Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Helge Deller <deller@gmx.de>
To:     linux-s390@vger.kernel.org, Josh Triplett <josh@joshtriplett.org>,
        x86@kernel.org, linux-snps-arc@lists.infradead.org,
        linux-fsdevel@vger.kernel.org, linux-parisc@vger.kernel.org,
        linux-kernel@vger.kernel.org, linuxppc-dev@lists.ozlabs.org,
        linux-arm-kernel@lists.infradead.org
Subject: [PATCH v3 0/4] Dump command line of faulting process to syslog
Date:   Mon,  8 Aug 2022 15:09:13 +0200
Message-Id: <20220808130917.30760-1-deller@gmx.de>
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

This patch series dumps the command line (including the program parameters=
) of
a faulting process to the syslog.

The motivation for this patch is that it's sometimes quite hard to find ou=
t and
annoying to not know which program *exactly* faulted when looking at the s=
yslog.

For example, a dump on parisc shows:
   do_page_fault() command=3D'cc1' type=3D15 address=3D0x00000000 in libc-=
2.33.so[f6abb000+184000]

-> We see the "cc1" compiler crashed, but it would be useful to know which=
 file was compiled.
With this patch you will see that cc1 crashed while compiling some haskell=
 code:

   cc1[13472] cmdline: /usr/lib/gcc/hppa-linux-gnu/12/cc1 -quiet @/tmp/ccR=
kFSfY -imultilib . -imultiarch hppa-linux-gnu -D USE_MINIINTERPRETER -D NO=
_REGS -D _HPUX_SOURCE -D NOSMP -D THREADED_RTS -include /build/ghc/ghc-9.0=
.2/includes/dist-install/build/ghcversion.h -iquote compiler/GHC/Iface -qu=
iet -dumpdir /tmp/ghc13413_0/ -dumpbase ghc_5.hc -dumpbase-ext .hc -O -Wim=
plicit -fno-PIC -fwrapv -fno-builtin -fno-strict-aliasing -o /tmp/ghc13413=
_0/ghc_5.s

Another example are the glibc testcases which always segfault in "ld.so.1"=
 with no other info:

   do_page_fault() command=3D'ld.so.1' type=3D15 address=3D0x565921d8 in l=
ibc.so[f7339000+1bb000]

-> With the patch you can see it was the "tst-safe-linking-malloc-hugetlb1=
" testcase:

   ld.so.1[1151] cmdline: /home/gnu/glibc/objdir/elf/ld.so.1 --library-pat=
h /home/gnu/glibc/objdir:/home/gnu/glibc/objdir/math:/home/gnu/
        /home/gnu/glibc/objdir/malloc/tst-safe-linking-malloc-hugetlb1

An example of a typical x86 fault shows up as:
   crash[2326]: segfault at 0 ip 0000561a7969c12e sp 00007ffe97a05630 erro=
r 6 in crash[561a7969c000+1000]
   Code: 68 ff ff ff c6 05 19 2f 00 00 01 5d c3 0f 1f 80 00 00 00 00 c3 0f=
 1f 80 00 00 ...

-> with this patch you now see the whole command line:
   crash[2326] cmdline: ./crash test_write_to_page_0

The patches are relatively small, and reuse functions which are used
to create the output for the /proc/<pid>/cmdline files.

The relevant changes are in patches #1 and #2.
Patch #3 adds the cmdline dump on x86.
Patch #4 drops code from arc which now becomes unnecessary as this is done=
 by generic code.

Helge

=2D--

Changes in v3:
- require task to be locked by caller, noticed by kernel test robot
- add parameter names in header files, noticed by kernel test robot

Changes in v2:
- Don't dump all or parts of the commandline depending on the
  kptr_restrict sysctl value (suggested by Josh Triplett).
- Patch sent to more arch mailing lists

Helge Deller (4):
  proc: Add get_task_cmdline_kernel() function
  lib/dump_stack: Add dump_stack_print_cmdline() and wire up in
    dump_stack_print_info()
  x86/fault: Dump command line of faulting process to syslog
  arc: Use generic dump_stack_print_cmdline() implementation

 arch/arc/kernel/troubleshoot.c | 24 -----------
 arch/x86/mm/fault.c            |  2 +
 fs/proc/base.c                 | 74 +++++++++++++++++++++++-----------
 include/linux/printk.h         |  5 +++
 include/linux/proc_fs.h        |  5 +++
 lib/dump_stack.c               | 34 ++++++++++++++++
 6 files changed, 97 insertions(+), 47 deletions(-)

=2D-
2.37.1