Commit Graph

1033 Commits

Author SHA1 Message Date
Dave Young
bbe10baeb2 Release 2.0.14-5 2017-03-09 16:59:31 +08:00
Pingfan Liu
14ec322b4b kdump-lib.sh: fix incorrect usage with pipe as input for grep -q in is_pcs_fence_kdump()
For -q option, as man grep says:  Exit immediately with zero status if
any match is found, even if an error was detected.
So when matching, the read side of pipe is closed by "grep -q", while
the write side still try to write more data, which cause SIGPIPE to the
process, and the shell can not exit with 0. It depends on the kernel's
implementation of pipe to decide how much data written by the producer
can trigger the bug.

Bash test script:
  #!/bin/sh
  set -o pipefail
  dd if=/dev/zero of=text.file bs=1M count=1
  sed -i '1s/^/keyword /' text.file
  cat text.file | grep -q keyword
  echo $?

Notice the "set -o pipefail" is set by dracut, so
mkdumprd -> dracut -> dracut-module-setup.sh -> is_pcs_fence_kdump()
trigger the bug.

Signed-off-by: Pingfan Liu <piliu@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
Acked-by: Pratyush Anand <panand@redhat.com>
2017-03-08 13:07:20 +08:00
Pingfan Liu
ad404977eb Document: fix incorrect link in fadump-how.txt
The file fadump-howto.txt has an incorrect link for further information
about sysrq usage. Fix it.

Signed-off-by: Pingfan Liu <piliu@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2017-03-08 13:07:20 +08:00
Dave Young
00bd88cd58 Release 2.0.14-4 2017-01-23 16:00:59 +08:00
Tong Li
1c27a3d827 drop kdump script rhcrashkernel-param in kexec-tools repo
Resolves: bz1399436

Since currently crashkernel= will be handled in kdump anaconda addon
we can safely remove rhcrashkernel-param callback.

Signed-off-by: Tong Li <tonli@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2017-01-23 15:59:47 +08:00
Xunlei Pang
2040103bd7 kdumpctl: sanity check of nr_cpus for x86_64 in case running out of vectors
Check the number of cpus for x86_64 kdump kernel to boot with.
We met an issue on x86_64: kdump runs out of vectors with the
default "nr_cpus=1", when requesting tons of irqs.

This patch detects such situation and warns users about the risk.

The total number of vectors percpu is 256 defined by x86 architecture.
The available vectors can be allocated to io devices percpu starts
from FIRST_EXTERNAL_VECTOR(see kernel code), and some high-numbered
ones are consumed by some system interrupts. As a result, the vectors
for io device are within [FIRST_EXTERNAL_VECTOR, FIRST_SYSTEM_VECTOR),
with one known exception, 0x80 within the range is reserved specially
as the syscall vector.

FIRST_EXTERNAL_VECTOR is invariably 32, while FIRST_SYSTEM_VECTOR can
vary between different kernel versions. E.g. FIRST_SYSTEM_VECTOR gets
0xef(with CONFIG_X86_LOCAL_APIC on)for linux-4.10, that is 17 vectors
reserved, considering it may increase in the future and the special
vectors, we use a flexible variance and assume there are 32 reserved
from FIRST_EXTERNAL_VECTOR. Then the max vectors for device interrupts
percpu is: (256-32)-32=192, we acquire the number N of device interrupts
from /proc/irq/, then the number of minimal cpus required is calculated:
(N + 192 - 1) / 192

Signed-off-by: Xunlei Pang <xlpang@redhat.com>
Acked-by: Pratyush Anand <panand@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2017-01-23 15:52:24 +08:00
Xunlei Pang
211b36b8f9 kdumpctl: change prepare_cmdline() to operate KDUMP_COMMANDLINE directly
Since KDUMP_COMMANDLINE is a global variable, prepare_cmdline can
modify it directly instead of echoing back the result. This change
enables it to output messages.

Changed some coding styles.

Signed-off-by: Xunlei Pang <xlpang@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
Acked-by: Pratyush Anand <panand@redhat.com>
2017-01-23 15:51:28 +08:00
Tong Li
8f07591ff1 use --hostonly-i18n for dracut
Resolves: bz1411240
Use --hostonly-i18n to force dracut to install only needed
keyboard and font files according to host's configuration, which
reduced initramfs's size by 2M on F25 x86_64.

Signed-off-by: Tong Li <tonli@redhat.com>
Reviewed-by: Xunlei Pang <xlpang@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2017-01-23 15:50:47 +08:00
Dave Young
769e1d53c6 Release 2.0.14-3 2017-01-04 15:00:49 +08:00
Dave Young
bd8a109b91 Rebase makedumpfile 1.6.1
Rebase makedumpfile to 1.6.1 and drop the upstreamed patches

Signed-off-by: Dave Young <dyoung@redhat.com>
2017-01-04 14:58:56 +08:00
Dave Young
55ff20016d Delete unused patches
Signed-off-by: Dave Young <dyoung@redhat.com>
2017-01-04 14:53:39 +08:00
Dave Young
abfed0b6d6 Release 2.0.14-2
Fix build error, forgot to update anaconda addon file in rpm spec.
Update changelog etc.
2016-12-20 17:45:45 +08:00
Dave Young
6cd80235c6 Update addon tarball in rpm spec
Previously forgot to update addon file name in rpm spec..
Signed-off-by: Dave Young <dyoung@redhat.com>
2016-12-20 17:43:09 +08:00
Dave Young
986a4999d4 Release 2.0.14-1
Rebase kexec-tools to upstream 2.0.14

Signed-off-by: Dave Young <dyoung@redhat.com>
2016-12-20 17:09:40 +08:00
Dave Young
456574e2c1 Update kdump-anaconda-addon
Resolves: bz1403658
commit d3e2e0ebcb2f9ba803576e48e7b9c752cc7e9f1a
Author: Tong Li <tonli@redhat.com>
Date:   Mon Dec 12 13:10:08 2016 +0800

    Fix 'an unknown error has occurred' issue when selecting languages using non-latin characters

    Now when kdump_anaconda_addon is enabled and languages which use non-latin
    characters are selected in anaconda, e.g. Chinese and Japanese, it will
    raise an error and unable to continue to finish the installation process.
    This is because 'gettext.ldgettext' will return a byte object when
    translation includes non-latin character, while anaconda's core code
    requires a string. To fix this, we apply the mothod used by pyanaconda,
    which is invoking gettext after getting a translation instance. This can
    make sure that a str object will be returned.

Signed-off-by: Dave Young <dyoung@redhat.com>
2016-12-20 17:05:22 +08:00
Dave Young
9cfdb303ff Uploaded new 2.0.14 kexec-tools, drop 2.0.13 in sources file
Signed-off-by: Dave Young <dyoung@redhat.com>
2016-12-20 16:58:16 +08:00
Dave Young
b676637e4b Update sources file with sha512 sum
new fedpkg need sha512 sum, see below error:
$ fedpkg upload kexec-tools-2.0.14.tar.xz
Could not execute upload: Can not upload a new source file with a sha512 hash, as the "/home/dyoung/fedpkg/kexec-tools/sources" file contains at least one line with a md5 hash.

Please redo the whole "/home/dyoung/fedpkg/kexec-tools/sources" file using:
    `/bin/fedpkg new-sources file1 file2 ...`

Thus update the sources file here.

Signed-off-by: Dave Young <dyoung@redhat.com>
2016-12-20 16:55:48 +08:00
Dave Young
8387948a7f Drop old sources in sources file
Signed-off-by: Dave Young <dyoung@redhat.com>
2016-12-20 16:53:03 +08:00
Dave Young
8cfb265790 Release 2.0.13-9 2016-11-28 10:43:20 +08:00
Dave Young
e2de6e1e9a rename function kdump_to_udev_name
kdump_to_udev_name function name cause confusion to people, change it to
kdump_get_persistent_dev which sounds better.

Signed-off-by: Dave Young <dyoung@redhat.com>
Reviewed-by: Xunlei Pang <xlpang@redhat.com>
2016-11-28 10:41:22 +08:00
Dave Young
3742e9d0c3 Raw dump: use by-id as persistent policy in 2nd kernel
Although we use by-id in mkdumprd as persistent policy for the dump target
checking, finally it is not used in kdump 2nd kernel because we call dracut
function in module-setup.sh without persistent policy specified that means
kdump will copy default "by-uuid" dev name.

Though by-uuid usually works and it is still better to fix it as raw disk
uuid make no sense.

Also do not need to call bind mount adjust function for raw dump, here add
another switch case for raw dump and cleanup the functions with short
variable names to keep code shorter.

Signed-off-by: Dave Young <dyoung@redhat.com>
Reviewed-by: Xunlei Pang <xlpang@redhat.com>
2016-11-28 10:41:22 +08:00
Dave Young
631d979eb3 drop dracut duplicate functions
We maintained several kdump specific functions which are duplicate with the
similar versions in dracut,  Dracut upstream splitted dracut init stuff from
dracut-functions.sh so that we can source it now.

Notes about kdump_get_presistent_dev:
Dracut now has a persistent_policy feature, for kdump when we dump to
raw disks we do not care the filesystem uuid and labels so we prefer to
search disk id instead. For raw disk set the persistent_policy before calling
get_persistent_dev ensure kdump logic still work.

Tested filesystem and raw dump in kvm guests.

[Xunlei: drop other functions other than get_persistent_dev.]

Signed-off-by: Dave Young <dyoung@redhat.com>
Reviewed-by: Xunlei Pang <xlpang@redhat.com>
2016-11-28 10:41:22 +08:00
Xunlei Pang
391969ced7 dracut-kdump: use POSIX shell syntax
kdump.sh may run under sh/dash in kdump kernel.

Signed-off-by: Xunlei Pang <xlpang@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2016-11-28 10:41:15 +08:00
Tong Li
ac1eb7edce Correct two typos in kdumpctl and kdump.conf
Signed-off-by: Tong Li <tonli@redhat.com>
Reviewed-by: Xunlei Pang <xlpang@redhat.com>
2016-11-28 10:41:05 +08:00
Dave Young
4137ad43c3 Release 2.0.13-8 2016-11-11 14:15:42 +08:00
Dave Young
9a1312a624 kexec/arch/i386: Add support for KASLR memory randomization
Backport upstream kexec-tools commit for correct kaslr page_offset value
commit 9f62cbddddfc93d78d9aafbddf3e1208cb242f7b
Author: Thomas Garnier <thgarnie@google.com>
Date:   Tue Sep 13 15:10:05 2016 +0800

    kexec/arch/i386: Add support for KASLR memory randomization

    Multiple changes were made on KASLR (right now in linux-next). One of
    them is randomizing the virtual address of the physical mapping, vmalloc
    and vmemmap memory sections. It breaks kdump ability to read physical
    memory.

    This change identifies if KASLR memories randomization is used by
    checking if the page_offset_base variable exists. It search for the
    correct PAGE_OFFSET value by looking at the loaded memory section and
    find the lowest aligned on PUD (the randomization level).

    Related commits on linux-next:
     - 0483e1fa6e09d4948272680f691dccb1edb9677f: Base for randomization
     - 021182e52fe01c1f7b126f97fd6ba048dc4234fd: Enable for PAGE_OFFSET

    Signed-off-by: Thomas Garnier <thgarnie@google.com>
    Signed-off-by: Simon Horman <horms@verge.net.au>

Signed-off-by: Dave Young <dyoung@redhat.com>
Acked-by: Baoquan He <bhe@redhat.com>
2016-11-11 14:14:59 +08:00
Dave Young
50c70debae Update kdump anaconda addon
*) Add support for new Anaconda addon methods (#1288636)
*) update latin translation (#1053550)

Signed-off-by: Dave Young <dyoung@redhat.com>
2016-11-11 11:03:59 +08:00
Hari Bathini
6a5e908d85 fadump: restore default initrd when fadump mode is disabled
When fadump mode is enabled, the default initrd is rebuilt with kdump
dracut module. As the default initrd is altered, the original default
initrd is backed up. But we are not restoring it when fadump mode is
disabled. This patch tries to restore the backed up default initrd on
disabling fadump mode.

Signed-off-by: Hari Bathini <hbathini@linux.vnet.ibm.com>
Acked-by: Dave Young <dyoung@redhat.com>
2016-11-11 11:01:13 +08:00
Hari Bathini
78e985e51c kdump/fadump: fix network interface name when switching from fadump to kdump
When a remote dump target is specified, kdump dracut module prefixes
'kdump-' to network interface name (ifname) as kernel assigned names
are not persistent. In fadump mode, kdump dracut module is added to
the default initrd, which adds the 'kdump-' prefix to the ifname of
the prodcution kernel itself. If fadump mode is disabled after this,
kdump dracut module picks the ifname that is already prefixed with
'kdump-' in the production kernel and adds another 'kdump-' to it,
making the ifname something like kdump-kdump-eth0 for kdump kernel.
Eventually, kdump kernel fails with below traces:

  dracut-initqueue[246]: RTNETLINK answers: Network is unreachable
  dracut-initqueue[246]: arping: Device kdump-kdump-eth0 not available.

The ip command shows the below:

  kdump:/# ip addr show kdump-kdump-eth0
  2: kdump-kdump-eth: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 \
  qdisc pfifo_fast state UNKNOWN qlen 1000
      link/ether 22:82:87:7b:98:02 brd ff:ff:ff:ff:ff:ff
      inet6 2002:903:15f:550:2082:87ff:fe7b:9802/64 scope global \
      mngtmpaddr dynamic
         valid_lft 2591890sec preferred_lft 604690sec
      inet6 fe80::2082:87ff:fe7b:9802/64 scope link
         valid_lft forever preferred_lft forever
  kdump:/#

The trailing 0 from kdump-kdump-eth0 is missing in the ifname, probably
truncated owing to ifname length limit, while setting.

This patch fixes this by avoiding addition of the prefix 'kdump-' when
such prefix is already present in the ifname.

Signed-off-by: Hari Bathini <hbathini@linux.vnet.ibm.com>
Acked-by: Dave Young <dyoung@redhat.com>
2016-11-11 11:00:22 +08:00
Tong Li
892bea7aae kdumpctl: filter 'root' kernel parameter when running in live images
Kernels of live images are booted with a kernel parameter which looks
like "root=live:CDLABEL=Fedora-WS-Live-25_A-2". This argument can't be
recognized by dracut during kdump process and will cause failure
of kdump if users didn't set KUDMP_COMMANDLINE in /etc/sysconfig/kdump.
So we should filter out 'root' when we find such a parameter in
/proc/cmdline to make kdump work correctly in live images.

Signed-off-by: Tong Li <tonli@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2016-11-11 10:56:35 +08:00
Tong Li
7e53e9fe38 Documentation: step by step guide on confiuring kdump in live images
This is a short document about how to setup kdump on live images. All
steps were tested on Fedora 25 Alpha LiveCD and saved vmcore captured
by kdump to a USB stick successfully.

Signed-off-by: Tong Li <tonli@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2016-11-11 10:56:11 +08:00
Dave Young
11785bdffb Release 2.0.13-7 2016-10-27 14:52:32 +08:00
Pratyush Anand
938280a5e9 x86_64: fix page_offset calculation
V1 of the KASLR patch which was backported in 2.0.13-6 had wrong
calculation for page_offset. This patch takes the correct difference.

Signed-off-by: Pratyush Anand <panand@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2016-10-27 14:52:15 +08:00
Dave Young
f77aa85025 Release 2.0.13-6 2016-10-27 14:44:27 +08:00
Pingfan Liu
aee378209e Add initial upstream support for kexec on aarch64
Fix Bug 925630 - kexec-tools: support for arm64
https://bugzilla.redhat.com/show_bug.cgi?id=925630
involves three things:
  1. back porting upstream code to enable the function of kexec-tools on arm64
     patchset backported from upstream:
        commit abdfe97736f89d9bc73662b9134604b0229a599e
        commit 522df5f7217fda01ece3f6ac3e9987b0320c2bb0
        commit 217bcc00c9309416a6c6cd0584196559d28a9259
  2. fix the arm related building issue by using autoreconf in spec file

  3. patches to fix the issue of higher version gcc used by koji (not upstrem yet,
     and the corresponding fix in kernel side is in other package)
        kexec-tools-2.0.13-arm64-Add-support-for-additional-relocations-in-the-kexec-purgatory-code.patch
        kexec-tools-2.0.13-arm64-Add-support-of-R_AARCH64_PREL32-relocation-in-.patch
[panand: apply patches in 3 of above paragraph]

Signed-off-by: Pingfan Liu <piliu@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
Acked-by: Pratyush Anand <panand@redhat.com>
2016-10-27 14:44:13 +08:00
Dave Young
658db69940 Support kdump for kaslr enabled Fedora kernel
Since in Fedora 25 kernel kaslr is enabled (x86) but makedumpfile can not save
a correct vmcore, so it means kdump default setup will not work.

Pratyush posted a patch series to upstream which can fix the issue. Let's merge them in F25, will get the normal fixes after it being merged in upstream, we hopefully can rebase soon in rawhide.

This is an urgent fix for F25 since F25 freeze is this week.

Signed-off-by: Dave Young <dyoung@redhat.com>
Acked-by: Pratyush Anand <panand@redhat.com>
Acked-by: Baoquan He <bhe@redhat.com>
2016-10-27 14:44:02 +08:00
Dave Young
f583213ea4 Release 2.0.13-5 2016-09-16 15:43:52 +08:00
Pratyush Anand
4db9e59e89 kdumpctl: fix target identification for systems without initrd
We get following error on the systems that have everything built-in and no
initrd is used.

	Kernel dev name of /dev/root is not found.
	Dump target /dev/root is probably not mounted.

It happens because `df $path` gets /dev/root from /proc/self/mountinfo.

Fix this by identifying real target device when `df $path` returns
Filesystem as /dev/root.

Reported-and-tested-by: Michael Holzheu <holzheu@linux.vnet.ibm.com>
Signed-off-by: Pratyush Anand <panand@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2016-09-16 15:40:48 +08:00
Pratyush Anand
1edfff7809 kdumpctl: remove duplicate statement
Since we also check for mount point of $_target after if/else loop, so
there is no need to do the same thing specifically in else loop as well.

Remove those duplicate statement from else loop.

Signed-off-by: Pratyush Anand <panand@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2016-09-16 15:40:48 +08:00
Pratyush Anand
1bb23e7536 kdumpctl: check /etc/fstab modification only when it exists
On a diskless client /etc/fstab does not exist. Therefore check
modification time of this file for rebuild only if it exists.

Also use --fstab option with findmnt only when /etc/fstab exists.

Signed-off-by: Pratyush Anand <panand@redhat.com>
Reviewed-by: Xunlei Pang <xlpang@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2016-09-16 15:40:48 +08:00
Pratyush Anand
9b95184251 kdumpctl: Kill duplicate code related to file modication check
commit "28e8c4b5ac89 kdumpctl: Move file modification check logic in
check_system_modified()" copied file modification check logic instead of
moving.
Kill the duplicate logic from original calling function check_rebuild().

Signed-off-by: Pratyush Anand <panand@redhat.com>
Reviewed-by: Xunlei Pang <xlpang@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2016-09-16 15:40:48 +08:00
Dave Young
fdd79a70f6 Release 2.0.13-4 2016-08-26 14:05:37 +08:00
Xunlei Pang
d614274092 spec: Update to add dependency on dracut version >= 044-117
Fedora dracut-network version >= 044-117 must be installed to work with
this set of kexec-tools patches. Therefore adding dependency for them.

Acked-by: Dave Young <dyoung@redhat.com>
Signed-off-by: Xunlei Pang <xlpang@redhat.com>
2016-08-26 14:03:48 +08:00
Xunlei Pang
8eb199cc91 mkdumprd: Do not add "nfs" dracut module explicitly
Kdump explicitly adds "nfs" dracut module in case of nfs dumping,
actually in case of nfs dump, nfs is a mount target, and will be
added into host_fs_types[], thus dracut will add it automatically,
according to 95nfs/module-setup.sh check().

So, we can safely remove all the add_dracut_module "nfs".

Acked-by: Dave Young <dyoung@redhat.com>
Signed-off-by: Xunlei Pang <xlpang@redhat.com>
2016-08-26 14:03:48 +08:00
Xunlei Pang
b22ed8dd3a kexec-kdump-howto: Add doc about the special mount information via "dracut_args"
Update "kexec-kdump-howto" to illustrate the usage of special mount
information via "dracut_args".

Reviewed-by: Pratyush Anand <panand@redhat.com>
Suggested-by: Dave Young <dyoung@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
Signed-off-by: Xunlei Pang <xlpang@redhat.com>
2016-08-26 14:03:48 +08:00
Xunlei Pang
74c6f46429 Support special mount information via "dracut_args"
There are some complaints about nfs kdump that users must mount
nfs beforehand, which may cause some overhead to nfs server.
For example, there're thounsands of diskless clients deployed with
nfs dumping, each time the client is boot up, it will trigger
kdump rebuilding so will mount nfs, thus resulting in thousands
of nfs request concurrently imposed on the same nfs server.

We introduce a new way of specifying mount information via the
already-existent "dracut_args" directive(so avoid adding extra
directives in /etc/kdump.conf), we will skip all the filesystem
mounting and checking stuff for it. So it can be used in the
above-mentioned nfs scenario to avoid severe nfs server overhead.

Specifically, if there is any "--mount" information specified via
"dracut_args" in /etc/kdump.conf, always use it as the final mount
without any validation(mounting or checking like mount options,
fs size, etc), so users are expected to ensure its correctness.

NOTE:
-Only one mount target is allowed using "dracut_args" globally.
-Dracut will create <mountpoint> if it doesn't exist in kdump kernel,
 <mountpoint> must be specified as an absolute path.
-Users should do a test first and ensure it works because kdump does
 not prepare the mount or check all the validity.

Reviewed-by: Pratyush Anand <panand@redhat.com>
Suggested-by: Dave Young <dyoung@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
Signed-off-by: Xunlei Pang <xlpang@redhat.com>
2016-08-26 14:03:48 +08:00
Dave Young
8e669fa5dd Release 2.0.13-3 2016-08-09 15:42:46 +08:00
Dave Young
97581f1a43 Fix armv7 build failure
Upstream kexec-tools 2.0.13 tarball does not pack arm phys_to_virt.h and
iomem.h, add them to avoid armv7 build failure.
2016-08-09 15:42:39 +08:00
Dave Young
3638e9e3a5 Release 2.0.13-2 2016-08-09 14:34:10 +08:00
Dave Young
ba543e01ed 2.0.13-1 build fail due to patches can not applied. Drop the old patches now.
They have been included in 2.0.13
2016-08-09 14:34:01 +08:00