import crash-7.3.1-5.el8

This commit is contained in:
CentOS Sources 2022-02-11 05:24:41 +00:00 committed by Stepan Oksanichenko
parent 6615b535b0
commit bd0c6a5112
12 changed files with 857 additions and 1 deletions

View File

@ -0,0 +1,70 @@
From 5c4f786450ea61b87d4db0092288df83dd5cb454 Mon Sep 17 00:00:00 2001
From: Qi Zheng <zhengqi.arch@bytedance.com>
Date: Tue, 21 Dec 2021 15:40:31 +0800
Subject: [PATCH 01/11] Fix pvops Xen detection for arm machine
Since the xen_start_info on the arm/arm64 platform points to a static
variable '_xen_start_info'(see its definition as below), which makes
that the address of xen_start_info will never be null.
arch/arm/xen/enlighten.c:40:static struct start_info _xen_start_info;
arch/arm/xen/enlighten.c:41:struct start_info *xen_start_info = &_xen_start_info;
arch/arm/xen/enlighten.c:42:EXPORT_SYMBOL(xen_start_info);
As a result, the is_pvops_xen() in commit 4badc6229c69 ("Fix pvops
Xen detection for kernels >= v4.20") always returns TRUE because it
can always read out the non-null address of xen_start_info, finally
the following error will be reported on arm/arm64 platform(non-Xen
environment) because p2m_mid_missing and xen_p2m_addr are not defined:
crash: cannot resolve "p2m_top"
For the arm/arm64 platform, fix it by using xen_vcpu_info instead of
xen_start_info to detect Xen dumps.
In addition, also explicitly narrow the scope of the xen_start_info
check to x86 with the machine_type(), there is no need to check it on
other architectures.
Fixes: 4badc6229c69 ("Fix pvops Xen detection for kernels >= v4.20")
Signed-off-by: Qi Zheng <zhengqi.arch@bytedance.com>
Acked-by: Kazuhito Hagio <k-hagio-ab@nec.com>
Signed-off-by: Lianbo Jiang <lijiang@redhat.com>
---
kernel.c | 20 +++++++++++++++-----
1 file changed, 15 insertions(+), 5 deletions(-)
diff --git a/kernel.c b/kernel.c
index 8ae9e0c169ff..a637dd0eb8f8 100644
--- a/kernel.c
+++ b/kernel.c
@@ -10754,11 +10754,21 @@ is_pvops_xen(void)
STREQ(sym, "paravirt_patch_default")))
return TRUE;
- if (symbol_exists("xen_start_info") &&
- readmem(symbol_value("xen_start_info"), KVADDR, &addr,
- sizeof(void *), "xen_start_info", RETURN_ON_ERROR) &&
- addr != 0)
- return TRUE;
+ if (machine_type("X86") || machine_type("X86_64")) {
+ if (symbol_exists("xen_start_info") &&
+ readmem(symbol_value("xen_start_info"), KVADDR, &addr,
+ sizeof(void *), "xen_start_info", RETURN_ON_ERROR) &&
+ addr != 0)
+ return TRUE;
+ }
+
+ if (machine_type("ARM") || machine_type("ARM64")) {
+ if (symbol_exists("xen_vcpu_info") &&
+ readmem(symbol_value("xen_vcpu_info"), KVADDR, &addr,
+ sizeof(void *), "xen_vcpu_info", RETURN_ON_ERROR) &&
+ addr != 0)
+ return TRUE;
+ }
return FALSE;
}
--
2.20.1

View File

@ -0,0 +1,101 @@
From 78255e3b33f8d51eb893e662dd1b05a008246b9d Mon Sep 17 00:00:00 2001
From: Lianbo Jiang <lijiang@redhat.com>
Date: Fri, 24 Dec 2021 18:56:35 +0800
Subject: [PATCH 02/11] Handle blk_mq_ctx member changes for kernels 5.16-rc1
and later
Kernel commit 9a14d6ce4135 ("block: remove debugfs blk_mq_ctx
dispatched/merged/completed attributes") removed the member
rq_dispatched and rq_completed from struct blk_mq_ctx. Without
the patch, "dev -d|-D" options will fail with the following error:
crash> dev -d
MAJOR GENDISK NAME REQUEST_QUEUE TOTAL ASYNC SYNC
dev: invalid structure member offset: blk_mq_ctx_rq_dispatched
FILE: dev.c LINE: 4229 FUNCTION: get_one_mctx_diskio()
Signed-off-by: Lianbo Jiang <lijiang@redhat.com>
Signed-off-by: Kazuhito Hagio <k-hagio-ab@nec.com>
---
dev.c | 57 +++++++++++++++++++++++++++++++++++++++------------------
1 file changed, 39 insertions(+), 18 deletions(-)
diff --git a/dev.c b/dev.c
index effe789f38d8..a493e51ac95c 100644
--- a/dev.c
+++ b/dev.c
@@ -4246,6 +4246,10 @@ get_mq_diskio(unsigned long q, unsigned long *mq_count)
unsigned long mctx_addr;
struct diskio tmp;
+ if (INVALID_MEMBER(blk_mq_ctx_rq_dispatched) ||
+ INVALID_MEMBER(blk_mq_ctx_rq_completed))
+ return;
+
memset(&tmp, 0x00, sizeof(struct diskio));
readmem(q + OFFSET(request_queue_queue_ctx), KVADDR, &queue_ctx,
@@ -4475,24 +4479,41 @@ display_one_diskio(struct iter *i, unsigned long gendisk, ulong flags)
&& (io.read + io.write == 0))
return;
- fprintf(fp, "%s%s%s %s%s%s%s %s%5d%s%s%s%s%s",
- mkstring(buf0, 5, RJUST|INT_DEC, (char *)(unsigned long)major),
- space(MINSPACE),
- mkstring(buf1, VADDR_PRLEN, LJUST|LONG_HEX, (char *)gendisk),
- space(MINSPACE),
- mkstring(buf2, 10, LJUST, disk_name),
- space(MINSPACE),
- mkstring(buf3, VADDR_PRLEN <= 11 ? 11 : VADDR_PRLEN,
- LJUST|LONG_HEX, (char *)queue_addr),
- space(MINSPACE),
- io.read + io.write,
- space(MINSPACE),
- mkstring(buf4, 5, RJUST|INT_DEC,
- (char *)(unsigned long)io.read),
- space(MINSPACE),
- mkstring(buf5, 5, RJUST|INT_DEC,
- (char *)(unsigned long)io.write),
- space(MINSPACE));
+ if (use_mq_interface(queue_addr) &&
+ (INVALID_MEMBER(blk_mq_ctx_rq_dispatched) ||
+ INVALID_MEMBER(blk_mq_ctx_rq_completed)))
+ fprintf(fp, "%s%s%s %s%s%s%s %s%s%s",
+ mkstring(buf0, 5, RJUST|INT_DEC, (char *)(unsigned long)major),
+ space(MINSPACE),
+ mkstring(buf1, VADDR_PRLEN, LJUST|LONG_HEX, (char *)gendisk),
+ space(MINSPACE),
+ mkstring(buf2, 10, LJUST, disk_name),
+ space(MINSPACE),
+ mkstring(buf3, VADDR_PRLEN <= 11 ? 11 : VADDR_PRLEN,
+ LJUST|LONG_HEX, (char *)queue_addr),
+ space(MINSPACE),
+ mkstring(buf4, 17, RJUST, "(not supported)"),
+ space(MINSPACE));
+
+ else
+ fprintf(fp, "%s%s%s %s%s%s%s %s%5d%s%s%s%s%s",
+ mkstring(buf0, 5, RJUST|INT_DEC, (char *)(unsigned long)major),
+ space(MINSPACE),
+ mkstring(buf1, VADDR_PRLEN, LJUST|LONG_HEX, (char *)gendisk),
+ space(MINSPACE),
+ mkstring(buf2, 10, LJUST, disk_name),
+ space(MINSPACE),
+ mkstring(buf3, VADDR_PRLEN <= 11 ? 11 : VADDR_PRLEN,
+ LJUST|LONG_HEX, (char *)queue_addr),
+ space(MINSPACE),
+ io.read + io.write,
+ space(MINSPACE),
+ mkstring(buf4, 5, RJUST|INT_DEC,
+ (char *)(unsigned long)io.read),
+ space(MINSPACE),
+ mkstring(buf5, 5, RJUST|INT_DEC,
+ (char *)(unsigned long)io.write),
+ space(MINSPACE));
if (VALID_MEMBER(request_queue_in_flight)) {
if (!use_mq_interface(queue_addr)) {
--
2.20.1

View File

@ -0,0 +1,59 @@
From c48177972f351d7853abb2a57709628c75ee38bc Mon Sep 17 00:00:00 2001
From: Lianbo Jiang <lijiang@redhat.com>
Date: Thu, 6 Jan 2022 22:34:26 +0800
Subject: [PATCH 03/11] Fix for "timer -r" option to display all the per-CPU
clocks
Currently, the hrtimer_max_clock_bases is hard-coded to 3, which
makes that crash only prints three clocks, and the rest of clocks
are not displayed.
Without the patch:
crash> timer -r -C 11
CPU: 11 HRTIMER_CPU_BASE: ffff9a775f95ee00
CLOCK: 0 HRTIMER_CLOCK_BASE: ffff9a775f95ee80 [ktime_get]
(empty)
CLOCK: 1 HRTIMER_CLOCK_BASE: ffff9a775f95ef00 [ktime_get_real]
(empty)
CLOCK: 2 HRTIMER_CLOCK_BASE: ffff9a775f95ef80 [ktime_get_boottime]
(empty)
With the patch:
crash> timer -r -C 11
CPU: 11 HRTIMER_CPU_BASE: ffff9a775f95ee00
CLOCK: 0 HRTIMER_CLOCK_BASE: ffff9a775f95ee80 [ktime_get]
(empty)
CLOCK: 1 HRTIMER_CLOCK_BASE: ffff9a775f95ef00 [ktime_get_real]
(empty)
CLOCK: 2 HRTIMER_CLOCK_BASE: ffff9a775f95ef80 [ktime_get_boottime]
(empty)
...
CLOCK: 7 HRTIMER_CLOCK_BASE: ffff9a775f95f200 [ktime_get_clocktai]
(empty)
Signed-off-by: Lianbo Jiang <lijiang@redhat.com>
---
kernel.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/kernel.c b/kernel.c
index a637dd0eb8f8..a44a9c52ace0 100644
--- a/kernel.c
+++ b/kernel.c
@@ -7672,7 +7672,8 @@ dump_hrtimer_data(const ulong *cpus)
if (VALID_STRUCT(hrtimer_clock_base)) {
hrtimer_max_clock_bases = 2;
if (symbol_exists("ktime_get_boottime"))
- hrtimer_max_clock_bases = 3;
+ hrtimer_max_clock_bases = MEMBER_SIZE("hrtimer_cpu_base", "clock_base") /
+ SIZE(hrtimer_clock_base);
} else if (VALID_STRUCT(hrtimer_base)) {
max_hrtimer_bases = 2;
} else
--
2.20.1

View File

@ -0,0 +1,69 @@
From 1706f8b6ab50cd25e8fdabe8d50a37ce89bd60e0 Mon Sep 17 00:00:00 2001
From: Lianbo Jiang <lijiang@redhat.com>
Date: Thu, 6 Jan 2022 12:01:17 +0800
Subject: [PATCH 04/11] Fix for "bt -v" option to display the stack-end address
correctly
The "bt -v" command prints incorrect stack-end address when the
"CONFIG_THREAD_INFO_IN_TASK=y" is enabled in kernel, the "bt -v"
command output shows that the value stored at 0xffff8dee0312c198
is 0xffffffffc076400a, however, the value stored actually at
0xffff8dee0312c198 is NULL(0x0000000000000000), the stack-end
address is incorrect.
Without the patch:
crash> bt -v
PID: 28642 TASK: ffff8dee0312c180 CPU: 0 COMMAND: "insmod"
possible stack overflow: ffff8dee0312c198: ffffffffc076400a != STACK_END_MAGIC
^^^^^^^^^^^^^^^^
crash> rd 0xffff8dee0312c198
ffff8dee0312c198: 0000000000000000 ........
^^^^^^^^^^^^^^^^
With the patch:
crash> bt -v
PID: 28642 TASK: ffff8dee0312c180 CPU: 0 COMMAND: "insmod"
possible stack overflow: ffff991340bc0000: ffffffffc076400a != STACK_END_MAGIC
crash> rd 0xffff991340bc0000
ffff991340bc0000: ffffffffc076400a .@v.....
Signed-off-by: Lianbo Jiang <lijiang@redhat.com>
---
task.c | 10 ++++++++--
1 file changed, 8 insertions(+), 2 deletions(-)
diff --git a/task.c b/task.c
index bb6a5da8ad33..b5ddc88e0acb 100644
--- a/task.c
+++ b/task.c
@@ -11202,7 +11202,7 @@ check_stack_overflow(void)
{
int i, overflow, cpu_size, cpu, total;
char buf[BUFSIZE];
- ulong magic, task, stackbase;
+ ulong magic, task, stackbase, location;
struct task_context *tc;
if (!tt->stack_end_magic &&
@@ -11286,9 +11286,15 @@ check_stack_end_magic:
if (magic != STACK_END_MAGIC) {
if (!overflow)
print_task_header(fp, tc, 0);
+
+ if (tt->flags & THREAD_INFO_IN_TASK)
+ location = task_to_stackbase(tc->task);
+ else
+ location = tc->thread_info + SIZE(thread_info);
+
fprintf(fp,
" possible stack overflow: %lx: %lx != STACK_END_MAGIC\n",
- tc->thread_info + SIZE(thread_info), magic);
+ location, magic);
overflow++, total++;
}
--
2.20.1

View File

@ -0,0 +1,35 @@
From f5637f341533ef2b28e2d6a6b12fcfb00d0fff2d Mon Sep 17 00:00:00 2001
From: Lianbo Jiang <lijiang@redhat.com>
Date: Mon, 10 Jan 2022 17:25:06 +0800
Subject: [PATCH 05/11] Fix for HZ calculation on Linux 5.14 and later
Kernel commit 3e9a99eba058 ("block/mq-deadline: Rename dd_init_queue()
and dd_exit_queue()") renamed dd_init_queue to dd_init_sched. Without
the patch, the 'help -m' may print incorrect hz value as follows:
crash> help -m | grep hz
hz: 1000 <---The correct hz value on ppc64le machine is 100.
^^^^
Fixes: b93027ce5c75 ("Add alternate HZ calculation using write_expire")
Signed-off-by: Lianbo Jiang <lijiang@redhat.com>
---
task.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/task.c b/task.c
index b5ddc88e0acb..76e184ae70b1 100644
--- a/task.c
+++ b/task.c
@@ -440,6 +440,8 @@ task_init(void)
}
} else if ((symbol_exists("dd_init_queue") &&
gdb_set_crash_scope(symbol_value("dd_init_queue"), "dd_init_queue")) ||
+ (symbol_exists("dd_init_sched") &&
+ gdb_set_crash_scope(symbol_value("dd_init_sched"), "dd_init_sched")) ||
(symbol_exists("deadline_init_queue") &&
gdb_set_crash_scope(symbol_value("deadline_init_queue"), "deadline_init_queue"))) {
char buf[BUFSIZE];
--
2.20.1

View File

@ -0,0 +1,40 @@
From a392b27653e4e75460753522af0f006006b4dc4e Mon Sep 17 00:00:00 2001
From: Alexander Egorenkov <egorenar@linux.ibm.com>
Date: Mon, 6 Dec 2021 16:04:19 +0100
Subject: [PATCH 06/11] memory: Handle struct slab changes on Linux 5.17-rc1
and later
Since kernel commit d122019bf061 ("mm: Split slab into its own type"),
the struct slab is used for both SLAB and SLUB. Therefore, don't depend
on the non-presence of the struct slab to decide whether SLAB implementation
should be chosen and use the member variable "cpu_slab" of the struct
kmem_cache instead, it should be present only in SLUB.
Without the patch, crash fails to start with the error message:
crash: invalid structure member offset: kmem_cache_s_num
FILE: memory.c LINE: 9619 FUNCTION: kmem_cache_init()
Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Lianbo Jiang <lijiang@redhat.com>
---
memory.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/memory.c b/memory.c
index 86c02c132890..5af45fd7d834 100644
--- a/memory.c
+++ b/memory.c
@@ -576,7 +576,8 @@ vm_init(void)
STRUCT_SIZE_INIT(cpucache_s, "cpucache_s");
} else if (!VALID_STRUCT(kmem_slab_s) &&
- !VALID_STRUCT(slab_s) &&
+ !VALID_STRUCT(slab_s) &&
+ !MEMBER_EXISTS("kmem_cache", "cpu_slab") &&
(VALID_STRUCT(slab) || (vt->flags & SLAB_OVERLOAD_PAGE))) {
vt->flags |= PERCPU_KMALLOC_V2;
--
2.20.1

View File

@ -0,0 +1,75 @@
From fa0b6453a05c5600849e4e531c94594ed9c90270 Mon Sep 17 00:00:00 2001
From: Lianbo Jiang <lijiang@redhat.com>
Date: Mon, 17 Jan 2022 15:14:00 +0800
Subject: [PATCH 07/11] Move the initialization of "boot_date" to task_init()
The "boot_date" is initialized conditionally in the cmd_log(), which may
display incorrect "boot_date" value with the following command before
running the "log -T" command:
crash> help -k | grep date
date: Wed Dec 22 13:39:29 IST 2021
boot_date: Thu Jan 1 05:30:00 IST 1970
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
The calculation of "boot_date" depends on the HZ value, and the HZ will
be calculated in task_init() at the latest, so let's move it here.
Signed-off-by: Lianbo Jiang <lijiang@redhat.com>
---
kernel.c | 18 +++---------------
task.c | 10 ++++++++++
2 files changed, 13 insertions(+), 15 deletions(-)
diff --git a/kernel.c b/kernel.c
index a44a9c52ace0..9afddc0c918c 100644
--- a/kernel.c
+++ b/kernel.c
@@ -5026,21 +5026,9 @@ cmd_log(void)
if (argerrs)
cmd_usage(pc->curcmd, SYNOPSIS);
- if (msg_flags & SHOW_LOG_CTIME) {
- if (pc->flags & MINIMAL_MODE) {
- error(WARNING, "the option '-T' is not available in minimal mode\n");
- return;
- }
-
- if (kt->boot_date.tv_sec == 0) {
- ulonglong uptime_jiffies;
- ulong uptime_sec;
-
- get_uptime(NULL, &uptime_jiffies);
- uptime_sec = (uptime_jiffies)/(ulonglong)machdep->hz;
- kt->boot_date.tv_sec = kt->date.tv_sec - uptime_sec;
- kt->boot_date.tv_nsec = 0;
- }
+ if (msg_flags & SHOW_LOG_CTIME && pc->flags & MINIMAL_MODE) {
+ error(WARNING, "the option '-T' is not available in minimal mode\n");
+ return;
}
if (msg_flags & SHOW_LOG_AUDIT) {
diff --git a/task.c b/task.c
index 76e184ae70b1..263a8344dd94 100644
--- a/task.c
+++ b/task.c
@@ -692,6 +692,16 @@ task_init(void)
stack_overflow_check_init();
+ if (machdep->hz) {
+ ulonglong uptime_jiffies;
+ ulong uptime_sec;
+
+ get_uptime(NULL, &uptime_jiffies);
+ uptime_sec = (uptime_jiffies)/(ulonglong)machdep->hz;
+ kt->boot_date.tv_sec = kt->date.tv_sec - uptime_sec;
+ kt->boot_date.tv_nsec = 0;
+ }
+
tt->flags |= TASK_INIT_DONE;
}
--
2.20.1

View File

@ -0,0 +1,79 @@
From bbd5a5c1f5db3bde04628e75396155260333e53e Mon Sep 17 00:00:00 2001
From: Kazuhito Hagio <k-hagio-ab@nec.com>
Date: Wed, 19 Jan 2022 16:24:49 +0900
Subject: [PATCH 08/11] Remove ptype command from "ps -t" option to reduce
memory and time
With some vmlinux e.g. RHEL9 ones, the first execution of the gdb ptype
command heavily consumes memory and time. The "ps -t" option uses it in
start_time_timespec(), and it can be replaced with the crash macros.
This can reduce about 1.4 GB memory and 6 seconds time comsumption in
the following test:
$ echo "ps -t" | time crash vmlinux vmcore
Without the patch:
11.60user 0.43system 0:11.94elapsed 100%CPU (0avgtext+0avgdata 1837964maxresident)k
0inputs+400outputs (0major+413636minor)pagefaults 0swaps
With the patch:
5.40user 0.16system 0:05.46elapsed 101%CPU (0avgtext+0avgdata 417896maxresident)k
0inputs+384outputs (0major+41528minor)pagefaults 0swaps
Although the ptype command and similar ones cannot be fully removed,
but removing some of them will make the use of crash safer, especially
for an automatic crash reporter.
Signed-off-by: Kazuhito Hagio <k-hagio-ab@nec.com>
Signed-off-by: Lianbo Jiang <lijiang@redhat.com>
---
task.c | 25 +++++--------------------
1 file changed, 5 insertions(+), 20 deletions(-)
diff --git a/task.c b/task.c
index 263a8344dd94..a79ed0d96fb5 100644
--- a/task.c
+++ b/task.c
@@ -4662,8 +4662,6 @@ show_task_times(struct task_context *tcp, ulong flags)
static int
start_time_timespec(void)
{
- char buf[BUFSIZE];
-
switch(tt->flags & (TIMESPEC | NO_TIMESPEC | START_TIME_NSECS))
{
case TIMESPEC:
@@ -4677,24 +4675,11 @@ start_time_timespec(void)
tt->flags |= NO_TIMESPEC;
- open_tmpfile();
- sprintf(buf, "ptype struct task_struct");
- if (!gdb_pass_through(buf, NULL, GNU_RETURN_ON_ERROR)) {
- close_tmpfile();
- return FALSE;
- }
-
- rewind(pc->tmpfile);
- while (fgets(buf, BUFSIZE, pc->tmpfile)) {
- if (strstr(buf, "start_time;")) {
- if (strstr(buf, "struct timespec")) {
- tt->flags &= ~NO_TIMESPEC;
- tt->flags |= TIMESPEC;
- }
- }
- }
-
- close_tmpfile();
+ if (VALID_MEMBER(task_struct_start_time) &&
+ STREQ(MEMBER_TYPE_NAME("task_struct", "start_time"), "timespec")) {
+ tt->flags &= ~NO_TIMESPEC;
+ tt->flags |= TIMESPEC;
+ }
if ((tt->flags & NO_TIMESPEC) && (SIZE(task_struct_start_time) == 8)) {
tt->flags &= ~NO_TIMESPEC;
--
2.20.1

View File

@ -0,0 +1,150 @@
From d52cccfaa96ed6f61ff9d53da88715296e31db80 Mon Sep 17 00:00:00 2001
From: Tao Liu <ltao@redhat.com>
Date: Fri, 21 Jan 2022 13:43:09 +0800
Subject: [PATCH 09/11] Improve the ps performance for vmcores with large
number of threads
Previously, the ps command will iterate over all threads which
have the same tgid, to accumulate their rss value, in order to
get a thread/process's final rss value as part of the final output.
For non-live systems, the rss accumulation values are identical for
threads which have the same tgid, so there is no need to do the
iteration and accumulation repeatly, thus a lot of readmem calls are
skipped. Otherwise it will be the performance bottleneck if the
vmcores have a large number of threads.
In this patch, the rss accumulation value will be stored in a cache,
next time a thread with the same tgid will take it directly without
the iteration.
For example, we can monitor the performance issue when a vmcore has
~65k processes, most of which are threads for several specific
processes. Without the patch, it will take ~7h for ps command
to finish. With the patch, ps command will finish in 1min.
Signed-off-by: Tao Liu <ltao@redhat.com>
Signed-off-by: Lianbo Jiang <lijiang@redhat.com>
---
defs.h | 1 +
memory.c | 70 +++++++++++++++++++++++++++++++-------------------------
task.c | 1 +
3 files changed, 41 insertions(+), 31 deletions(-)
diff --git a/defs.h b/defs.h
index 41b6cbc6cc85..77e76f27cddb 100644
--- a/defs.h
+++ b/defs.h
@@ -830,6 +830,7 @@ struct task_context { /* context stored for each task */
struct tgid_context { /* tgid and task stored for each task */
ulong tgid;
ulong task;
+ long rss_cache;
};
struct task_table { /* kernel/local task table data */
diff --git a/memory.c b/memory.c
index 5af45fd7d834..e80c59ea4534 100644
--- a/memory.c
+++ b/memory.c
@@ -4665,7 +4665,7 @@ void
get_task_mem_usage(ulong task, struct task_mem_usage *tm)
{
struct task_context *tc;
- long rss = 0;
+ long rss = 0, rss_cache = 0;
BZERO(tm, sizeof(struct task_mem_usage));
@@ -4730,38 +4730,46 @@ get_task_mem_usage(ulong task, struct task_mem_usage *tm)
(last->tgid == (last + 1)->tgid))
last++;
- while (first <= last)
- {
- /* count 0 -> filepages */
- if (!readmem(first->task +
- OFFSET(task_struct_rss_stat) +
- OFFSET(task_rss_stat_count), KVADDR,
- &sync_rss,
- sizeof(int),
- "task_struct rss_stat MM_FILEPAGES",
- RETURN_ON_ERROR))
- continue;
-
- rss += sync_rss;
-
- /* count 1 -> anonpages */
- if (!readmem(first->task +
- OFFSET(task_struct_rss_stat) +
- OFFSET(task_rss_stat_count) +
- sizeof(int),
- KVADDR, &sync_rss,
- sizeof(int),
- "task_struct rss_stat MM_ANONPAGES",
- RETURN_ON_ERROR))
- continue;
-
- rss += sync_rss;
-
- if (first == last)
- break;
- first++;
+ /*
+ * Using rss cache for dumpfile is more beneficial than live debug
+ * because its value never changes in dumpfile.
+ */
+ if (ACTIVE() || last->rss_cache == UNINITIALIZED) {
+ while (first <= last)
+ {
+ /* count 0 -> filepages */
+ if (!readmem(first->task +
+ OFFSET(task_struct_rss_stat) +
+ OFFSET(task_rss_stat_count), KVADDR,
+ &sync_rss,
+ sizeof(int),
+ "task_struct rss_stat MM_FILEPAGES",
+ RETURN_ON_ERROR))
+ continue;
+
+ rss_cache += sync_rss;
+
+ /* count 1 -> anonpages */
+ if (!readmem(first->task +
+ OFFSET(task_struct_rss_stat) +
+ OFFSET(task_rss_stat_count) +
+ sizeof(int),
+ KVADDR, &sync_rss,
+ sizeof(int),
+ "task_struct rss_stat MM_ANONPAGES",
+ RETURN_ON_ERROR))
+ continue;
+
+ rss_cache += sync_rss;
+
+ if (first == last)
+ break;
+ first++;
+ }
+ last->rss_cache = rss_cache;
}
+ rss += last->rss_cache;
tt->last_tgid = last;
}
}
diff --git a/task.c b/task.c
index a79ed0d96fb5..864c838637ee 100644
--- a/task.c
+++ b/task.c
@@ -2947,6 +2947,7 @@ add_context(ulong task, char *tp)
tg = tt->tgid_array + tt->running_tasks;
tg->tgid = *tgid_addr;
tg->task = task;
+ tg->rss_cache = UNINITIALIZED;
if (do_verify && !verify_task(tc, do_verify)) {
error(INFO, "invalid task address: %lx\n", tc->task);
--
2.20.1

View File

@ -0,0 +1,59 @@
From 1a1fd21c625cb2ca335e626eb50426f13c4160f7 Mon Sep 17 00:00:00 2001
From: Kazuhito Hagio <k-hagio-ab@nec.com>
Date: Wed, 26 Jan 2022 06:07:00 +0000
Subject: [PATCH 10/11] arm64: Fix segfault by "bt" command with offline cpus
Currently on arm64, NT_PRSTATUS notes in dumpfile are not mapped to
online cpus and machine_specific->panic_task_regs correctly. As a
result, the "bt" command can cause a segmentation fault.
crash> bt -c 0
PID: 0 TASK: ffff8000117fa240 CPU: 0 COMMAND: "swapper/0"
Segmentation fault (core dumped)
To fix this,
1) make map_cpus_to_prstatus_kdump_cmprs() map the notes to
dd->nt_prstatus_percpu also on arm64, and
2) move arm64_get_crash_notes() to machdep_init(POST_INIT) in order
to apply the mapping to machine_specific->panic_task_regs.
Resolves: https://github.com/crash-utility/crash/issues/105
Reported-by: xuchunmei000 <xuchunmei@linux.alibaba.com>
Signed-off-by: Kazuhito Hagio <k-hagio-ab@nec.com>
Tested-by: David Wysochanski <dwysocha@redhat.com>
Signed-off-by: Lianbo Jiang <lijiang@redhat.com>
---
arm64.c | 2 +-
diskdump.c | 3 +--
2 files changed, 2 insertions(+), 3 deletions(-)
diff --git a/arm64.c b/arm64.c
index 23c3d75d85aa..4f2c2b5104a1 100644
--- a/arm64.c
+++ b/arm64.c
@@ -472,7 +472,7 @@ arm64_init(int when)
arm64_stackframe_init();
break;
- case POST_VM:
+ case POST_INIT:
/*
* crash_notes contains machine specific information about the
* crash. In particular, it contains CPU registers at the time
diff --git a/diskdump.c b/diskdump.c
index 112f769f8949..690b42443ed2 100644
--- a/diskdump.c
+++ b/diskdump.c
@@ -111,8 +111,7 @@ map_cpus_to_prstatus_kdump_cmprs(void)
if (pc->flags2 & QEMU_MEM_DUMP_COMPRESSED) /* notes exist for all cpus */
goto resize_note_pointers;
- if (!(online = get_cpus_online()) || (online == kt->cpus) ||
- machine_type("ARM64"))
+ if (!(online = get_cpus_online()) || (online == kt->cpus))
goto resize_note_pointers;
if (CRASHDEBUG(1))
--
2.20.1

View File

@ -0,0 +1,89 @@
From 86446eaba408807e00cf2310d5748aa6b7511284 Mon Sep 17 00:00:00 2001
From: Kazuhito Hagio <k-hagio-ab@nec.com>
Date: Wed, 2 Feb 2022 02:14:56 +0000
Subject: [PATCH 11/11] Fix for "kmem -s|-S" and "bt -F[F]" on Linux 5.17-rc1
Since the following kernel commits split slab info from struct page
into struct slab, crash cannot get several slab related offsets from
struct page.
d122019bf061 ("mm: Split slab into its own type")
07f910f9b729 ("mm: Remove slab from struct page")
Without the patch, "kmem -s|-S" and "bt -F[F]" options cannot work
correctly with the following errors:
crash> kmem -s kmem_cache
CACHE OBJSIZE ALLOCATED TOTAL SLABS SSIZE NAME
kmem: page_to_nid: invalid page: ffff9454afc35020
kmem: kmem_cache: cannot gather relevant slab data
ffff945140042000 216 ? ? ? 8k kmem_cache
crash> bt -F
...
bt: invalid structure member offset: page_slab
FILE: memory.c LINE: 9477 FUNCTION: vaddr_to_kmem_cache()
Signed-by: Kazuhito Hagio <k-hagio-ab@nec.com>
Signed-off-by: Lianbo Jiang <lijiang@redhat.com>
---
memory.c | 13 +++++++++++++
1 file changed, 13 insertions(+)
diff --git a/memory.c b/memory.c
index e80c59ea4534..8448ddc3a16c 100644
--- a/memory.c
+++ b/memory.c
@@ -421,6 +421,8 @@ vm_init(void)
MEMBER_OFFSET_INIT(page_prev, "page", "prev");
if (INVALID_MEMBER(page_next))
ANON_MEMBER_OFFSET_INIT(page_next, "page", "next");
+ if (INVALID_MEMBER(page_next))
+ MEMBER_OFFSET_INIT(page_next, "slab", "next");
MEMBER_OFFSET_INIT(page_list, "page", "list");
if (VALID_MEMBER(page_list)) {
@@ -747,11 +749,15 @@ vm_init(void)
MEMBER_OFFSET_INIT(kmem_cache_random, "kmem_cache", "random");
MEMBER_OFFSET_INIT(kmem_cache_cpu_freelist, "kmem_cache_cpu", "freelist");
MEMBER_OFFSET_INIT(kmem_cache_cpu_page, "kmem_cache_cpu", "page");
+ if (INVALID_MEMBER(kmem_cache_cpu_page))
+ MEMBER_OFFSET_INIT(kmem_cache_cpu_page, "kmem_cache_cpu", "slab");
MEMBER_OFFSET_INIT(kmem_cache_cpu_node, "kmem_cache_cpu", "node");
MEMBER_OFFSET_INIT(kmem_cache_cpu_partial, "kmem_cache_cpu", "partial");
MEMBER_OFFSET_INIT(page_inuse, "page", "inuse");
if (INVALID_MEMBER(page_inuse))
ANON_MEMBER_OFFSET_INIT(page_inuse, "page", "inuse");
+ if (INVALID_MEMBER(page_inuse))
+ MEMBER_OFFSET_INIT(page_inuse, "slab", "inuse");
MEMBER_OFFSET_INIT(page_offset, "page", "offset");
if (INVALID_MEMBER(page_offset))
ANON_MEMBER_OFFSET_INIT(page_offset, "page", "offset");
@@ -763,6 +769,9 @@ vm_init(void)
if (INVALID_MEMBER(page_slab))
ANON_MEMBER_OFFSET_INIT(page_slab, "page", "slab_cache");
}
+ if (INVALID_MEMBER(page_slab))
+ MEMBER_OFFSET_INIT(page_slab, "slab", "slab_cache");
+
MEMBER_OFFSET_INIT(page_slab_page, "page", "slab_page");
if (INVALID_MEMBER(page_slab_page))
ANON_MEMBER_OFFSET_INIT(page_slab_page, "page", "slab_page");
@@ -772,10 +781,14 @@ vm_init(void)
MEMBER_OFFSET_INIT(page_freelist, "page", "freelist");
if (INVALID_MEMBER(page_freelist))
ANON_MEMBER_OFFSET_INIT(page_freelist, "page", "freelist");
+ if (INVALID_MEMBER(page_freelist))
+ MEMBER_OFFSET_INIT(page_freelist, "slab", "freelist");
if (INVALID_MEMBER(kmem_cache_objects)) {
MEMBER_OFFSET_INIT(kmem_cache_oo, "kmem_cache", "oo");
/* NOTE: returns offset of containing bitfield */
ANON_MEMBER_OFFSET_INIT(page_objects, "page", "objects");
+ if (INVALID_MEMBER(page_objects))
+ ANON_MEMBER_OFFSET_INIT(page_objects, "slab", "objects");
}
if (VALID_MEMBER(kmem_cache_node)) {
ARRAY_LENGTH_INIT(len, NULL, "kmem_cache.node", NULL, 0);
--
2.20.1

View File

@ -4,7 +4,7 @@
Summary: Kernel analysis utility for live systems, netdump, diskdump, kdump, LKCD or mcore dumpfiles Summary: Kernel analysis utility for live systems, netdump, diskdump, kdump, LKCD or mcore dumpfiles
Name: crash Name: crash
Version: 7.3.1 Version: 7.3.1
Release: 3%{?dist} Release: 5%{?dist}
License: GPLv3 License: GPLv3
Group: Development/Debuggers Group: Development/Debuggers
Source0: https://github.com/crash-utility/crash/archive/crash-%{version}.tar.gz Source0: https://github.com/crash-utility/crash/archive/crash-%{version}.tar.gz
@ -21,6 +21,17 @@ Patch1: rhel8_build.patch
Patch2: rhel8_freepointer.patch Patch2: rhel8_freepointer.patch
Patch3: 0001-arm64-Support-overflow-stack-panic.patch Patch3: 0001-arm64-Support-overflow-stack-panic.patch
Patch4: 0002-defs.h-fix-breakage-of-compatibility-of-struct-symbo.patch Patch4: 0002-defs.h-fix-breakage-of-compatibility-of-struct-symbo.patch
Patch5: 0001-Fix-pvops-Xen-detection-for-arm-machine.patch
Patch6: 0002-Handle-blk_mq_ctx-member-changes-for-kernels-5.16-rc.patch
Patch7: 0003-Fix-for-timer-r-option-to-display-all-the-per-CPU-cl.patch
Patch8: 0004-Fix-for-bt-v-option-to-display-the-stack-end-address.patch
Patch9: 0005-Fix-for-HZ-calculation-on-Linux-5.14-and-later.patch
Patch10: 0006-memory-Handle-struct-slab-changes-on-Linux-5.17-rc1-.patch
Patch11: 0007-Move-the-initialization-of-boot_date-to-task_init.patch
Patch12: 0008-Remove-ptype-command-from-ps-t-option-to-reduce-memo.patch
Patch13: 0009-Improve-the-ps-performance-for-vmcores-with-large-nu.patch
Patch14: 0010-arm64-Fix-segfault-by-bt-command-with-offline-cpus.patch
Patch15: 0011-Fix-for-kmem-s-S-and-bt-F-F-on-Linux-5.17-rc1.patch
%description %description
The core analysis suite is a self-contained tool that can be used to The core analysis suite is a self-contained tool that can be used to
@ -46,6 +57,17 @@ offered by Mission Critical Linux, or the LKCD kernel patch.
%patch2 -p1 -b rhel8_freepointer.patch %patch2 -p1 -b rhel8_freepointer.patch
%patch3 -p1 %patch3 -p1
%patch4 -p1 %patch4 -p1
%patch5 -p1
%patch6 -p1
%patch7 -p1
%patch8 -p1
%patch9 -p1
%patch10 -p1
%patch11 -p1
%patch12 -p1
%patch13 -p1
%patch14 -p1
%patch15 -p1
%build %build
cp %{SOURCE1} . cp %{SOURCE1} .
@ -76,6 +98,14 @@ rm -rf %{buildroot}
%{_includedir}/* %{_includedir}/*
%changelog %changelog
* Tue Feb 08 2022 Lianbo Jiang <lijiang@redhat.com> - 7.3.1-5
- Rebuild for osci badfuncs issue
* Mon Feb 07 2022 Lianbo Jiang <lijiang@redhat.com> - 7.3.1-4
- Fix segfault on aarch64 for "bt -a|-c" command
- Fix HZ calculation on Linux 5.14 and later
- Fix for "timer -r" option to display all the per-CPU clocks
* Mon Dec 13 2021 Lianbo Jiang <lijiang@redhat.com> - 7.3.1-3 * Mon Dec 13 2021 Lianbo Jiang <lijiang@redhat.com> - 7.3.1-3
- Fix segmentation fault caused by crash extension modules - Fix segmentation fault caused by crash extension modules
- Support the overflow stack exception handling on aarch64 - Support the overflow stack exception handling on aarch64