Closed Bug 1154416 Opened 10 years ago Closed 10 years ago

Attempt to pthread close an invalid file descriptor found during stability test

Categories

(Firefox OS Graveyard :: Stability, defect)

Product:
Firefox OS Graveyard
Component:
Stability
Platform:
ARM
Gonk (Firefox OS)
Type:
defect
Priority:
Not set
Severity:
critical

Tracking

(Not tracked)

Status:
RESOLVED DUPLICATE of bug 1154411

People

(Reporter: ggrisco, Assigned: sotaro, NeedInfo)

References

Details

(Keywords: crash, Whiteboard: [b2g-crash][caf-crash 621][caf priority: p1][CR 822151])

Crash Data

Attachments

(6 files)

EXTRA file attachment -
141.12 KB, text/plain
Details
decoded minidump -
164.08 KB, text/plain
Details
EXTRA file attachment - AU_LINUX_GECKO_LF.BR.1.2.3.00.00.00.000.129
141.12 KB, text/plain
Details
decoded minidump - AU_LINUX_GECKO_LF.BR.1.2.3.00.00.00.000.129
164.08 KB, text/plain
Details
EXTRA file attachment - AU_LINUX_GECKO_LF.BR.1.2.3.00.00.00.000.132
143.23 KB, text/plain
Details
decoded minidump - AU_LINUX_GECKO_LF.BR.1.2.3.00.00.00.000.132
156.54 KB, text/plain
Details
CAF debug code signals an abort when an attempt to close an invalid file descriptor is made. This instance is being found during stability test. cafbot will upload logs.
Attached file decoded minidump - —
Whiteboard: [CR 822151]
Whiteboard: [CR 822151] → [caf priority: p1][CR 822151]
See Also: → 1154411
fabrice/sotaro, need some help on initial investigation here as well. Similar to :https://bugzilla.mozilla.org/show_bug.cgi?id=1154411#c6
Flags: needinfo?(sotaro.ikeda.g)
Flags: needinfo?(fabrice.desre)
Hi! Rex, Please help to take a look. Thanks -- Keven
Flags: needinfo?(rhung)
As shown in the following, closing an invalid FD triggers calling abort(). However, the reason why b2g process should be aborted for this symptom is still under checking because I cannot find noselect.c from msm8909 source tree for now. Will update this part later. // 0be37360-0a8f-c058-60e4ff60-37e054d7.extra 740 04-14 14:48:34.538 18661 18661 E libnoselect: -1 = close(63) failed. errno=9 741 04-14 14:48:34.538 18661 18661 E Gecko : mozalloc_abort: Redirecting call to abort() to mozalloc_abort // 0be37360-0a8f-c058-60e4ff60-37e054d7.txt 10 Thread 0 (crashed) 11 0 plugin-container!mozalloc_abort [mozalloc_abort.cpp : 37 + 0x4] 12 r0 = 0x00000046 r1 = 0xbe9b6598 r2 = 0x00000025 r3 = 0x00000000 13 r4 = 0x00000009 r5 = 0xffffffff r6 = 0xb6fe8e38 r7 = 0xb29205b8 14 r8 = 0x000014fe r9 = 0x00053e26 r10 = 0xbe9b6af0 r12 = 0xb29205b8 15 fp = 0x00000000 sp = 0xbe9b6a08 lr = 0xb6feadb3 pc = 0xb6feadb6 16 Found by: given as instruction pointer in context 17 1 libnoselect.so!close [noselect.c : 89 + 0x3] ^^^^^^^^^^ Should check here to know why an abortion should be triggered. 18 r3 = 0x2dc8b506 r4 = 0x00000009 r5 = 0xffffffff r6 = 0xb6fe8e38 19 r7 = 0xb29205b8 r8 = 0x000014fe r9 = 0x00053e26 r10 = 0xbe9b6af0 20 fp = 0x00000000 sp = 0xbe9b6a10 pc = 0xb6fd0667 21 Found by: call frame info 22 2 libmozglue.so!__wrap_close [Nuwa.cpp : 1360 + 0x3] 23 r0 = 0x0000003f r1 = 0x00000009 r4 = 0xb1a7cc20 r5 = 0x0000003f 24 r6 = 0x0000003f r7 = 0xb29205b8 r8 = 0x000014fe r9 = 0x00053e26 25 r10 = 0xbe9b6af0 fp = 0x00000000 sp = 0xbe9b6a28 pc = 0xb6f97ce1
(In reply to Rex Hung[:rhung] from comment #5) > As shown in the following, closing an invalid FD triggers calling abort(). > However, the reason why b2g process should be aborted for this symptom is ^^^^^^^^^^^ > still under checking because I cannot find noselect.c from msm8909 source > tree for now. Will update this part later. Correct my comment at comment 5: It should a pre-allocated app, instead of b2g process.
Flags: needinfo?(rhung)
From the crash log, it seems dup of Bug 1154411.
Flags: needinfo?(sotaro.ikeda.g)
The libnoselect library is added in CAF internal builds to intercept calls to close() and check for invalid file descriptor, aborting if such a case is detected. This bug and bug 1154411 are two different instances where this is being caught (probably from different sources).
Whiteboard: [caf priority: p1][CR 822151] → [b2g-crash][caf-crash 621][caf priority: p1][CR 822151]
Keywords: crash
(In reply to Sotaro Ikeda [:sotaro] from comment #7) > From the crash log, it seems dup of Bug 1154411. As mentioned in comment 8, this crash could be different root cause. Is there any evidence that the patch for bug 1154411 will fix this one too? If not, I hope there can still be analysis done on this issue.
Flags: needinfo?(sotaro.ikeda.g)
(In reply to Greg Grisco from comment #15) > (In reply to Sotaro Ikeda [:sotaro] from comment #7) > > From the crash log, it seems dup of Bug 1154411. > > As mentioned in comment 8, this crash could be different root cause. Is > there any evidence that the patch for bug 1154411 will fix this one too? If > not, I hope there can still be analysis done on this issue. All crash have the following in the stack it is same to Bug 1154411 and attachment 8593613 [details] [diff] [review] is a fix to ArrayBufferBuilder::mapToFileInPackage(). > libxul.so!mozilla::ArrayBufferBuilder::mapToFileInPackage
Flags: needinfo?(sotaro.ikeda.g)
as comment 16, make duplication.
Assignee: nobody → sotaro.ikeda.g
Status: NEW → RESOLVED
Closed: 10 years ago
Resolution: --- → DUPLICATE
No longer blocks: CAF-v2.2-metabug
already dup, remove nom
blocking-b2g: 2.2? → ---
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: