New issue 84989 by sits...@gmail.com: Bus error while in inflate_fast
http://code.google.com/p/chromium/issues/detail?id=84989
Chrome Version : 13.0.771.0 (Developer Build 85921 Linux) custom
OS Version: Slackware 13.1.0
URLs (if applicable) :
Other browsers tested:
Add OK or FAIL after other browsers where you have tested this issue:
Safari 5:
Firefox 4.x:
IE 7/8/9:
What steps will reproduce the problem?
1. Start chromium .
2. Use chromium by browsing to sites with flash on them.
What is the expected result?
Just flash plugin to crash.
What happens instead?
Whole browser crashes.
Please provide any additional information below. Attach a screenshot if
possible.
This seemed to start happening about a few weeks ago. Disabling all flags
doesn't seem to make things any more stable. The two core dumps I have both
have a back trace that starts like the following:
Core was generated by `/tmp/chrome-linux/chrome --allow-sandbox-debugging'.
Program terminated with signal 7, Bus error.
#0 0x00007f1299d2c397 in inflate_fast () from /usr/lib64/libz.so.1
(gdb) bt
#0 0x00007f1299d2c397 in inflate_fast () from /usr/lib64/libz.so.1
#1 0x00007f1299d29f91 in inflate () from /usr/lib64/libz.so.1
#2 0x0000000000f98a10 in net::GZipFilter::DoInflate(char*, int*) ()
#3 0x0000000000f98f1a in net::GZipFilter::ReadFilteredData(char*, int*) ()
#4 0x0000000000f97b87 in net::Filter::ReadData(char*, int*) ()
#5 0x0000000000f835aa in net::URLRequestJob::ReadFilteredData(int*) ()
#6 0x0000000000f83bcb in net::URLRequestJob::Read(net::IOBuffer*, int,
int*) ()
#7 0x0000000000f79d92 in net::URLRequest::Read(net::IOBuffer*, int, int*)
()
#8 0x0000000001f8f288 in ResourceDispatcherHost::Read(net::URLRequest*,
int*) ()
#9 0x0000000001f97b34 in
ResourceDispatcherHost::StartReading(net::URLRequest*) ()
#10 0x0000000001f9a266 in
ResourceDispatcherHost::OnResponseStarted(net::URLRequest*) ()
#11 0x0000000000f7abde in net::URLRequest::ResponseStarted() ()
#12 0x0000000000f830f1 in net::URLRequestJob::NotifyHeadersComplete() ()
#13 0x0000000001021920 in net::URLRequestHttpJob::NotifyHeadersComplete() ()
#14 0x000000000102221d in net::URLRequestHttpJob::SaveNextCookie() ()
#15 0x0000000001022326 in
net::URLRequestHttpJob::SaveCookiesAndNotifyHeadersComplete() ()
#16 0x000000000102259a in net::URLRequestHttpJob::OnStartCompleted(int) ()
#17 0x0000000000c5219e in (anonymous namespace)::TaskClosureAdapter::Run()
()
#18 0x0000000000c53026 in MessageLoop::RunTask(MessageLoop::PendingTask
const&)
#19 0x0000000000c544f0 in
MessageLoop::DeferOrRunPendingTask(MessageLoop::PendingTask const&) ()
#20 0x0000000000c54987 in MessageLoop::DoWork() ()
#21 0x0000000000c33289 in
base::MessagePumpLibevent::Run(base::MessagePump::Delegate*) ()
#22 0x0000000000c561f6 in MessageLoop::RunInternal() ()
#23 0x0000000000c563ac in MessageLoop::Run() ()
#24 0x0000000000c83f97 in base::Thread::ThreadMain() ()
#25 0x0000000000c83a72 in base::(anonymous namespace)::ThreadFunc(void*) ()
#26 0x00007f1298d62980 in start_thread () from /lib64/libpthread.so.0
#27 0x00007f1296e8c57d in clone () from /lib64/libc.so.6
OK looking at this a bit closer this doesn't seem plugin related. The crash
is 100% reproducible for me by visiting http://www.netvibes.com/ - I can do
this in an incognito session, with click-to-play enabled or with a totally
new profile and the crash still occurs. The crash is also still happening
with 14.0.786.0 (Developer Build 87970 Linux) custom . The crash also seems
to happen on other addresses like
http://build.chromium.org/f/chromium/snapshots/chromium-rel-linux-64/ .
Comment #4 on issue 84989 by ero...@chromium.org: Bus error while in
inflate_fast
http://code.google.com/p/chromium/issues/detail?id=84989
(No comment was entered for this change.)
thestig:
I managed to chase this down to a full /dev/shm/ . After clearing out all
the things that had been accumulating there this problem has gone away. I
guess other browsers like Firefox don't use /dev/shm so they would not
manifest the problem...
Comment #6 on issue 84989 by the...@chromium.org: Bus error while in
inflate_fast
http://code.google.com/p/chromium/issues/detail?id=84989
Ok, glad you found the problem. And yes, Chromium uses /dev/shm quite a bit.