Skip to contentSkip to conversation

3.10.9: Oops at elf_core_dump()

PATCH In review Martin MOKREJŠ 3 participants 9 messages Aug 29, 2013 – Sep 1, 2013

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

Hi,
I just got this stacktrace. Not sure whom to send it, poking throu MAINTAINERS
file and looking for ELF gave me nothing. ;-)

[105670.434336] BUG: unable to handle kernel NULL pointer dereference at           (null)
[105670.434366] IP: [<ffffffff812f7b42>] strlen+0x2/0x20
[105670.434385] PGD 18c8e5067 PUD 2b547e067 PMD 0 
[105670.434401] Oops: 0000 [#1] SMP 
[105670.434413] Modules linked in: iwldvm iwlwifi
[105670.434432] CPU: 0 PID: 7497 Comm: emerge Not tainted 3.10.9-default-pciehp #8
[105670.434451] Hardware name: Dell Inc. Vostro 3550/, BIOS A11 08/03/2012
[105670.434468] task: ffff88037df42f70 ti: ffff88018683c000 task.ti: ffff88018683c000
[105670.434487] RIP: 0010:[<ffffffff812f7b42>]  [<ffffffff812f7b42>] strlen+0x2/0x20
[105670.434509] RSP: 0018:ffff88018683d9f0  EFLAGS: 00010246
[105670.434523] RAX: 0000000000000000 RBX: 0000000000000003 RCX: ffff88037df42f70
[105670.434542] RDX: 00000000016e3610 RSI: 0000000000000000 RDI: 0000000000000000
[105670.434560] RBP: ffff88018683da08 R08: 0000000000000000 R09: 0000000000000000
[105670.434579] R10: 0000000000000000 R11: 0000000000000001 R12: ffff88018683db00
[105670.434598] R13: 00007ffffffff000 R14: 0000000000000004 R15: 0000000000000000
[105670.434617] FS:  00007f89b0989740(0000) GS:ffff88041d800000(0000) knlGS:0000000000000000
[105670.434637] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[105670.434652] CR2: 0000000000000000 CR3: 00000002b4c06000 CR4: 00000000000407f0
[105670.434671] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[105670.434690] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
[105670.434708] Stack:
[105670.434715]  ffffffff811bb0a3 0000000000000004 00000000000003d8 ffff88018683dc08
[105670.434738]  ffffffff811bbcbd ffffffff811bb913 0000000000000000 ffff88018683db28
[105670.434762]  ffff88037df42f70 ffff88018683ffff 0000000000000246 000f424200000000
[105670.434785] Call Trace:
[105670.434795]  [<ffffffff811bb0a3>] ? notesize.isra.11+0x13/0x30
[105670.434812]  [<ffffffff811bbcbd>] elf_core_dump+0xbfd/0x1570
[105670.434828]  [<ffffffff811bb913>] ? elf_core_dump+0x853/0x1570
[105670.434845]  [<ffffffff811c2bc5>] ? do_coredump+0xe25/0xff0
[105670.434861]  [<ffffffff810eb85d>] ? trace_hardirqs_on+0xd/0x10
[105670.434878]  [<ffffffff8116e0ff>] ? __sb_start_write+0xdf/0x1b0
[105670.434894]  [<ffffffff811c2bc5>] ? do_coredump+0xe25/0xff0
[105670.434911]  [<ffffffff81097209>] ? unshare_files+0x29/0xa0
[105670.434926]  [<ffffffff811c289c>] do_coredump+0xafc/0xff0
[105670.434943]  [<ffffffff810a63c8>] ? __sigqueue_free+0x38/0x40
[105670.434960]  [<ffffffff810a9961>] get_signal_to_deliver+0x1c1/0x5c0
[105670.434977]  [<ffffffff810a8721>] ? do_send_sig_info+0x61/0x90
[105670.434994]  [<ffffffff81002303>] do_signal+0x53/0x8e0
[105670.435008]  [<ffffffff810a8cd0>] ? kill_pgrp+0x60/0x60
[105670.435025]  [<ffffffff810c2b9e>] ? finish_task_switch+0x7e/0xe0
[105670.435043]  [<ffffffff81799750>] ? sysret_signal+0x5/0x47
[105670.435058]  [<ffffffff81002bef>] do_notify_resume+0x5f/0x70
[105670.435074]  [<ffffffff812fc71e>] ? trace_hardirqs_on_thunk+0x3a/0x3f
[105670.435092]  [<ffffffff817999e2>] int_signal+0x12/0x17
[105670.435106] Code: 48 89 e5 f6 82 40 c6 84 81 20 74 15 0f 1f 44 00 00 48 83 c0 01 0f b6 10 f6 82 40 c6 84 81 20 75 f0 5d c3 66 0f 1f 44 00 00 31 c0 <80> 3f 00 55 48 89 e5 74 11 48 89 f8 66 90 48 83 c0 01 80 38 00 
[105670.435238] RIP  [<ffffffff812f7b42>] strlen+0x2/0x20
[105670.435254]  RSP <ffff88018683d9f0>
[105670.435843] CR2: 0000000000000000
[105670.439699] ---[ end trace 9d67aee555e92d75 ]---

Martin

#1 lore raw
Quoted 10 lines · Martin MOKREJŠ

Hi,
I just got this stacktrace. Not sure whom to send it, poking throu MAINTAINERS
file and looking for ELF gave me nothing. ;-)

[105670.434336] BUG: unable to handle kernel NULL pointer dereference at           (null)
[105670.434366] IP: [<ffffffff812f7b42>] strlen+0x2/0x20
[105670.434385] PGD 18c8e5067 PUD 2b547e067 PMD 0 
[105670.434401] Oops: 0000 [#1] SMP 
[105670.434413] Modules linked in: iwldvm iwlwifi
[105670.434432] CPU: 0 PID: 7497 Comm: emerge Not tainted 3.10.9-default-pciehp #8

Is this reproducable?

thanks,

greg k-h

#2 lore raw

Got it for the first time. Actually, am doing something really unusual
(http://bugs.python.org/issue18843).

Am looking for an answer why I suffer memory corruption in python applicatuons. So I installed DUMA from http://duma.sourceforge.net and tried to recompile&reinstall failing python. In previous attempt it exited and per README instructions I increased the vm.max_map_count value.

# export LD_PRELOAD=/usr/lib64/libduma.so.0.0.0
# sysctl -w vm.max_map_count=1000000
# emerge dev-lang/python:2.7 
DUMA 2.5.15 (shared library, NO_LEAKDETECTION)
Copyright (C) 2006 Michael Eddington <[email protected]>
Copyright (C) 2002-2008 Hayati Ayguen <[email protected]>, Procitec GmbH
Copyright (C) 1987-1999 Bruce Perens <[email protected]>

* IMPORTANT: 11 news items need reading for repository 'gentoo'.
* Use eselect news to read news items.

* IMPORTANT: config file '5 (shared library, NO_LEAKDETECTION)
Copyright (C) 2006 Michael Eddington <[email protected]> Copyright (C) 2002-2008 Hayati Ayguen <[email protected]>, Procitec GmbH Copyright (C) 1987-1999 Bruce Perens <[email protected]>

' needs updating.
* See the CONFIGURATION FILES section of the emerge
* man page to learn how to update config files.
Calculating dependencies |
DUMA Aborting: mprotect() failed: Cannot allocate memory.
Check README section 'MEMORY USAGE AND EXECUTION SPEED'
if your (Linux) system may limit the number of different page mappings per process

[and it crashed, no ctrl+c working]

Sorry do not know what to say more. I just crashed teh kernel but except the Ooops it works so far. The core filesize is zero.
Martin

Quoted 18 lines · Greg KH
Quoted 10 lines · Martin MOKREJŠ

Hi,
I just got this stacktrace. Not sure whom to send it, poking throu MAINTAINERS
file and looking for ELF gave me nothing. ;-)

[105670.434336] BUG: unable to handle kernel NULL pointer dereference at           (null)
[105670.434366] IP: [<ffffffff812f7b42>] strlen+0x2/0x20
[105670.434385] PGD 18c8e5067 PUD 2b547e067 PMD 0 
[105670.434401] Oops: 0000 [#1] SMP 
[105670.434413] Modules linked in: iwldvm iwlwifi
[105670.434432] CPU: 0 PID: 7497 Comm: emerge Not tainted 3.10.9-default-pciehp #8

Is this reproducable?

thanks,

greg k-h

Signature · 7 lines
Martin Mokrejs, Ph.D.
Bioinformatics
Donovalska 1658
149 00 Prague
Czech Republic
http://www.iresite.org
http://www.iresite.org/~mmokrejs
#3 lore raw

So it happened again:

$ export LD_PRELOAD=/usr/lib64/libduma.so.0.0.0
$ python  memory-corruption-test.py
DUMA 2.5.15 (shared library, NO_LEAKDETECTION)
Copyright (C) 2006 Michael Eddington <[email protected]>
Copyright (C) 2002-2008 Hayati Ayguen <[email protected]>, Procitec GmbH
Copyright (C) 1987-1999 Bruce Perens <[email protected]>

DUMA 2.5.15 (shared library, NO_LEAKDETECTION)
Copyright (C) 2006 Michael Eddington <[email protected]> Copyright (C) 2002-2008 Hayati Ayguen <[email protected]>, Procitec GmbH Copyright (C) 1987-1999 Bruce Perens <[email protected]>

Finished one record
Finished one record
Finished one record
Finished one record
Finished one record
[cut]

DUMA Aborting: mprotect() failed: Cannot allocate memory.
Check README section 'MEMORY USAGE AND EXECUTION SPEED'
if your (Linux) system may limit the number of different page mappings per process
Fatal Python error: Illegal instruction

Current thread 0x00007fc9c4803740:
File "/usr/lib64/python2.7/site-packages/Bio/Blast/NCBIXML.py", line 106 in endElement
File "/mnt/1TB/var/tmp/portage/dev-lang/python-2.7.5-r2/work/Python-2.7.5/Modules/pyexpat.c", line 618 in EndElement File "/usr/lib64/python2.7/site-packages/Bio/Blast/NCBIXML.py", line 654 in parse
File "memory-corruption-test.py", line 55 in doparse
File "memory-corruption-test.py", line 104 in main
File "memory-corruption-test.py", line 109 in <module>

The stacktrace is little different but ... I think I need to find what resource to so that duma can keep running watching the python binary.

[112567.987073] BUG: unable to handle kernel NULL pointer dereference at           (null)
[112567.987684] IP: [<ffffffff812f7b42>] strlen+0x2/0x20
[112567.988282] PGD 28be2c067 PUD 3a7744067 PMD 0 
[112567.988879] Oops: 0000 [#2] SMP 
[112567.989468] Modules linked in: iwldvm iwlwifi
[112567.990057] CPU: 0 PID: 8822 Comm: python2.7 Tainted: G      D      3.10.9-default-pciehp #8
[112567.990655] Hardware name: Dell Inc. Vostro 3550/, BIOS A11 08/03/2012
[112567.991249] task: ffff8803b5eb0fd0 ti: ffff8803b36d4000 task.ti: ffff8803b36d4000
[112567.991845] RIP: 0010:[<ffffffff812f7b42>]  [<ffffffff812f7b42>] strlen+0x2/0x20
[112567.992443] RSP: 0018:ffff8803b36d59f0  EFLAGS: 00010246
[112567.993039] RAX: 0000000000000000 RBX: 0000000000000003 RCX: ffff8803b5eb0fd0
[112567.993643] RDX: 00000000016e3610 RSI: 0000000000000000 RDI: 0000000000000000
[112567.994249] RBP: ffff8803b36d5a08 R08: 00000000fffffffa R09: 0000000000000000
[112567.994854] R10: 0000000000000000 R11: 0000000000000000 R12: ffff8803b36d5b00
[112567.995459] R13: 00007ffffffff000 R14: 0000000000000004 R15: 0000000000000000
[112567.996060] FS:  00007fc9c4803740(0000) GS:ffff88041d800000(0000) knlGS:0000000000000000
[112567.996664] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[112567.997268] CR2: 0000000000000000 CR3: 00000003a3b84000 CR4: 00000000000407f0
[112567.997884] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[112567.998495] DR3: 0000000000000000 DR6: 00000000ffff4ff0 DR7: 0000000000000400
[112567.999100] Stack:
[112567.999698]  ffffffff811bb0a3 ffff8803b36d5e60 00000000000003d8 ffff8803b36d5c08
[112568.000315]  ffffffff811bbcbd ffffffff811bb913 0000000000000002 0000000000000000
[112568.000931]  ffff8803b5eb0fd0 ffff8803b36dffff 0000000000000246 000f424200000000
[112568.001548] Call Trace:
[112568.002156]  [<ffffffff811bb0a3>] ? notesize.isra.11+0x13/0x30
[112568.002774]  [<ffffffff811bbcbd>] elf_core_dump+0xbfd/0x1570
[112568.003392]  [<ffffffff811bb913>] ? elf_core_dump+0x853/0x1570
[112568.004012]  [<ffffffff81097209>] ? unshare_files+0x29/0xa0
[112568.004629]  [<ffffffff811c289c>] do_coredump+0xafc/0xff0
[112568.005247]  [<ffffffff810a63c8>] ? __sigqueue_free+0x38/0x40
[112568.005865]  [<ffffffff810a9961>] get_signal_to_deliver+0x1c1/0x5c0
[112568.006488]  [<ffffffff810b5d90>] ? pid_vnr+0x30/0x30
[112568.007108]  [<ffffffff81002303>] do_signal+0x53/0x8e0
[112568.007725]  [<ffffffff81002bef>] do_notify_resume+0x5f/0x70
[112568.008342]  [<ffffffff812fc71e>] ? trace_hardirqs_on_thunk+0x3a/0x3f
[112568.008965]  [<ffffffff817999e2>] int_signal+0x12/0x17
[112568.009583] Code: 48 89 e5 f6 82 40 c6 84 81 20 74 15 0f 1f 44 00 00 48 83 c0 01 0f b6 10 f6 82 40 c6 84 81 20 75 f0 5d c3 66 0f 1f 44 00 00 31 c0 <80> 3f 00 55 48 89 e5 74 11 48 89 f8 66 90 48 83 c0 01 80 38 00 
[112568.011042] RIP  [<ffffffff812f7b42>] strlen+0x2/0x20
[112568.011748]  RSP <ffff8803b36d59f0>
[112568.012445] CR2: 0000000000000000
[112568.013155] ---[ end trace 9d67aee555e92d76 ]---
Quoted 65 lines · Martin MOKREJŠ

Got it for the first time. Actually, am doing something really unusual
(http://bugs.python.org/issue18843).

Am looking for an answer why I suffer memory corruption in python applicatuons. So I installed DUMA from http://duma.sourceforge.net and tried to recompile&reinstall failing python. In previous attempt it exited and per README instructions I increased the vm.max_map_count value.

# export LD_PRELOAD=/usr/lib64/libduma.so.0.0.0
# sysctl -w vm.max_map_count=1000000
# emerge dev-lang/python:2.7 
DUMA 2.5.15 (shared library, NO_LEAKDETECTION)
Copyright (C) 2006 Michael Eddington <[email protected]>
Copyright (C) 2002-2008 Hayati Ayguen <[email protected]>, Procitec GmbH
Copyright (C) 1987-1999 Bruce Perens <[email protected]>

* IMPORTANT: 11 news items need reading for repository 'gentoo'.
* Use eselect news to read news items.

* IMPORTANT: config file '5 (shared library, NO_LEAKDETECTION)
Copyright (C) 2006 Michael Eddington <[email protected]> Copyright (C) 2002-2008 Hayati Ayguen <[email protected]>, Procitec GmbH Copyright (C) 1987-1999 Bruce Perens <[email protected]>

' needs updating.
* See the CONFIGURATION FILES section of the emerge
* man page to learn how to update config files.
Calculating dependencies |
DUMA Aborting: mprotect() failed: Cannot allocate memory.
Check README section 'MEMORY USAGE AND EXECUTION SPEED'
if your (Linux) system may limit the number of different page mappings per process

[and it crashed, no ctrl+c working]

Sorry do not know what to say more. I just crashed teh kernel but except the Ooops it works so far. The core filesize is zero.
Martin

Quoted 18 lines · Greg KH
Quoted 10 lines · Martin MOKREJŠ

Hi,
I just got this stacktrace. Not sure whom to send it, poking throu MAINTAINERS
file and looking for ELF gave me nothing. ;-)

[105670.434336] BUG: unable to handle kernel NULL pointer dereference at           (null)
[105670.434366] IP: [<ffffffff812f7b42>] strlen+0x2/0x20
[105670.434385] PGD 18c8e5067 PUD 2b547e067 PMD 0 
[105670.434401] Oops: 0000 [#1] SMP 
[105670.434413] Modules linked in: iwldvm iwlwifi
[105670.434432] CPU: 0 PID: 7497 Comm: emerge Not tainted 3.10.9-default-pciehp #8

Is this reproducable?

thanks,

greg k-h

Signature · 7 lines
Martin Mokrejs, Ph.D.
Bioinformatics
Donovalska 1658
149 00 Prague
Czech Republic
http://www.iresite.org
http://www.iresite.org/~mmokrejs
#4 lore raw
In reply to Greg KH
Quoted 5 lines · Greg KH
Quoted 4 lines · Martin MOKREJŠ

Hi,
I just got this stacktrace. Not sure whom to send it, poking throu MAINTAINERS
file and looking for ELF gave me nothing. ;-)

Quoted 8 lines
Quoted 6 lines
[105670.434336] BUG: unable to handle kernel NULL pointer dereference at           (null)
[105670.434366] IP: [<ffffffff812f7b42>] strlen+0x2/0x20
[105670.434385] PGD 18c8e5067 PUD 2b547e067 PMD 0 
[105670.434401] Oops: 0000 [#1] SMP 
[105670.434413] Modules linked in: iwldvm iwlwifi
[105670.434432] CPU: 0 PID: 7497 Comm: emerge Not tainted 3.10.9-default-pciehp #8

Is this reproducable?

Yes, and here is my analysis:

fill_files_note(&info->files) exits early because of too many VM areas, or due to memory pressure (vmalloc failing), leaving a NULL string in info->files, letting notesize() crash on it.

as root do:

echo 300000 > /proc/sys/vm/max_map_count

then, as a regular user:

ulimit -c unlimited
gcc prog.c -o prog
./prog

prog.c:
-------
int main(int argc, char *argv[])

{
	char *p, *t;
	int i;

	p = (void *)0x444400000000;

	for (i = 0; i < 200000; i++) {
		t = mmap(p, 0x1000, PROT_READ | PROT_WRITE | PROT_EXEC,
		     MAP_FIXED | MAP_ANONYMOUS | MAP_PRIVATE,
		     -1, 0);
		p = &p[0x2000];
	}

	*((char *)0x0) = 0;

	return 0;
}

And the result:

user@guestvm:~$ c[  380.520865] BUG: unable to handle kernel NULL pointer dereference at 0000000000000086
[  380.523196] IP: [<ffffffff812ee180>] strim+0x80/0x80
[  380.524477] PGD 3abc6067 PUD 3c7b4067 PMD 0 
[  380.525974] Oops: 0000 [#1] SMP 

Entering kdb (current=0xffff880033ee8000, pid 1716) on processor 0 Oops: (null) due to oops @ 0xffffffff812ee180

dCPU: 0 PID: 1716 Comm: a.out Not tainted 3.10.9-mod-nodbg+ #1
dHardware name: Bochs Bochs, BIOS Bochs 01/01/2011
dtask: ffff880033ee8000 ti: ffff880034eec000 task.ti: ffff880034eec000
dRIP: 0010:[<ffffffff812ee180>]  [<ffffffff812ee180>] strim+0x80/0x80
dRSP: 0000:ffff880034eeda30  EFLAGS: 00010292
dRAX: 0000000000c353c0 RBX: 00000000ffff8800 RCX: ffff880033ee8000
dRDX: 0000000000493f78 RSI: 00000000ffff8800 RDI: 0000000000000086
dRBP: ffff880034eeda48 R08: 00000000fffffffd R09: 0000000000000000
dR10: 0000000000000000 R11: ffffffff812e6c4e R12: ffff880034eedb78
dR13: 00007ffffffff000 R14: 0000000000000000 R15: ffffffff81802708
dFS:  00007fd9f8bf7740(0000) GS:ffff88003f200000(0000) knlGS:0000000000000000
dCS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
dCR2: 0000000000000086 CR3: 000000003ae91000 CR4: 00000000001407f0
dDR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
dDR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400

dStack:
ffffffff811d29a5 ffff880033ee8000 00000000000003d8 ffff880034eedc38 ffffffff811d3579 ffff880034eeda88 ffffffff8108625a ffff88003a522300 ffff880033ee8000 0000000000493f78 0000ffff00030d51 ffff880000493f78
dCall Trace:
d [<ffffffff811d29a5>] ? notesize.isra.9+0x15/0x30
d [<ffffffff811d3579>] elf_core_dump+0xbb9/0x1460
d [<ffffffff8108625a>] ? finish_task_switch+0x4a/0x100
d [<ffffffff8164054d>] ? schedule+0x5d/0x60
d [<ffffffff81084a23>] ? __wake_up+0x53/0x70
d [<ffffffff811dbaee>] do_coredump+0xb8e/0xef0
d [<ffffffff8106632d>] ? __sigqueue_free+0x3d/0x50
d [<ffffffff81069bcf>] get_signal_to_deliver+0x53f/0x5d0 d [<ffffffff81637c03>] ? bad_area+0x44/0x4c
d [<ffffffff810123c7>] do_signal+0x57/0x570
d [<ffffffff8108cf0d>] ? __dequeue_entity+0x3d/0x50
d [<ffffffff81637eda>] ? printk+0x61/0x63
d [<ffffffff8108625a>] ? finish_task_switch+0x4a/0x100
d [<ffffffff8164030b>] ? __schedule+0x6bb/0x800
d [<ffffffff8101291e>] do_notify_resume+0x3e/0x90
d [<ffffffff81641b3c>] retint_signal+0x48/0x8c

On some systems the requirements for max_map_count are really large, so we can't avoid it. So, binfmt_elf.c should be fixed.

Signature · 1 line
Dan Aloni
#5 lore raw

PATCH 1/1 Prevent a coredump with a large vm_map_count from Oopsing

A high setting of max_map_count, and a process core-dumping with a large enough vm_map_count could result in an NT_FILE note not being written, and the kernel crashing immediately later because it has assumed otherwise.

Reproduction of the bug described here:

    https://lkml.org/lkml/2013/8/30/50

Issue originating in 2aa362c49 (from Oct 4, 2012).

This patch make that section optional in that case. fill_files_note() should signify the error, and also let the info struct in elf_core_dump() be zero-initialized so that we can check for the optionally written note.

Cc'ed original signers.

Cc'ed Al Viro because it is trivially relies on his linux-next tree changes.

Cc Al Viro <[email protected]>
Cc Denys Vlasenko <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 fs/binfmt_elf.c | 33 +++++++++++++++++++++------------
 1 file changed, 21 insertions(+), 12 deletions(-)
fs/binfmt_elf.c +21 -12
index dc82279..e1a323a 100644
@@ -1429,7 +1429,7 @@ static void fill_siginfo_note(struct memelfnote *note, user_siginfo_t *csigdata,
  *   long file_ofs
  * followed by COUNT filenames in ASCII: "FILE1" NUL "FILE2" NUL...
  */
-static void fill_files_note(struct memelfnote *note)
+static int fill_files_note(struct memelfnote *note)
 {
 	struct vm_area_struct *vma;
 	unsigned count, size, names_ofs, remaining, n;
@@ -1444,11 +1444,11 @@ static void fill_files_note(struct memelfnote *note)
 	names_ofs = (2 + 3 * count) * sizeof(data[0]);
  alloc:
 	if (size >= MAX_FILE_NOTE_SIZE) /* paranoia check */
-		goto err;
+		return -E2BIG;
 	size = round_up(size, PAGE_SIZE);
 	data = vmalloc(size);
 	if (!data)
-		goto err;
+		return -ENOMEM;
 
 	start_end_ofs = data + 2;
 	name_base = name_curpos = ((char *)data) + names_ofs;
@@ -1501,7 +1501,7 @@ static void fill_files_note(struct memelfnote *note)
 
 	size = name_curpos - (char *)data;
 	fill_note(note, "CORE", NT_FILE, size, data);
- err: ;
+	return 0;
 }
 
 #ifdef CORE_DUMP_USE_REGSET
@@ -1623,6 +1623,7 @@ static int fill_note_info(struct elfhdr *elf, int phdrs,
 	struct elf_prpsinfo *psinfo;
 	struct core_thread *ct;
 	unsigned int i;
+	int ret;
 
 	info->size = 0;
 	info->thread = NULL;
@@ -1702,8 +1703,9 @@ static int fill_note_info(struct elfhdr *elf, int phdrs,
 	fill_auxv_note(&info->auxv, current->mm);
 	info->size += notesize(&info->auxv);
 
-	fill_files_note(&info->files);
-	info->size += notesize(&info->files);
+	ret = fill_files_note(&info->files);
+	if (!ret)
+		info->size += notesize(&info->files);
 
 	return 1;
 }
@@ -1735,7 +1737,7 @@ static int write_note_info(struct elf_note_info *info,
 			return 0;
 		if (first && !writenote(&info->auxv, cprm))
 			return 0;
-		if (first && !writenote(&info->files, cprm))
+		if (first && info->files.data && !writenote(&info->files, cprm))
 			return 0;
 
 		for (i = 1; i < info->thread_notes; ++i)
@@ -1822,6 +1824,7 @@ static int elf_dump_thread_status(long signr, struct elf_thread_status *t)
 
 struct elf_note_info {
 	struct memelfnote *notes;
+	struct memelfnote *notes_files;
 	struct elf_prstatus *prstatus;	/* NT_PRSTATUS */
 	struct elf_prpsinfo *psinfo;	/* NT_PRPSINFO */
 	struct list_head thread_list;
@@ -1865,6 +1868,7 @@ static int fill_note_info(struct elfhdr *elf, int phdrs,
 			  siginfo_t *siginfo, struct pt_regs *regs)
 {
 	struct list_head *t;
+	int ret;
 
 	if (!elf_note_info_init(info))
 		return 0;
@@ -1912,9 +1916,13 @@ static int fill_note_info(struct elfhdr *elf, int phdrs,
 
 	fill_siginfo_note(info->notes + 2, &info->csigdata, siginfo);
 	fill_auxv_note(info->notes + 3, current->mm);
-	fill_files_note(info->notes + 4);
+	info->numnote = 4;
 
-	info->numnote = 5;
+	ret = fill_files_note(info->notes + info->numnote);
+	if (!ret) {
+		info->notes_files = info->notes + info->numnote;
+		info->numnote++;
+	}
 
 	/* Try to dump the FPU. */
 	info->prstatus->pr_fpvalid = elf_core_copy_task_fpregs(current, regs,
@@ -1976,8 +1984,9 @@ static void free_note_info(struct elf_note_info *info)
 		kfree(list_entry(tmp, struct elf_thread_status, list));
 	}
 
-	/* Free data allocated by fill_files_note(): */
-	vfree(info->notes[4].data);
+	/* Free data possibly allocated by fill_files_note(): */
+	if (info->notes_files)
+		vfree(info->notes_files->data);
 
 	kfree(info->prstatus);
 	kfree(info->psinfo);
@@ -2059,7 +2068,7 @@ static int elf_core_dump(struct coredump_params *cprm)
 	struct vm_area_struct *vma, *gate_vma;
 	struct elfhdr *elf = NULL;
 	loff_t offset = 0, dataoff;
-	struct elf_note_info info;
+	struct elf_note_info info = {0, };
 	struct elf_phdr *phdr4note = NULL;
 	struct elf_shdr *shdr4extnum = NULL;
 	Elf_Half e_phnum;
Signature · 1 line
1.8.1.4
#6 lore raw

Re: [PATCH linux-next] Prevent a coredump with a large vm_map_count from Oopsing

Hi Dan,
thank you for your work on my issue. I would like to test it on 3.10.9 where
I faced the problem initially.

linux-3.10.9 # patch -p1 < ../patches/vm_map_count.patch patching file fs/binfmt_elf.c
Hunk #1 succeeded at 1415 (offset -14 lines).
Hunk #2 succeeded at 1430 (offset -14 lines).
Hunk #3 succeeded at 1487 (offset -14 lines).
Hunk #4 succeeded at 1609 (offset -14 lines).
Hunk #5 succeeded at 1689 (offset -14 lines).
Hunk #6 FAILED at 1737.
Hunk #7 succeeded at 1810 (offset -14 lines).
Hunk #8 succeeded at 1854 (offset -14 lines).
Hunk #9 succeeded at 1902 (offset -14 lines).
Hunk #10 succeeded at 1970 (offset -14 lines).
Hunk #11 FAILED at 2068.
2 out of 11 hunks FAILED -- saving rejects to file fs/binfmt_elf.c.rej #

Thank you.

Quoted 149 lines · Dan Aloni

A high setting of max_map_count, and a process core-dumping with a large enough vm_map_count could result in an NT_FILE note not being written, and the kernel crashing immediately later because it has assumed otherwise.

Reproduction of the bug described here:

    https://lkml.org/lkml/2013/8/30/50

Issue originating in 2aa362c49 (from Oct 4, 2012).

This patch make that section optional in that case. fill_files_note() should signify the error, and also let the info struct in elf_core_dump() be zero-initialized so that we can check for the optionally written note.

Cc'ed original signers.

Cc'ed Al Viro because it is trivially relies on his linux-next tree changes.

Cc Al Viro <[email protected]>
Cc Denys Vlasenko <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 fs/binfmt_elf.c | 33 +++++++++++++++++++++------------
 1 file changed, 21 insertions(+), 12 deletions(-)
fs/binfmt_elf.c +21 -12
index dc82279..e1a323a 100644
@@ -1429,7 +1429,7 @@ static void fill_siginfo_note(struct memelfnote *note, user_siginfo_t *csigdata,
  *   long file_ofs
  * followed by COUNT filenames in ASCII: "FILE1" NUL "FILE2" NUL...
  */
-static void fill_files_note(struct memelfnote *note)
+static int fill_files_note(struct memelfnote *note)
 {
 	struct vm_area_struct *vma;
 	unsigned count, size, names_ofs, remaining, n;
@@ -1444,11 +1444,11 @@ static void fill_files_note(struct memelfnote *note)
 	names_ofs = (2 + 3 * count) * sizeof(data[0]);
  alloc:
 	if (size >= MAX_FILE_NOTE_SIZE) /* paranoia check */
-		goto err;
+		return -E2BIG;
 	size = round_up(size, PAGE_SIZE);
 	data = vmalloc(size);
 	if (!data)
-		goto err;
+		return -ENOMEM;
 
 	start_end_ofs = data + 2;
 	name_base = name_curpos = ((char *)data) + names_ofs;
@@ -1501,7 +1501,7 @@ static void fill_files_note(struct memelfnote *note)
 
 	size = name_curpos - (char *)data;
 	fill_note(note, "CORE", NT_FILE, size, data);
- err: ;
+	return 0;
 }
 
 #ifdef CORE_DUMP_USE_REGSET
@@ -1623,6 +1623,7 @@ static int fill_note_info(struct elfhdr *elf, int phdrs,
 	struct elf_prpsinfo *psinfo;
 	struct core_thread *ct;
 	unsigned int i;
+	int ret;
 
 	info->size = 0;
 	info->thread = NULL;
@@ -1702,8 +1703,9 @@ static int fill_note_info(struct elfhdr *elf, int phdrs,
 	fill_auxv_note(&info->auxv, current->mm);
 	info->size += notesize(&info->auxv);
 
-	fill_files_note(&info->files);
-	info->size += notesize(&info->files);
+	ret = fill_files_note(&info->files);
+	if (!ret)
+		info->size += notesize(&info->files);
 
 	return 1;
 }
@@ -1735,7 +1737,7 @@ static int write_note_info(struct elf_note_info *info,
 			return 0;
 		if (first && !writenote(&info->auxv, cprm))
 			return 0;
-		if (first && !writenote(&info->files, cprm))
+		if (first && info->files.data && !writenote(&info->files, cprm))
 			return 0;
 
 		for (i = 1; i < info->thread_notes; ++i)
@@ -1822,6 +1824,7 @@ static int elf_dump_thread_status(long signr, struct elf_thread_status *t)
 
 struct elf_note_info {
 	struct memelfnote *notes;
+	struct memelfnote *notes_files;
 	struct elf_prstatus *prstatus;	/* NT_PRSTATUS */
 	struct elf_prpsinfo *psinfo;	/* NT_PRPSINFO */
 	struct list_head thread_list;
@@ -1865,6 +1868,7 @@ static int fill_note_info(struct elfhdr *elf, int phdrs,
 			  siginfo_t *siginfo, struct pt_regs *regs)
 {
 	struct list_head *t;
+	int ret;
 
 	if (!elf_note_info_init(info))
 		return 0;
@@ -1912,9 +1916,13 @@ static int fill_note_info(struct elfhdr *elf, int phdrs,
 
 	fill_siginfo_note(info->notes + 2, &info->csigdata, siginfo);
 	fill_auxv_note(info->notes + 3, current->mm);
-	fill_files_note(info->notes + 4);
+	info->numnote = 4;
 
-	info->numnote = 5;
+	ret = fill_files_note(info->notes + info->numnote);
+	if (!ret) {
+		info->notes_files = info->notes + info->numnote;
+		info->numnote++;
+	}
 
 	/* Try to dump the FPU. */
 	info->prstatus->pr_fpvalid = elf_core_copy_task_fpregs(current, regs,
@@ -1976,8 +1984,9 @@ static void free_note_info(struct elf_note_info *info)
 		kfree(list_entry(tmp, struct elf_thread_status, list));
 	}
 
-	/* Free data allocated by fill_files_note(): */
-	vfree(info->notes[4].data);
+	/* Free data possibly allocated by fill_files_note(): */
+	if (info->notes_files)
+		vfree(info->notes_files->data);
 
 	kfree(info->prstatus);
 	kfree(info->psinfo);
@@ -2059,7 +2068,7 @@ static int elf_core_dump(struct coredump_params *cprm)
 	struct vm_area_struct *vma, *gate_vma;
 	struct elfhdr *elf = NULL;
 	loff_t offset = 0, dataoff;
-	struct elf_note_info info;
+	struct elf_note_info info = {0, };
 	struct elf_phdr *phdr4note = NULL;
 	struct elf_shdr *shdr4extnum = NULL;
 	Elf_Half e_phnum;
Signature · 7 lines
Martin Mokrejs, Ph.D.
Bioinformatics
Donovalska 1658
149 00 Prague
Czech Republic
http://www.iresite.org
http://www.iresite.org/~mmokrejs
#7 lore raw

Re: [PATCH linux-next] Prevent a coredump with a large vm_map_count from Oopsing

Quoted 3 lines · Martin MOKREJŠ

Hi Dan,
thank you for your work on my issue. I would like to test it on 3.10.9 where
I faced the problem initially.

Sure, see the attached patch for 3.10.9.

Signature · 1 line
Dan Aloni
Quoted 1 line

From e323d3b4fdc1e61c3c39dfb3733d8b8c56f63b00 Mon Sep 17 00:00:00 2001

From: Dan Aloni <[email protected]>
Date: Sat, 31 Aug 2013 00:13:43 +0300
Subject: [PATCH 1/1] Prevent a coredump with a large max_map_count from
Oopsing

A high setting of max_map_count, and a process core-dumping with a large enough vm_map_count could result in a NT_FILE note not being written, and the kernel crashing immediately later because it has assumed otherwise.

Reproduction of the bug described here:

    https://lkml.org/lkml/2013/8/30/50

This patch make that section optional in that case. fill_files_note() should signify the error, and also let the info struct in elf_core_dump() be zero-initialized so that we can check for optionally written note.

Cc'ed original signers.

Cc Denys Vlasenko <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Linus Torvalds <[email protected]>
---
 fs/binfmt_elf.c | 34 ++++++++++++++++++++++------------
 1 file changed, 22 insertions(+), 12 deletions(-)
fs/binfmt_elf.c +22 -12
index f8a0b0e..1c4a425 100644
@@ -1415,7 +1415,7 @@ static void fill_siginfo_note(struct memelfnote *note, user_siginfo_t *csigdata,
  *   long file_ofs
  * followed by COUNT filenames in ASCII: "FILE1" NUL "FILE2" NUL...
  */
-static void fill_files_note(struct memelfnote *note)
+static int fill_files_note(struct memelfnote *note)
 {
 	struct vm_area_struct *vma;
 	unsigned count, size, names_ofs, remaining, n;
@@ -1430,11 +1430,11 @@ static void fill_files_note(struct memelfnote *note)
 	names_ofs = (2 + 3 * count) * sizeof(data[0]);
  alloc:
 	if (size >= MAX_FILE_NOTE_SIZE) /* paranoia check */
-		goto err;
+		return -E2BIG;
 	size = round_up(size, PAGE_SIZE);
 	data = vmalloc(size);
 	if (!data)
-		goto err;
+		return -ENOMEM;
 
 	start_end_ofs = data + 2;
 	name_base = name_curpos = ((char *)data) + names_ofs;
@@ -1487,7 +1487,7 @@ static void fill_files_note(struct memelfnote *note)
 
 	size = name_curpos - (char *)data;
 	fill_note(note, "CORE", NT_FILE, size, data);
- err: ;
+	return 0;
 }
 
 #ifdef CORE_DUMP_USE_REGSET
@@ -1609,6 +1609,7 @@ static int fill_note_info(struct elfhdr *elf, int phdrs,
 	struct elf_prpsinfo *psinfo;
 	struct core_thread *ct;
 	unsigned int i;
+	int ret;
 
 	info->size = 0;
 	info->thread = NULL;
@@ -1688,8 +1689,9 @@ static int fill_note_info(struct elfhdr *elf, int phdrs,
 	fill_auxv_note(&info->auxv, current->mm);
 	info->size += notesize(&info->auxv);
 
-	fill_files_note(&info->files);
-	info->size += notesize(&info->files);
+	ret = fill_files_note(&info->files);
+	if (!ret)
+		info->size += notesize(&info->files);
 
 	return 1;
 }
@@ -1721,7 +1723,8 @@ static int write_note_info(struct elf_note_info *info,
 			return 0;
 		if (first && !writenote(&info->auxv, file, foffset))
 			return 0;
-		if (first && !writenote(&info->files, file, foffset))
+		if (first && info->files.data && !writenote(&info->files,
+							    file, foffset))
 			return 0;
 
 		for (i = 1; i < info->thread_notes; ++i)
@@ -1808,6 +1811,7 @@ static int elf_dump_thread_status(long signr, struct elf_thread_status *t)
 
 struct elf_note_info {
 	struct memelfnote *notes;
+	struct memelfnote *notes_files;
 	struct elf_prstatus *prstatus;	/* NT_PRSTATUS */
 	struct elf_prpsinfo *psinfo;	/* NT_PRPSINFO */
 	struct list_head thread_list;
@@ -1851,6 +1855,7 @@ static int fill_note_info(struct elfhdr *elf, int phdrs,
 			  siginfo_t *siginfo, struct pt_regs *regs)
 {
 	struct list_head *t;
+	int ret;
 
 	if (!elf_note_info_init(info))
 		return 0;
@@ -1898,9 +1903,13 @@ static int fill_note_info(struct elfhdr *elf, int phdrs,
 
 	fill_siginfo_note(info->notes + 2, &info->csigdata, siginfo);
 	fill_auxv_note(info->notes + 3, current->mm);
-	fill_files_note(info->notes + 4);
+	info->numnote = 4;
 
-	info->numnote = 5;
+	ret = fill_files_note(info->notes + info->numnote);
+	if (!ret) {
+		info->notes_files = info->notes + info->numnote;
+		info->numnote++;
+	}
 
 	/* Try to dump the FPU. */
 	info->prstatus->pr_fpvalid = elf_core_copy_task_fpregs(current, regs,
@@ -1962,8 +1971,9 @@ static void free_note_info(struct elf_note_info *info)
 		kfree(list_entry(tmp, struct elf_thread_status, list));
 	}
 
-	/* Free data allocated by fill_files_note(): */
-	vfree(info->notes[4].data);
+	/* Free data possibly allocated by fill_files_note(): */
+	if (info->notes_files)
+		vfree(info->notes_files->data);
 
 	kfree(info->prstatus);
 	kfree(info->psinfo);
@@ -2046,7 +2056,7 @@ static int elf_core_dump(struct coredump_params *cprm)
 	struct vm_area_struct *vma, *gate_vma;
 	struct elfhdr *elf = NULL;
 	loff_t offset = 0, dataoff, foffset;
-	struct elf_note_info info;
+	struct elf_note_info info = {0, };
 	struct elf_phdr *phdr4note = NULL;
 	struct elf_shdr *shdr4extnum = NULL;
 	Elf_Half e_phnum;
Signature · 1 line
1.8.1.4
#8 lore raw

Re: [PATCH linux-next] Prevent a coredump with a large vm_map_count from Oopsing

Quoted 6 lines · Dan Aloni
Quoted 3 lines · Martin MOKREJŠ

Hi Dan,
thank you for your work on my issue. I would like to test it on 3.10.9 where
I faced the problem initially.

Sure, see the attached patch for 3.10.9.

Thanks, it works for my case. You can add my Reported-by: and Tested-by:. ;-)