<div dir="ltr">The decision of doing away with -O2 is beyond our control, and we shouldn't have code which depend on optimization to be disabled to behave properly. Representing -errno as return is the cleanest fix (that's how other projects which use setcontext/getcontext are behaving too.) If there are any new further issues which arise from setcontext/getcontext, I'm tempted to change the internal implementation to use a a vanilla pthread pool.<div>
<br></div><div>Avati</div></div><div class="gmail_extra"><br><br><div class="gmail_quote">On Wed, Dec 11, 2013 at 10:29 PM, Anand Subramanian <span dir="ltr"><<a href="mailto:ansubram@redhat.com" target="_blank">ansubram@redhat.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Is doing away with -O2 an option that was ever considered or is it that we simply must have O2 on?<br>
<br>
(I understand that turning off O2 can open some so-far-unexposed can of worms and a lot of soaking maybe required, and also that we may have had a good set of perf related reasons to have settled on -O2 in the first place, but wanted to understand nevertheless...)<span class="HOEnZb"><font color="#888888"><br>
<br>
Anand</font></span><div class="HOEnZb"><div class="h5"><br>
<br>
<br>
On 12/11/2013 02:21 PM, Pranith Kumar Karampuri wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
hi,<br>
We found a day-1 bug when syncop_xxx() infra is used inside a synctask with compilation optimization (CFLAGS -O2). This bug has been dormant for at least 2 years.<br>
There are around ~400(rebalance, replace-brick, bd, self-heal-daemon, quota, fuse lock/fd migration) places where syncop is used in the code base all of which are potential candidates which can take the hit.<br>
<br>
I sent first round of patch at <a href="http://review.gluster.com/6475" target="_blank">http://review.gluster.com/6475</a> to catch regressions upstream.<br>
These are the files that are affected by the changes I introduced to fix this:<br>
<br>
api/src/glfs-fops.c | 36 ++++++++++++++++++++++++++++++<u></u>++++<br>
api/src/glfs-handleops.c | 15 ++++++++++++++<br>
api/src/glfs-internal.h | 7 +++++++<br>
api/src/glfs-resolve.c | 10 ++++++++++<br>
libglusterfs/src/syncop.c | 117 ++++++++++++++++++++++++++++++<u></u>++++++++++++++++++++++++++++++<u></u>++++++++++++------------------<u></u>-------------------<br>
xlators/cluster/afr/src/afr-<u></u>self-heald.c | 45 +++++++++++++++++++++++++++++-<u></u>------------<br>
xlators/cluster/afr/src/pump.c | 12 ++++++++++--<br>
xlators/cluster/dht/src/dht-<u></u>helper.c | 24 +++++++++++++++--------<br>
xlators/cluster/dht/src/dht-<u></u>rebalance.c | 168 ++++++++++++++++++++++++++++++<u></u>++++++++++++++++++++++++++++++<u></u>++++++++++++++++++++++++++++++<u></u>++++--------------------------<u></u>------------------------------<u></u>-------<br>
xlators/cluster/dht/src/dht-<u></u>selfheal.c | 6 ++++--<br>
xlators/features/locks/src/<u></u>posix.c | 3 ++-<br>
xlators/features/qemu-block/<u></u>src/bdrv-xlator.c | 15 ++++----------<br>
xlators/features/qemu-block/<u></u>src/qb-coroutines.c | 14 ++++++++++----<br>
xlators/mount/fuse/src/fuse-<u></u>bridge.c | 16 ++++++++++-----<br>
<br>
Please review your respective component for these changes in gerrit.<br>
<br>
Thanks<br>
Pranith.<br>
<br>
Detailed explanation of the Root cause:<br>
We found the bug in 'gf_defrag_migrate_data' in rebalance operation:<br>
<br>
Lets look at interesting parts of the function:<br>
<br>
int<br>
gf_defrag_migrate_data (xlator_t *this, gf_defrag_info_t *defrag, loc_t *loc,<br>
dict_t *migrate_data)<br>
{<br>
.....<br>
code section - [ Loop ]<br>
while ((ret = syncop_readdirp (this, fd, 131072, offset, NULL,<br>
&entries)) != 0) {<br>
.....<br>
code section - [ ERRNO-1 ] (errno of readdirp is stored in readdir_operrno by a thread)<br>
/* Need to keep track of ENOENT errno, that means, there is no<br>
need to send more readdirp() */<br>
readdir_operrno = errno;<br>
.....<br>
code section - [ SYNCOP-1 ] (syncop_getxattr is called by a thread)<br>
ret = syncop_getxattr (this, &entry_loc, &dict,<br>
GF_XATTR_LINKINFO_KEY);<br>
code section - [ ERRNO-2] (checking for failures of syncop_getxattr(). This may not always be executed in same thread which executed [SYNCOP-1])<br>
if (ret < 0) {<br>
if (errno != ENODATA) {<br>
loglevel = GF_LOG_ERROR;<br>
defrag->total_failures += 1;<br>
.....<br>
}<br>
<br>
the function above could be executed by thread(t1) till [SYNCOP-1] and code from [ERRNO-2] can be executed by a different thread(t2) because of the way syncop-infra schedules the tasks.<br>
<br>
when the code is compiled with -O2 optimization this is the assembly code that is generated:<br>
[ERRNO-1]<br>
1165 readdir_operrno = errno; <<---- errno gets expanded as *(__errno_location())<br>
0x00007fd149d48b60 <+496>: callq 0x7fd149d410c0 <__errno_location@plt><br>
0x00007fd149d48b72 <+514>: mov %rax,0x50(%rsp) <<------ Address returned by __errno_location() is stored in a special location in stack for later use.<br>
0x00007fd149d48b77 <+519>: mov (%rax),%eax<br>
0x00007fd149d48b79 <+521>: mov %eax,0x78(%rsp)<br>
....<br>
[ERRNO-2]<br>
1281 if (errno != ENODATA) {<br>
0x00007fd149d492ae <+2366>: mov 0x50(%rsp),%rax <<----- Because it already stored the address returned by __errno_location(), it just dereferences the address to get the errno value. BUT THIS CODE NEED NOT BE EXECUTED BY SAME THREAD!!!<br>
0x00007fd149d492b3 <+2371>: mov $0x9,%ebp<br>
0x00007fd149d492b8 <+2376>: mov (%rax),%edi<br>
0x00007fd149d492ba <+2378>: cmp $0x3d,%edi<br>
<br>
The problem is that __errno_location() value of t1 and t2 are different. So [ERRNO-2] ends up reading errno of t1 instead of errno of t2 even though t2 is executing [ERRNO-2] code section.<br>
<br>
When code is compiled without any optimization for [ERRNO-2]:<br>
1281 if (errno != ENODATA) {<br>
0x00007fd58e7a326f <+2237>: callq 0x7fd58e797300 <__errno_location@plt><<--- As it is calling __errno_location() again it gets the location from t2 so it works as intended.<br>
0x00007fd58e7a3274 <+2242>: mov (%rax),%eax<br>
0x00007fd58e7a3276 <+2244>: cmp $0x3d,%eax<br>
0x00007fd58e7a3279 <+2247>: je 0x7fd58e7a32a1 <gf_defrag_migrate_data+2287><br>
<br>
Fix:<br>
We decided to make syncop_xxx() return (-errno) value as the return value in case of errors and all the functions which make syncop_xxx() will need to use (-ret) to figure out the reason for failure in case of syncop_xxx() failures.<br>
<br>
Pranith<br>
<br>
______________________________<u></u>_________________<br>
Gluster-devel mailing list<br>
<a href="mailto:Gluster-devel@nongnu.org" target="_blank">Gluster-devel@nongnu.org</a><br>
<a href="https://lists.nongnu.org/mailman/listinfo/gluster-devel" target="_blank">https://lists.nongnu.org/<u></u>mailman/listinfo/gluster-devel</a><br>
</blockquote>
<br>
<br>
______________________________<u></u>_________________<br>
Gluster-devel mailing list<br>
<a href="mailto:Gluster-devel@nongnu.org" target="_blank">Gluster-devel@nongnu.org</a><br>
<a href="https://lists.nongnu.org/mailman/listinfo/gluster-devel" target="_blank">https://lists.nongnu.org/<u></u>mailman/listinfo/gluster-devel</a><br>
</div></div></blockquote></div><br></div>