search

agusneos / lusca-cache

Automatically exported from code.google.com/p/lusca-cache
0 stars 0 forks source link

FATAL: Received Segment Violation...dying. - revision 13801 #3

Open GoogleCodeExporter opened 9 years ago

GoogleCodeExporter commented 9 years ago 
FATAL: Received Segment Violation...dying.
2009/02/05 17:06:31| storeDirWriteCleanLogs: Starting...
2009/02/05 17:06:31| WARNING: Closing open FD 15
2009/02/05 17:06:31| commSetEvents: epoll_ctl(EPOLL_CTL_DEL): failed on
fd=15: (1) Operation not permitted
2009/02/05 17:06:31| WARNING: Closing open FD 16
2009/02/05 17:06:31| commSetEvents: epoll_ctl(EPOLL_CTL_DEL): failed on
fd=16: (1) Operation not permitted

-------------------------

(gdb) bt
#0 0x000000368e830155 in raise () from /lib64/libc.so.6
#1 0x000000368e831bf0 in abort () from /lib64/libc.so.6
#2 0x000000000046de1e in death (sig=<value optimized out>) at tools.c:326
#3 <signal handler called>
#4 0x000000368e878827 in strncmp () from /lib64/libc.so.6
#5 0x00000000004656a6 in storeLocateVaryRead (data=0x1a3010c8, nr={node =
0x1a5434f0, offset = 0}, size=222) at store.c:1010
#6 0x0000000000466031 in storeClientCallback (sc=0x1a52de98, sz=222) at
store_client.c:156
#7 0x0000000000479eea in storeAufsReadDone (fd=<value optimized out>,
my_data=0x1a452bb8,
buf=0x1a4c65d0 "Key: D675BFBB0CE7C424BFC0D1D145A392C9\nAccept-Encoding:
gzip, deflate\nVaryData: accept-encoding=\"gzip,%20deflate\"\nKey:
A13D7D131373C1E77A47B25B780C4504\nAccept-Encoding: gzip,deflate\nVaryData:
accept-en"..., len=<value optimized out>, errflag=0) at
aufs/store_io_aufs.c:400
#8 0x000000000047c06a in aioCheckCallbacks () at aufs/async_io.c:319
#9 0x000000000047b81b in squidaio_fdhandler (fd=8, data=<value optimized
out>) at aufs/aiops.c:257
#10 0x00000000004a0459 in comm_select (msec=<value optimized out>) at
comm_generic.c:264
#11 0x00000000004477eb in main (argc=<value optimized out>,
argv=0x7fffe9f97d68) at main.c:933

Original issue reported on code.google.com by ajcor...@gmail.com on 5 Feb 2009 at 7:11

GoogleCodeExporter commented 9 years ago 
I think accept-encoding String may have a NULL buffer :(

Rev 13802 introduces strNCmpNull, like the store.c hacky routine before it, 
which implements the given 
behaviour. I hope. :)

Original comment by adrian.c...@gmail.com on 5 Feb 2009 at 7:36

GoogleCodeExporter commented 9 years ago 
This seems to have fixed the crashes. The Vary support should now be checked to 
ensure that my modifications 
in this part of the codebase haven't introduced any Vary regressions.

Original comment by adrian.c...@gmail.com on 6 Feb 2009 at 3:48