I'm not aware of any fixes for this specific issue. I had the original author of StatCache take a look at @BBlack's comments and he said it shouldn't be possible without some kind of memory corruption, which could've been fixed (or perturbed out of existence).
- Queries
- All Stories
- Search
- Advanced Search
- Transactions
- Transaction Logs
Feed Advanced Search
Advanced Search
Advanced Search
Jul 22 2015
Jul 22 2015
Mar 3 2015
Mar 3 2015
facebook-github-bot <githubbot@fb.com> committed rOSHDf4e77b17c79c: Clean up some tracing and asserts (authored by swtaarrs).
Clean up some tracing and asserts
facebook-github-bot <githubbot@fb.com> committed rOSHD684240e28c02: Don't use -Og for debug build, and add -ggdb (authored by swtaarrs).
Don't use -Og for debug build, and add -ggdb
facebook-github-bot <githubbot@fb.com> committed rOSHDb62bae3e2ac5: RBTrace should be NNT (authored by swtaarrs).
RBTrace should be NNT
facebook-github-bot <githubbot@fb.com> committed rOSHD1151f50b55f9: Clean up CONTRIBUTING.md (authored by swtaarrs).
Clean up CONTRIBUTING.md
facebook-github-bot <githubbot@fb.com> committed rOSHD615409bf8e84: Clean up uses of Type::unbox() in MInstrTranslator (authored by swtaarrs).
Clean up uses of Type::unbox() in MInstrTranslator
facebook-github-bot <githubbot@fb.com> committed rOSHD6e8657716739: Remove JIT::RuntimeType (authored by swtaarrs).
Remove JIT::RuntimeType
facebook-github-bot <githubbot@fb.com> committed rOSHDd7a3be1d72bc: Fix some Type::Null stuff, LbGbl guard relaxation (authored by swtaarrs).
Fix some Type::Null stuff, LbGbl guard relaxation
JoelMarcey <joelm@fb.com> committed rOSHD541170360b80: Fix bug in cgStoreTypedValue (authored by swtaarrs).
Fix bug in cgStoreTypedValue
facebook-github-bot <githubbot@fb.com> committed rOSHDfb58c2e57887: Expand documentation on Value::merge in refcount-opts (authored by swtaarrs).
Expand documentation on Value::merge in refcount-opts
facebook-github-bot <githubbot@fb.com> committed rOSHD74b1a7596602: Disable flaky zend test (authored by swtaarrs).
Disable flaky zend test
facebook-github-bot <githubbot@fb.com> committed rOSHD94db5a52496e: Re-record assetic (authored by swtaarrs).
Re-record assetic
facebook-github-bot <githubbot@fb.com> committed rOSHDc3953723a876: CallsNative flag on ir opcodes should imply MayRaiseError by default (authored by swtaarrs).
CallsNative flag on ir opcodes should imply MayRaiseError by default
facebook-github-bot <githubbot@fb.com> committed rOSHD2c22fff6dbb6: Expand jit documentation a bit (authored by swtaarrs).
Expand jit documentation a bit
facebook-github-bot <githubbot@fb.com> committed rOSHD605393280800: Initial coding conventions doc (authored by swtaarrs).
Initial coding conventions doc
facebook-github-bot <githubbot@fb.com> committed rOSHD85bcb83efc31: Disable flaky zend tests (authored by swtaarrs).
Disable flaky zend tests
facebook-github-bot <githubbot@fb.com> committed rOSHD3b5e28c1f1f9: Disable flaky zend test (authored by swtaarrs).
Disable flaky zend test
facebook-github-bot <githubbot@fb.com> committed rOSHD810d3cccf288: Fix relaxation of specialized boxed types (authored by swtaarrs).
Fix relaxation of specialized boxed types
facebook-github-bot <githubbot@fb.com> committed rOSHDea5a7a4a8644: Disable flaky zend test (authored by swtaarrs).
Disable flaky zend test
Sara Golemon <sgolemon@fb.com> committed rOSHDabf5df92ed31: More ir guard relaxation cleanup (authored by swtaarrs).
More ir guard relaxation cleanup
Sara Golemon <sgolemon@fb.com> committed rOSHD4037b66b320e: Change AssertType and AssertStk to refine their source type through typeParam (authored by swtaarrs).
Change AssertType and AssertStk to refine their source type through typeParam
Sara Golemon <sgolemon@fb.com> committed rOSHD77632ba4312c: Don't assume predicted output type for interpOne (authored by swtaarrs).
Don't assume predicted output type for interpOne
Sara Golemon <sgolemon@fb.com> committed rOSHD2140496762b0: Add a mode to always translate with and without region translator (authored by swtaarrs).
Add a mode to always translate with and without region translator
Sara Golemon <sgolemon@fb.com> committed rOSHD8bae0a6b2065: Disable flaky zend test (authored by swtaarrs).
Disable flaky zend test
Sara Golemon <sgolemon@fb.com> committed rOSHD64d90cf69cbe: Support multiple parallel configs in server mode tests (authored by swtaarrs).
Support multiple parallel configs in server mode tests
Sara Golemon <sgolemon@fb.com> committed rOSHD22cecaa32e93: Disable flaky zend test (authored by swtaarrs).
Disable flaky zend test
Sara Golemon <sgolemon@fb.com> committed rOSHD06acf96fb1e8: Punt to interpreter if region translation fails in tracelet/legacy mode (authored by swtaarrs).
Punt to interpreter if region translation fails in tracelet/legacy mode
Sara Golemon <sgolemon@fb.com> committed rOSHD2f4e4c02743e: Delete a .expect-trimmed file, add .gitignore rule (authored by swtaarrs).
Delete a .expect-trimmed file, add .gitignore rule
Sara Golemon <sgolemon@fb.com> committed rOSHDc072db13c786: Don't use Set in test/run (authored by swtaarrs).
Don't use Set in test/run
Sara Golemon <sgolemon@fb.com> committed rOSHD712de2c2f9b5: Add more detailed logging when we can't open the central repo (authored by swtaarrs).
Add more detailed logging when we can't open the central repo
Sara Golemon <sgolemon@fb.com> committed rOSHDadf430e96dc9: Don't write to a freed buffer (authored by swtaarrs).
Don't write to a freed buffer
Sara Golemon <sgolemon@fb.com> committed rOSHDe4992ae4eb01: Update NEWS for new release (authored by swtaarrs).
Update NEWS for new release
Sara Golemon <sgolemon@fb.com> committed rOSHDf6493ff67ae0: Clear out request-local globals at the beginning of each request (authored by swtaarrs).
Clear out request-local globals at the beginning of each request
Sara Golemon <sgolemon@fb.com> committed rOSHD6a41c10cdf2b: Don't truncate HphpArray's allocation size to 32 bits (authored by swtaarrs).
Don't truncate HphpArray's allocation size to 32 bits
Sara Golemon <sgolemon@fb.com> committed rOSHD25c9bf57ef9c: Clean up http header initialization code (authored by swtaarrs).
Clean up http header initialization code
Sara Golemon <sgolemon@fb.com> committed rOSHDa1ee28fa40d7: Don't close std(in|out|err) on request shutdown (authored by swtaarrs).
Don't close std(in|out|err) on request shutdown
Sara Golemon <sgolemon@fb.com> committed rOSHD7a7820b2e8c2: Use the correct closure class name in the emitter (authored by swtaarrs).
Use the correct closure class name in the emitter
Sara Golemon <sgolemon@fb.com> committed rOSHDcf38b4f72c12: Add some type-related asserts (authored by swtaarrs).
Add some type-related asserts
Jordan DeLong <jdelong@fb.com> committed rOSHF900b4f9b7f14: Remove an unused variable in FormatValue<double> (authored by swtaarrs).
Remove an unused variable in FormatValue<double>
Feb 23 2015
Feb 23 2015
With everything waiting on a lock I'd suspect some kind of deadlock, but a lot of those thread backtraces don't make any sense. The one for Thread 4 in particular claims to be in code that only runs when you're building a repo ahead of time. I also don't know what's up with the threads that claim to be calling __ll_lock_wait from vector::emplace_back.
Dec 19 2014
Dec 19 2014
swtaarrs added a comment to T73486: HHVM: segfault when serializing/unserializing large preprocessor cache items.
Someone's already doing option (b) internally, which will at least turn segfaults into fatal errors. I'll let you know when the fix makes it out to master (it's pretty simple). He's also going to look at (c) but that might not end up happening if it's too complex or makes things slower.
Nov 6 2014
Nov 6 2014
When I run the latest repro in an asan build, it reliably yells about a double free regardless of $n. This diff applied on top of the previous patch should fix the crash, and also another leak: https://gist.github.com/swtaarrs/2577519dbcf6e6e062a2
Nov 5 2014
Nov 5 2014
Crashes are still happening with the new patch, which I kinda expected. We poked at a core file for a while and it looks like it's crashing on a double free of an <a> element with these parents: <body><div id="mainpage"><table id="mp-tfa"><tr><td><p><a href="/wiki/প্রামাণ্যচিত্র">প্রামাণ্যচিত্র</a>. The document also contains headers for "featured article" and "did you know", strongly suggesting it has something to do with this code: https://github.com/wikimedia/mediawiki-extensions-MobileFrontend/blob/master/includes/MobileFormatter.php.
Nov 4 2014
Nov 4 2014
I rebased Tim's fix onto current master and the small repro is segfaulting now. I took a deeper look at the code and think I might have a more robust fix. I'll update as soon as I know more.
Oct 28 2014
Oct 28 2014
Here's one of those graphs, from running the repro case for a while:
loadhtml-heap.pdf13 KBDownload
. 90% of the heap is from DOMDocument::loadHTMLI can't find any evidence of that path getting hit; it looks like all the leaked memory is still coming from the call to htmlParseDocument in dom_load_html (I've been looking at call graphs from pprof's --pdf option).
Here's some stuff I wrote up before seeing @ori's most recent test case:
Oct 27 2014
Oct 27 2014
Awesome, thanks for the investigation. I can take a look at fixing this unless one of you already is.
Content licensed under Creative Commons Attribution-ShareAlike (CC BY-SA) 4.0 unless otherwise noted; code licensed under GNU General Public License (GPL) 2.0 or later and other open source licenses. By using this site, you agree to the Terms of Use, Privacy Policy, and Code of Conduct. · Wikimedia Foundation · Privacy Policy · Code of Conduct · Terms of Use · Disclaimer · CC-BY-SA · GPL