1595596 - Crash in [@ LZ4F_decodeHeader]

Status: RESOLVED FIXED

(Toolkit :: Startup and Profile System, defect, P2)

Description

[[:philipp]]

This bug is for crash report bp-d9fab019-262c-41e6-8079-056ae0191111.

Top 10 frames of crashing thread:

0 mozglue.dll static unsigned int LZ4F_decodeHeader mfbt/lz4/lz4frame.c:1119
1 mozglue.dll LZ4F_decompress mfbt/lz4/lz4frame.c:1395
2 mozglue.dll mozilla::Compression::LZ4FrameDecompressionContext::Decompress mfbt/Compression.cpp:181
3 xul.dll mozilla::scache::StartupCache::GetBuffer startupcache/StartupCache.cpp:387
4 xul.dll ReadCachedScript js/xpconnect/loader/mozJSLoaderUtils.cpp:25
5 xul.dll nsresult mozJSComponentLoader::ObjectForLocation js/xpconnect/loader/mozJSComponentLoader.cpp:820
6 xul.dll mozJSComponentLoader::Import js/xpconnect/loader/mozJSComponentLoader.cpp:1340
7 xul.dll static bool mozilla::dom::module_getter::ModuleGetter dom/base/ChromeUtils.cpp:513
8 xul.dll js::InternalCallOrConstruct js/src/vm/Interpreter.cpp:550
9 xul.dll js::CallGetter js/src/vm/Interpreter.cpp:760

this is a low volume crash (so far only in windows) apparently related to the new way that startup cache is working after bug 1550108.

the crash reason field for these reports indicate that some external factors are playing a role in this:
EXCEPTION_IN_PAGE_ERROR_READ / STATUS_IO_DEVICE_ERROR
EXCEPTION_IN_PAGE_ERROR_READ / STATUS_DEVICE_DATA_ERROR
EXCEPTION_IN_PAGE_ERROR_READ / STATUS_DEVICE_HARDWARE_ERROR

Comment 1

[Pascal Chevrel:pascalc (PTO until April 26)]

Doug, could you have a look at this regression caused by 1550108? Thanks

Comment 2

[Alex Thayer [:alexical] (she/her)]

Think I have a fix for this. However, I'm curious if similar crashes have been showing up for a while under the ScriptPreloader. My understanding of what's causing this makes me think that they ought to.

Comment 3

[Alex Thayer [:alexical] (she/her)]

Attachment: Bug 1595596 - Always wait on Prefetch thread before touching mCacheData

This is not, AFAICT, causing crashes - it is just an oversight noticed while
investigating the crash in the bug.

Comment 4

[Alex Thayer [:alexical] (she/her)]

Attachment: Bug 1595596 - Use MMAP_FAULT_HANDLER in StartupCache

Please double check that I am using this correctly. I believe we are
seeing the crash in the linked bug because we are not handling hardware
faults when reading from the memory mapped file. This patch just wraps
all accesses in the MMAP_FAULT_HANDLER_ macros.

Depends on D53042

Comment 5

[Pascal Chevrel:pascalc (PTO until April 26)]

We have shipped our last beta for 71, since there is a patch in the bug, I am setting the status as fix-optional for 71 in case there is a safe uplift possible in a potential dot release as a ride-along.

Comment 6

[BugBot [:suhaib / :marco/ :calixte]]

The priority flag is not set for this bug.
:mossop, could you have a look please?

For more information, please visit auto_nag documentation.

Comment 7

[BugBot [:suhaib / :marco/ :calixte]]

There are some r+ patches which didn't land and no activity in this bug for 2 weeks.
:dthayer, could you have a look please?
For more information, please visit auto_nag documentation.

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/1e21eefd5fce
Always wait on Prefetch thread before touching mCacheData r=froydnj
https://hg.mozilla.org/integration/autoland/rev/34c018c96749
Use MMAP_FAULT_HANDLER in StartupCache r=aklotz

Comment 9

[Dorel Luca [:dluca]]

Backed out 2 changesets (bug 1595596) for build bustage in build/src/mozglue/misc/MmapFaultHandler.cpp. CLOSED TREE

Log:
https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=295481970&repo=autoland&lineNumber=5455

Push with failures:
https://treeherder.mozilla.org/#/jobs?repo=autoland&group_state=expanded&selectedJob=295481970&revision=34c018c96749690a3994b506f1e5113a67647acf

Backout:
https://hg.mozilla.org/integration/autoland/rev/9a71a1f4ed3343ff46ddf7d2e65e0d283bc741f2

Comment 10

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/941df4ca0eb8
Always wait on Prefetch thread before touching mCacheData r=froydnj
https://hg.mozilla.org/integration/autoland/rev/322f2ff6b80a
Use MMAP_FAULT_HANDLER in StartupCache r=aklotz

Comment 11

[Cristian Brindusan [:cbrindusan]]

https://hg.mozilla.org/mozilla-central/rev/941df4ca0eb8
https://hg.mozilla.org/mozilla-central/rev/322f2ff6b80a

Comment 12

[Ryan VanderMeulen [:RyanVM]]

Crash volume looks pretty low for this, so let's just let it ride the 77 train to release. Free free to nominate for Beta approval if you feel strongly otherwise, however.