| Summary: | www-client/chromium-20.0.1132.43 - ld catches a SIGBUS while linking out/Release/chrome | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Honza <hkmaly> |
| Component: | Current packages | Assignee: | Chromium Project <chromium> |
| Status: | RESOLVED WORKSFORME | ||
| Severity: | normal | ||
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | x86 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Attachments: | build log | ||
|
Description
Honza
2012-07-15 16:10:35 UTC
... tried adding 1GB of swap and compiled second time. This time it worked, despite only 80k of swap being used. Ok, not sure how this should be solved ... some warning message? Or should it be crossreported as a ld bug of not being able to clearly tell problem is not having swap? (In any case, I hope this bugreport will help someone.) ... of course now the chromium is not working, but that's probably unrelated (reported upstream as https://code.google.com/p/chromium/issues/detail?id=137438 ...) you didn't attach the full build log. I guess the failure probably due to the final linkage consume too much memory, thus trigger OOM error. Classical OOM errors are written into dmesg, this wasn't. But yes, considering the swap helped it probably WAS amount-of-RAM related ... only adding physical RAM wouldn't help as it already has 4GB (or, to be exact, adding physical RAM would not help better that adding swap). Created attachment 318386 [details]
build log
Full build log. Only 40MB (before bzip2ing) with nothing interesting except the parts already posted. Enjoy.
Maybe you simply ran out of memory. (In reply to comment #6) > Maybe you simply ran out of memory. By 80k? Also, I think I still had something in cache (I mean, top was showing something in "buffers" and "cached"). could you try 20.0.1132.57,see bug#426204,we stablize it recently also it compiles fine here (In reply to comment #7) > (In reply to comment #6) > > Maybe you simply ran out of memory. > > By 80k? Also, I think I still had something in cache (I mean, top was > showing something in "buffers" and "cached"). (In reply to comment #8) > could you try 20.0.1132.57,see bug#426204,we stablize it recently > also it compiles fine here > (In reply to comment #7) > > (In reply to comment #6) > > > Maybe you simply ran out of memory. > > > > By 80k? Also, I think I still had something in cache (I mean, top was > > showing something in "buffers" and "cached"). To my suprise, it compiles fine even without swap. (In reply to comment #9) > To my suprise, it compiles fine even without swap. Could you confirm this is fixed for you? I'm not going to compile chromium several times just to get statistically significant sample. Both "fixed" and "not reproducible always" is possible now. I'll return to this bug if it happens again for example after next GLSA or if I broke the installed chromium by upgrading some library it depends on, but I hope neither happen for some time. Or I'll return if some solution requiring recompilation will be offered on https://code.google.com/p/chromium/issues/detail?id=137438 (which is problem I have with that chromium now, but as it has workaround it's not so critical). We can only work with the information given, and that really doesn't tell us much. Same problem on different computer with 12GB RAM (or is it same computer I only added more memory since?). Different chromium (33.0.1750.146), different gcc (4.6.3), same binutils. After adding 100MB of swap (which was never used) it compiled ok. I don't expect you to do something about it, looks like it's random so it would be near impossible to debug unless someone will come with smaller example than chromium. |
