: 0837-coredump-put-context-array-into-a-struct.patch Patch0837: 0837-ci-mergify-Update-policy-Drop-LGTM-checks.patch
Patch0838: 0838-coredump-do-not-allow-user-to-access-coredumps-with-.patch Patch0838: 0838-time-util-fix-buffer-over-run.patch
Patch0839: 0839-basic-recognize-pdfs-filesystem-as-a-network-filesys.patch
Patch0840: 0840-core-move-reset_arguments-to-the-end-of-main-s-finis.patch
Patch0841: 0841-manager-move-inc.-of-n_reloading-into-a-function.patch
Patch0842: 0842-core-Add-new-DBUS-properties-UnitsReloadStartTimesta.patch
Patch0843: 0843-core-Indicate-the-time-when-the-manager-started-load.patch
Patch0844: 0844-core-do-not-touch-run-systemd-systemd-units-load-fro.patch
Patch0845: 0845-sysctl-downgrade-message-when-we-have-no-permission.patch
Patch0846: 0846-core-respect-SELinuxContext-for-socket-creation.patch
Patch0847: 0847-manager-use-target-process-context-to-set-socket-con.patch
Patch0848: 0848-virt-detect-Amazon-EC2-Nitro-instance.patch
Patch0849: 0849-machine-id-setup-generate-machine-id-from-DMI-produc.patch
Patch0850: 0850-virt-use-string-table-to-detect-VM-or-container.patch
Patch0851: 0851-fileio-introduce-read_full_virtual_file-for-reading-.patch
Patch0852: 0852-Use-BIOS-characteristics-to-distinguish-EC2-bare-met.patch
Patch0853: 0853-device-drop-refuse_after.patch
Patch0854: 0854-manager-limit-access-to-private-dbus-socket.patch
Patch0855: 0855-journalctl-do-not-treat-EINTR-as-an-error-when-waiti.patch
Patch0856: 0856-core-bring-manager_startup-and-manager_reload-more-i.patch
Patch0857: 0857-pam-add-a-call-to-pam_namespace.patch
Patch0858: 0858-virt-Support-detection-for-ARM64-Hyper-V-guests.patch
Patch0859: 0859-virt-Fix-the-detection-for-Hyper-V-VMs.patch
Patch0860: 0860-basic-add-STRERROR-wrapper-for-strerror_r.patch
Patch0861: 0861-coredump-put-context-array-into-a-struct.patch
Patch0862: 0862-coredump-do-not-allow-user-to-access-coredumps-with-.patch
Patch0863: 0863-logind-remember-our-idle-state-and-use-it-to-detect-.patch
Patch0864: 0864-test-import-logind-test-from-debian-ubuntu-test-suit.patch
Patch0865: 0865-test-introduce-inst_recursive-helper-function.patch
Patch0866: 0866-tests-verify-that-Lock-D-Bus-signal-is-sent-when-Idl.patch
Patch0867: 0867-systemctl-simplify-halt_main.patch
Patch0868: 0868-systemctl-shutdown-don-t-fallback-on-auth-fail.patch
Patch0869: 0869-systemctl-reintroduce-the-original-halt_main.patch
Patch0870: 0870-systemctl-preserve-old-behavior-unless-requested.patch
Patch0871: 0871-pam_systemd-suppress-LOG_DEBUG-log-messages-if-debug.patch
Patch0872: 0872-udev-net_id-introduce-naming-scheme-for-RHEL-8.8.patch
Patch0873: 0873-journald-add-API-to-move-logging-from-var-to-run-aga.patch
Patch0874: 0874-journalctl-add-new-relinquish-and-smart-relinquish-o.patch
Patch0875: 0875-units-automatically-revert-to-run-logging-on-shutdow.patch
Patch0876: 0876-pstore-Tool-to-archive-contents-of-pstore.patch
Patch0877: 0877-meson-drop-redundant-line.patch
Patch0878: 0878-pstore-drop-unnecessary-initializations.patch
Patch0879: 0879-pstopre-fix-return-value-of-list_files.patch
Patch0880: 0880-pstore-remove-temporary-file-on-failure.patch
Patch0881: 0881-pstore-do-not-add-FILE-journal-entry-if-content_size.patch
Patch0882: 0882-pstore-run-only-when-sys-fs-pstore-is-not-empty.patch
Patch0883: 0883-pstore-fix-use-after-free.patch
Patch0884: 0884-pstore-refuse-to-run-if-arguments-are-specified.patch
Patch0885: 0885-pstore-allow-specifying-src-and-dst-dirs-are-argumen.patch
Patch0886: 0886-pstore-rework-memory-handling-for-dmesg.patch
Patch0887: 0887-pstore-fixes-for-dmesg.txt-reconstruction.patch
Patch0888: 0888-pstore-Don-t-start-systemd-pstore.service-in-contain.patch
Patch0889: 0889-units-pull-in-systemd-pstore.service-from-sysinit.ta.patch
Patch0890: 0890-units-drop-dependency-on-systemd-remount-fs.service-.patch
Patch0891: 0891-units-make-sure-systemd-pstore-stops-at-shutdown.patch
Patch0892: 0892-pstore-Run-after-modules-are-loaded.patch
Patch0893: 0893-pstore-do-not-try-to-load-all-known-pstore-modules.patch
Patch0894: 0894-logind-session-make-stopping-of-idle-session-visible.patch
Patch0895: 0895-journald-Increase-stdout-buffer-size-sooner-when-alm.patch
Patch0896: 0896-journald-rework-end-of-line-marker-handling-to-use-a.patch
Patch0897: 0897-journald-use-the-fact-that-client_context_release-re.patch
Patch0898: 0898-journald-rework-pid-change-handling.patch
Patch0899: 0899-test-Add-a-test-case-for-15654.patch
Patch0900: 0900-test-Stricter-test-case-for-15654-Add-more-checks.patch
Patch0901: 0901-man-document-the-new-_LINE_BREAK-type.patch
Patch0902: 0902-journald-server-always-create-state-file-in-signal-h.patch
Patch0903: 0903-journald-server-move-relinquish-code-into-function.patch
Patch0904: 0904-journald-server-always-touch-state-file-in-signal-ha.patch
%ifarch %{ix86} x86_64 aarch64 %ifarch %{ix86} x86_64 aarch64
%global have_gnu_efi 1 %global have_gnu_efi 1
@ -1305,7 +1371,8 @@ python3 %{SOURCE2} %buildroot <<EOF
EOF EOF
%check %check
%meson_test # Add --num-processes 1 as workaround for issues on ppc64le - AttributeError: 'NoneType' object has no attribute '_add_reader' - https://github.com/python/cpython/issues/82200
%meson_test --num-processes 1
############################################################################################# #############################################################################################
@ -1517,28 +1584,96 @@ fi
%files tests -f .file-list-tests %files tests -f .file-list-tests
%changelog %changelog
* Mon Jan 30 2023 systemd maintenance team <systemd-maint@redhat.com> - 239-68.4 * Tue Mar 14 2023 systemd maintenance team <systemd-maint@redhat.com> - 239-74
- basic: add STRERROR() wrapper for strerror_r() (#2155519) - journald-server: always create state file in signal handler (#2174645)
- coredump: put context array into a struct (#2155519) - journald-server: move relinquish code into function (#2174645)
- coredump: do not allow user to access coredumps with changed uid/gid/capabilities (#2155519) - journald-server: always touch state file in signal handler (#2174645)
* Fri Jan 27 2023 systemd maintenance team <systemd-maint@redhat.com> - 239-68.3 * Mon Feb 27 2023 systemd maintenance team <systemd-maint@redhat.com> - 239-73
- core: bring manager_startup() and manager_reload() more inline (#2164049) - journald: add API to move logging from /var to /run again (#1873540)
- journalctl: add new --relinquish and --smart-relinquish options (#1873540)
- units: automatically revert to /run logging on shutdown if necessary (#1873540)
- pstore: Tool to archive contents of pstore (#2158832)
- meson: drop redundant line (#2158832)
- pstore: drop unnecessary initializations (#2158832)
- pstopre: fix return value of list_files() (#2158832)
- pstore: remove temporary file on failure (#2158832)
- pstore: do not add FILE= journal entry if content_size == 0 (#2158832)
- pstore: run only when /sys/fs/pstore is not empty (#2158832)
- pstore: fix use after free (#2158832)
- pstore: refuse to run if arguments are specified (#2158832)
- pstore: allow specifying src and dst dirs are arguments (#2158832)
- pstore: rework memory handling for dmesg (#2158832)
- pstore: fixes for dmesg.txt reconstruction (#2158832)
- pstore: Don't start systemd-pstore.service in containers (#2158832)
- units: pull in systemd-pstore.service from sysinit.target (#2158832)
- units: drop dependency on systemd-remount-fs.service from systemd-pstore.service (#2158832)
- units: make sure systemd-pstore stops at shutdown (#2158832)
- pstore: Run after modules are loaded (#2158832)
- pstore: do not try to load all known pstore modules (#2158832)
- logind-session: make stopping of idle session visible to admins (#2156780)
- journald: Increase stdout buffer size sooner, when almost full (#2029426)
- journald: rework end of line marker handling to use a field table (#2029426)
- journald: use the fact that client_context_release() returns NULL (#2029426)
- journald: rework pid change handling (#2029426)
- test: Add a test case for #15654 (#2029426)
- test: Stricter test case for #15654 (Add more checks) (#2029426)
- man: document the new _LINE_BREAK= type (#2029426)
* Mon Nov 21 2022 systemd maintenance team <systemd-maint@redhat.com> - 239-68.2 * Fri Feb 17 2023 systemd maintenance team <systemd-maint@redhat.com> - 239-72
- basic: recognize pdfs filesystem as a network filesystem (#2143100) - test: import logind test from debian/ubuntu test suite (#1866955)
- test: introduce inst_recursive() helper function (#1866955)
- tests: verify that Lock D-Bus signal is sent when IdleAction=lock (#1866955)
- systemctl: simplify halt_main() (#2053273)
- systemctl: shutdown don't fallback on auth fail (#2053273)
- systemctl: reintroduce the original halt_main() (#2053273)
- systemctl: preserve old behavior unless requested (#2053273)
- pam_systemd: suppress LOG_DEBUG log messages if debugging is off (#2170084)
- udev/net_id: introduce naming scheme for RHEL-8.8 (#2170499)
- pam: add a call to pam_namespace (#1861836)
* Mon Nov 07 2022 systemd maintenance team <systemd-maint@redhat.com> - 239-68.1 * Tue Jan 31 2023 systemd maintenance team <systemd-maint@redhat.com> - 239-71
- time-util: fix buffer-over-run (#2139390) - manager: limit access to private dbus socket (#2119405)
- core: move reset_arguments() to the end of main's finish (#2127170) - journalctl: do not treat EINTR as an error when waiting for events (#2161683)
- core: bring manager_startup() and manager_reload() more inline (#2059633)
- pam: add a call to pam_namespace (#1861836)
- virt: Support detection for ARM64 Hyper-V guests (#2158307)
- virt: Fix the detection for Hyper-V VMs (#2158307)
- basic: add STRERROR() wrapper for strerror_r() (#2155520)
- coredump: put context array into a struct (#2155520)
- coredump: do not allow user to access coredumps with changed uid/gid/capabilities (#2155520)
* Tue Sep 27 2022 systemd maintenance team <systemd-maint@redhat.com> - 239-68 * Mon Jan 16 2023 systemd maintenance team <systemd-maint@redhat.com> - 239-70
- basic: recognize pdfs filesystem as a network filesystem (#2094661)
- core: move reset_arguments() to the end of main's finish (#2127131)
- manager: move inc. of n_reloading into a function (#2136869)
- core: Add new DBUS properties UnitsReloadStartTimestamp and UnitsLoadTimestampMontonic (#2136869)
- core: Indicate the time when the manager started loading units the last time (#2136869)
- core: do not touch /run/systemd/systemd-units-load from user session instances (#2136869)
- sysctl: downgrade message when we have no permission (#2158160)
- core: respect SELinuxContext= for socket creation (#2136738)
- manager: use target process context to set socket context (#2136738)
- virt: detect Amazon EC2 Nitro instance (#2117948)
- machine-id-setup: generate machine-id from DMI product ID on Amazon EC2 (#2117948)
- virt: use string table to detect VM or container (#2117948)
- fileio: introduce read_full_virtual_file() for reading virtual files in sysfs, procfs (#2117948)
- Use BIOS characteristics to distinguish EC2 bare-metal from VMs (#2117948)
- device: drop refuse_after (#2043524)
* Tue Nov 08 2022 systemd maintenance team <systemd-maint@redhat.com> - 239-69
- logind: optionally watch utmp for login data (#2122288) - logind: optionally watch utmp for login data (#2122288)
- logind: add hashtable for finding session by leader PID (#2122288) - logind: add hashtable for finding session by leader PID (#2122288)
- core/load-fragment: move config_parse_sec_fix_0 to src/shared (#2122288) - core/load-fragment: move config_parse_sec_fix_0 to src/shared (#2122288)
- sd-event: add relative timer calls (#2122288) - sd-event: add relative timer calls (#2122288)
- logind: add option to stop idle sessions after specified timeout (#2122288) - logind: add option to stop idle sessions after specified timeout (#2122288)
- logind: schedule idle check full interval from now if we couldn't figure out atime timestamp (#2122288) - logind: schedule idle check full interval from now if we couldn't figure out atime timestamp (#2122288)
- ci(lint): add shell linter - Differential ShellCheck (#2122499)
- meson: do not compare objects of different types (#2122499)
- journal-remote: use MHD_HTTP_CONTENT_TOO_LARGE as MHD_HTTP_PAYLOAD_TOO_LARGE is deprecated since 0.9.74 (#2122499)
- Fix build with µhttpd 0.9.71 (#2122499)
- ci: replace LGTM with CodeQL (#2122499)
- ci(mergify): Update policy - Drop LGTM checks (#2122499)
- time-util: fix buffer-over-run (#2139391)
* Fri Aug 26 2022 systemd maintenance team <systemd-maint@redhat.com> - 239-67 * Fri Aug 26 2022 systemd maintenance team <systemd-maint@redhat.com> - 239-67
- resolved: pin stream while calling callbacks for it (#2110549) - resolved: pin stream while calling callbacks for it (#2110549)