| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch does the following:
1. Addresses CID: 1124815,124816,1124833,1291724,1325535,1325536,1357858
- by adding some null checks
- by handling return values from functions
- by using an appropriate buffer length in strncpy
2. Cleans up some commented code
Change-Id: I5a7079f34e3e460d5a6267734c3bc84bf4ad72f5
updates: bz#789278
Signed-off-by: Vijay Bellur <vbellur@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
It's not needed.
There's a good chance the compiler is smart enough to remove it
anyway, but it can't hurt - I hope.
Compile-tested only!
Change-Id: Id7c054e146ba630227affa591007803f3046416b
updates: bz#1193929
Signed-off-by: Yaniv Kaul <ykaul@redhat.com>
|
|
|
|
|
|
| |
Change-Id: Ie4fe18d5094c051fa20de71f7fc841085cc6aaee
Fixes: bz#1614142
Signed-off-by: Nigel Babu <nigelb@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
While I was reading code, I saw that socket_submit_request()
and socket_submit_reply() are identical except for @a_byte and
the source of @msg object being different.
This patch moves all of the common code to a new function with
the differing vars passed as parameters by the callers.
Change-Id: I7a62ae72f10c34dc8de01b250d89a77ec5ab490d
fixes: bz#1608991
Signed-off-by: Krutika Dhananjay <kdhananj@redhat.com>
|
|
|
|
|
|
| |
Change-Id: I73a113e2d40f508fd53b273a990a2371692c87bf
fixes: bz#1607689
Signed-off-by: Zhang Huan <zhanghuan@open-fs.com>
|
|
|
|
|
|
| |
Change-Id: I3fa97b99bf23459cf548205d75d2cc7936b2310e
fixes: bz#1607689
Signed-off-by: Zhang Huan <zhanghuan@open-fs.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Newer FreeBSD versions (noticed with 10.3-RELEASE) provide a event.h
file that on occasion gets included instead of the libglusterfs file.
When this happens, 'struct event_pool' will not be defined and building
will fail with errors like:
autoscale-threads.c:18:55: error: incomplete definition of type 'struct event_pool'
int thread_count = pool->eventthreadcount;
~~~~^
autoscale-threads.c:17:16: note: forward declaration of 'struct event_pool'
struct event_pool *pool = ctx->event_pool;
^
This problem is caused by 'pkg-config --cflags uuid' that adds
/usr/local/include to the GF_CPPFLAGS. The use of libuuid is preferred
so that the contrib/uuid/ directory can be removed.
By renaming event.h to gf-event.h there is no conflict between the
different event.h files anymore and compiling on FreeBSD works without
issues.
Change-Id: Ie69f6b8a4f8f8e9630d39a86693eb74674f0f763
Updates: bz#1607319
Signed-off-by: Niels de Vos <ndevos@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Patch attempts to use the epoll infra for handling SSL connections
as well instead of the socket_poller() thread func.
This essentially makes priv->own_thread flag redundant.
SSL_connect()/SSL_accept() is now non-blocking which has done away
with the localised poll() in ssl_do(). So, ssl_do() has been updated
appropriately.
own_thread and coincidently socket_poller() thread for SSL processing
is now deprecated.
Change-Id: I1ce54c06ddb643c16baa143598e7e4fbf16bae0a
fixes: bz#1561332
Signed-off-by: Milind Changire <mchangir@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Problem: gfapi client is getting crashed in rpc_clnt_connection_cleanup
at the time of destroying saved_frames
Solution: gfapi client is getting crashed because saved_frame ptr is
already freed in rpc_clnt_destroy.To avoid the same update
code in rpc_clnt_destroy
Change-Id: Id8cce102b49f26cfd86ef88257032ed98f43192b
fixes: bz#1607783
Signed-off-by: Mohit Agrawal <moagrawal@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Please review, it's not always just the comments that were fixed.
I've had to revert of course all calls to creat() that were changed
to create() ...
Only compile-tested!
Change-Id: I7d02e82d9766e272a7fd9cc68e51901d69e5aab5
updates: bz#1193929
Signed-off-by: Yaniv Kaul <ykaul@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Problem: br-state-check.t is getting crash
Solution: Check condition in rpcsvc_request_create
before allocate memory from rxpool
BUG: 1597776
Change-Id: I4fde1ade6073f603c32453f1840395db9a9155b7
fixes: bz#1597776
Signed-off-by: Mohit Agrawal <moagrawa@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
| |
This patch fixes compile warnings that appear with newer compilers. The
solution applied is only to remove the warnings, but it doesn't always
solve the problem in the best way. It assumes that the problem will never
happen, as the previous code assumed.
Change-Id: I6e8470d6c2e2dbd3bd7d324b5fd2f92ffdc3d6ec
updates: bz#1193929
Signed-off-by: Xavi Hernandez <xhernandez@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The state management of "connected" in rpc is ad-hoc as far as the
responsibility goes. Note that there is nothing wrong with
functionality itself. rpc layer manages this state in disconnect
codepath and has exposed an api to manage this one from
consumers. Note that rpc layer never sets "connected" to true by
itself, which forces the consumers to use this api to get a working
rpc connection. The situation is best captured from a comment in code
from Jeff Darcy in glusterfsd/src/gf-attach.c:
-/*
- * In a sane world, the generic RPC layer would be capable of tracking
- * connection status by itself, with no help from us. It might invoke our
- * callback if we had registered one, but only to provide information. Sadly,
- * we don't live in that world. Instead, the callback *must* exist and *must*
- * call rpc_clnt_{set,unset}_connected, because that's the only way those
- * fields get set (with RPC both above and below us on the stack). If we don't
- * do that, then rpc_clnt_submit doesn't think we're connected even when we
- * are. It calls the socket code to reconnect, but the socket code tracks this
- * stuff in a sane way so it knows we're connected and returns EINPROGRESS.
- * Then we're stuck, connected but unable to use the connection. To make it
- * work, we define and register this trivial callback.
- */
Also, consumers of rpc know about state of connection only through the
notifications sent by rpc-clnt. So, consumers don't have any extra
information to manage the state and hence letting them manage the
state is counter intuitive. This patch cleans that up and instead
moves the responsibility of state management of rpc layer into
itself.
Change-Id: I31e641a60795fc480ca753917f4b2579f1e05094
Signed-off-by: Raghavendra G <rgowdapp@redhat.com>
Fixes: bz#1585585
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The auth_value was being reset to AUTH_GLUSTERFS_v2
during rpc disconnect. It shoud not be reset. The
disconnect during portmap request can race with
handshake. If handshake happens first and
disconnect later, auth_value would set to default
value and it never sets back to actual auth_value
fixes: bz#1579276
Change-Id: Ib46c9e01a97f6defb3fd1e0423fdb4b899b4a361
Signed-off-by: Kotresh HR <khiremat@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Problem:
On node reboot, when glusterd starts volumes rapidly, there's a flood of
connections from the bricks to glusterd and from the self-heal daemons
to the bricks. This causes SYN Flooding and dropped connections when the
listen-backlog is not enough to hold the pending connections to
compensate for the rate at which connections are accepted by the RPC
layer.
Solution:
Increase the listen-backlog value to 1024. This is a partial solution.
Part of the solution is to rearm the listener socket early for quicker
accept() of connections.
See commit 6964640a977cb10c0c95a94e03c229918fa6eca8 (change 19833)
Change-Id: I62283d1f4990dd43839f9a6932cf8a36effd632c
fixes: bz#1564600
Signed-off-by: Milind Changire <mchangir@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Problem:
On node reboot, when glusterd starts volumes, a setup with a large
number of bricks might cause SYN Flooding and connections to be dropped
if the connections are not accepted quickly enough.
Solution:
accept() the connection and rearm the listener socket early to receive
more connection requests as soon as possible.
Change-Id: Ibed421e50284c3f7a8fcdb4de7ac86cf53d4b74e
fixes: bz#1564600
Signed-off-by: Milind Changire <mchangir@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
re-registering
> Reviewed-on: https://review.gluster.org/16849
> Reviewed-by: Shreyas Siravara <sshreyas@fb.com>
Change-Id: I05ed6b7c715a71e5819fbe8116e7c3146010f836
BUG: 1521030
Signed-off-by: Kevin Vigor <kvigor@fb.com>
Signed-off-by: Amar Tumballi <amarts@redhat.com>
|
|
|
|
|
|
|
|
|
| |
When a saved frame is to be forced unwind, there is no need to pass an
empty iovector without any data pointed to.
Change-Id: I6e858fb38644326e22239b83272b15db656035e5
BUG: 1523122
Signed-off-by: Zhang Huan <zhanghuan@open-fs.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
xdr_replymsg is called to decode reply message, and it returns failure
if the message is corrupted. However, retrieving return value from
the global errno is 0 even xdr_replymsg fails.
Fix this issue by simply returning a negative value if call to
xdr_replymsg fails.
Change-Id: I2b9a1dc97652fbb6cf6568ea617f120713784a55
BUG: 1523122
Signed-off-by: Zhang Huan <zhanghuan@open-fs.com>
|
|
|
|
|
|
|
|
|
|
| |
* Say which file had the problem
* Dump openssl error stack
Fixes gluster/glusterfs#431.
Change-Id: I66e9a0ae7758e9d7d8a5f19cc8ff898f01f2b491
Signed-off-by: Niklas Hambüchen <mail@nh2.me>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Problem: TLS verification fails while using intermediate CA
if mgmt SSL is enabled.
Solution: There are two main issue of TLS verification failing
1) not calling ssl_api to set cert_depth
2) The current code does not allow to set certificate depth
while MGMT SSL is enabled.
After apply this patch to set certificate depth user
need to set parameter option transport.socket.ssl-cert-depth <depth>
in /var/lib/glusterd/secure_acccess instead to set in
/etc/glusterfs/glusterd.vol. At the time of set secure_mgmt in ctx
we will check the value of cert-depth and save the value of cert-depth
in ctx.If user does not provide any value in cert-depth in that case
it will consider default value is 1
BUG: 1555154
Change-Id: I89e9a9e1026e37efb5c20f9ec62b1989ef644f35
Signed-off-by: Mohit Agrawal <moagrawa@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Removed unused struct member and its one time usage.
- cleaned up wrong white space
member 'client_latency' was not used otherwise since it was added by
commit 07cc8679cdf3b29680f4f105d0222da168d8bfc1
Author: Kevin Vigor <kvigor@fb.com>
Date: Tue Mar 21 08:23:25 2017 -0700
Halo Replication feature for AFR translator
Change-Id: Ibb0ea828d4090bbe8897f6af326b317884162a00
BUG: 1495153
Signed-off-by: Sven Fischer <sven@fischer-abc.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We have the following undefined symbol error from protocol/server.so:
glusterfs_mgmt_pmap_signout
glusterfs_autoscale_threads
See https://review.gluster.org/19225 (bz#1532238)
and https://review.gluster.org/19657 (bz#1550895)
(why are there two different bzs for the same bug?)
IMO this is a cleaner solution. I.e. moving the above two functions
to libgfrpc (.../rpc/rpc-lib/...)
I would also, for (foolish) consistency sake, like to see
glusterfs_mgmt_pmap_signin() moved from glusterfsd to libgfrpc as
well.
This works on f28/rawhide, with its new, more restrictive run-time
link semantics. The smoke and regression tests on earlier fedora and
centos will confirm that it works on those platforms too.
Change-Id: I9cfbd1cc15e7ebd9fc31b56ac791287fa2c584de
BUG: 1550895
Signed-off-by: Kaleb S. KEITHLEY <kkeithle@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Scale rpcsvc_request_handler threads to match the scaling of event
handler threads.
Please refer to https://bugzilla.redhat.com/show_bug.cgi?id=1467614#c51
for a discussion about why we need multi-threaded rpcsvc request
handlers.
Change-Id: Ib6838fb8b928e15602a3d36fd66b7ba08999430b
Signed-off-by: Milind Changire <mchangir@redhat.com>
|
|
|
|
|
|
|
| |
All socket options update for GD2
Change-Id: I227c16965e92018a5ab5aacd9c2617fb2735268c
Signed-off-by: Mohit Agrawal <moagrawa@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Clients will request for a list of volfile servers from glusterd2 by
setting a (optional) flag in GETSPEC RPC call. glusterd2 will check for
the presence of this flag and accordingly return a list of glusterd2
servers in GETSPEC RPC reply. Currently, this list of servers returned
only contains servers which have bricks belonging to the volume.
See:
https://github.com/gluster/glusterd2/issues/382
https://github.com/gluster/glusterfs/issues/351
Updates #351
Change-Id: I0eee3d0bf25a87627e562380ef73063926a16b81
Signed-off-by: Prashanth Pai <ppai@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Building with --with-default-ipv6 causes shared
components of gluster calling the rpcbind6 functions
to fail. Adding the symbols in the list is all that is
necessary. Building without ipv6 keeps the same behavior.
No test cases as this is a build-specific fix.
Change-Id: I248d3291bf17326b07d152d9b79cdcfaf9068f0d
BUG: 1544961
Signed-off-by: Sheena Artrip <sheenobu@fb.com>
|
|
|
|
|
|
|
| |
updates #384
Change-Id: Id80bf470988dbecc69779de9eb64088559cb1f6a
Signed-off-by: Amar Tumballi <amarts@redhat.com>
|
|
|
|
|
|
| |
Updates #353
Change-Id: I755b9208690be76935d763688fa414521eba3a40
Signed-off-by: Poornima G <pgurusid@redhat.com>
|
|
|
|
|
| |
Change-Id: I4e6439d0f1ab55528c0f7e23c437c7b62628b8e4
Signed-off-by: Anoop C S <anoopcs@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
`make V=1` is broken — no commands are printed whatsoever. At the
same time, `make V=0` *also* is broken in that no summary lines
("CC foo.o") are printed, either. Kill the annoying hardcoded
--quiet in configure.ac, since it seems to override everything that
automake offers.
Change-Id: I377c0e0469619a33586afb4a93dde6d241e7bc21
Fixes: #381
BUG: 1539023
Original-author: Jan Engelhardt <jengelh@inai.de>
Signed-off-by: Niels de Vos <ndevos@redhat.com>
[ndevos: silence rpc/xdr/gen rpgcen]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Another error snuck in with Change-Id I86f847dfd, or more
accurately I think, with Change-Id: Ic47065e9c2...
All libs, not just libgfrpc, need to be linked with libtirpc,
especially on systems that still have xdr functions in (g)libc
where you will get a mixture of calls to libtirpc functions
and glibc functions, with catastrophic results.
BUG: 1536186
Change-Id: I97dc39c7844f44c36fe210aa813480c219e1e415
Signed-off-by: Kaleb S. KEITHLEY <kkeithle@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
With this patchset, some major things are changed in XDR, mainly:
* Naming: Instead of gfs3/gfs4 settle for gfx_ for xdr structures
* add iattx as a separate structure, and add conversion methods
* the *_rsp structure is now changed, and is also reduced in number
(ie, no need for different strucutes if it is similar to other response).
* use proper XDR methods for sending dict on wire.
Also, with the change of xdr structure, there are changes needed
outside of xlator protocol layer to handle these properly. Mainly
because the abstraction was broken to support 0-copy RDMA with payload
for write and read FOP. This made transport layer know about the xdr
payload, hence with the change of xdr payload structure, transport layer
needed to know about the change.
Updates #384
Change-Id: I1448fbe9deab0a1b06cb8351f2f37488cefe461f
Signed-off-by: Amar Tumballi <amarts@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Introduce another authentication header which can now send more data.
This is useful because this data can be common for all the fops, and
we don't need to change all the signatures.
As part of this, made rpc-clnt.c little more modular to support multiple
authentication structures.
stack.h changes are placeholder for the ctime etc, can be moved later
based on need.
updates #384
Change-Id: I6111c13cfd2ec92e2b4e9295896bf62a8a33b2c7
Signed-off-by: Amar Tumballi <amarts@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This new type helps to avoid excessive logs. It should be
set only in case of
* volume graph building (graph.y)
* dict unserialize
(happens once a dictionary is received on wire in old protocol)
All other dict set and get should have proper check and warning
logs if there is a mismatch.
updates #220
Change-Id: I1cccb304a877aa80c07aaac95f10f5005e35b9c5
Signed-off-by: Amar Tumballi <amarts@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The locks xlator now is able to send a contention notification to
the current owner of the lock.
This is only a notification that can be used to improve performance
of some client side operations that might benefit from extended
duration of lock ownership. Nothing is done if the lock owner decides
to ignore the message and to not release the lock. For forced
release of acquired resources, leases must be used.
Change-Id: I7f1ad32a0b4b445505b09908a050080ad848f8e0
Signed-off-by: Xavier Hernandez <xhernandez@datalab.es>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Without an export map (at link time) libgrpc and libgfxdr export over
150 and 450 symbols each, respectively. Many are not used by anything
else. (Unclear what the unused symbols are, some may be simple
sloppiness, e.g. not declaring functions static that should be. Others
may be intra-library calls that can't be static but aren't part of the
API, per se.)
By linking with an export map the number of exported symbols is
reduced to ~60 and ~250 respectively.
This parallels the similar change made to libglusterfs recently
and the older changes to the xlators to minimize the symbols that
are visible (exported) from the .so.
And I don't know, do we want to go all the way to symbol versions?
For these libs? And for libglusterfs?
fixes gluster/glusterfs#392
Change-Id: I9cdc3eee10e5f1408d7e7f2f29fad597c97e4003
Signed-off-by: Kaleb S. KEITHLEY <kkeithle@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Now with https://review.gluster.org/#/c/19056/ merged, we perform
dlopen() with RTLD_LOCAL so that every shared library is expected to
be explicitly linked against other required libraries.
"undefined symbol ..." was seen while trying to access GlusterFS volumes
via NFS-Ganesha/Samba using libgfapi resulting in an unusable state.
This is a follow-up patch to https://review.gluster.org/#/c/17659/
to make sure that we link libgfrpc too while making socket and rdma
rpc-transport shared libraries.
Change-Id: I9943cdc449c257ded3cb9f9f2becdd5784d1d82d
BUG: 1532238
Signed-off-by: Anoop C S <anoopcs@redhat.com>
|
|
|
|
|
|
|
| |
This reverts commit 56e5fdae74845dfec0ff7ad0c8fee77695d36ad5.
Change-Id: Ia62cee5440bbe8e23f5da9cff692d792091d544a
Signed-off-by: Milind Changire <mchangir@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
Added 2 more types which are present in gluster codebase, mainly
IATT and UUID.
Updates #203
Change-Id: Ib6d6d6aefb88c3494fbf93dcbe08d9979484968f
Signed-off-by: Amar Tumballi <amarts@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Issues:
[1] https://download.gluster.org/pub/gluster/glusterfs/static-analysis/master/glusterfs-coverity/2017-12-25-0bc22bef/cov-errors.txt
Resolved:
[1] rdma.c:128 var_deref_op: Dereferencing null pointer "post".
[2] rdma.c:677: Potentially overflowing expression
[3] rdma.c:4250: freed_arg: "rpc_transport_unref" frees "peer->trans".
[4] rdma.c:4644: var_deref_op: Dereferencing null pointer "rdma_ctx".
[5] rdma.c:4945: cond_false: Condition "rdma_ctx != NULL", taking false branch.
Change-Id: Iec38f118d645df4131739da412a6c741ebbd2f85
BUG: 789278
Signed-off-by: Yi Wang <wangyi@storswift.com>
|
|
|
|
|
|
|
|
|
|
|
| |
If the transport object is freed in rpc_transport_unref,
a notify of RPC_TRANSPORT_CLEANUP is push to rpc_clnt_notify,
where the rpc_clnt(contains conn) is freed.
After that, using of conn after rpc_transport_unref is use after freed.
Change-Id: I5cac8a8e7ced7c1079930080a12abf02d46667d5
Signed-off-by: Kinglong Mee <mijinlong@open-fs.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
RTLD_LOCAL is the default value for symbol visibility flag of dlopen()
in Linux and NetBSD. Using it avoids conflicts during symbol resolution.
This also allows us to detect xlators that have not been explicitly
linked with libraries that they use. This used to go unnoticed
when RTLD_GLOBAL was being used.
BUG: 1193929
Change-Id: I50db6ea14ffdee96596060c4d6bf71cd3c432f7b
Signed-off-by: Prashanth Pai <ppai@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch brings data type awareness to dictionary,
and also makes sure valid data is properly sent to the
other side of the wire using XDR.
Next step is to allow people to add more data types
(for example, Bool, UUID, iatt etc), and then make
it part of every fop signature in wire.
Fixes #203
Change-Id: Ie0eee2db847bea2bf7dad80dec89ce3e7c5917c1
Signed-off-by: Amar Tumballi <amarts@redhat.com>
|
|
|
|
|
|
|
|
|
| |
This patch creates a new way of defining message id's that is easier
and less error prone because it doesn't require so many manual changes
each time a new component is defined or a new message created.
Change-Id: I71ba8af9ac068f5add7e74f316a2478bc991c67b
Signed-off-by: Xavier Hernandez <jahernan@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Patch attempts to use the epoll infra for handling SSL connections
as well instead of the socket_poller() thread func.
This essentially makes priv->own_thread flag redundant.
SSL_connect()/SSL_accept() is now non-blocking which has done away
with the localised poll() in ssl_do(). So, ssl_do() has been updated
appropriately.
own_thread and coincidently socket_poller() thread for SSL processing
is now deprecated.
Added a timeout to test whether seal-heal daemon is up and running
as per Ravi's suggestion.
Change-Id: If2b5d7b4fd19e321cb289e08d49a718d2161aafe
Signed-off-by: Milind Changire <mchangir@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Problem: In glusterfs code base we call mutex_lock/unlock to take
reference/dereference for a object.Sometime it could be
reason for lock contention also.
Solution: There is no need to use mutex to increase/decrease ref
counter, instead of using mutex use gcc builtin ATOMIC
operation.
Test: I have not observed yet how much performance gain after apply
this patch specific to glusterfs but i have tested same
with below small program(mutex and atomic both) and
get good difference.
static int numOuterLoops;
static void *
threadFunc(void *arg)
{
int j;
for (j = 0; j < numOuterLoops; j++) {
__atomic_add_fetch (&glob, 1,__ATOMIC_ACQ_REL);
}
return NULL;
}
int
main(int argc, char *argv[])
{
int opt, s, j;
int numThreads;
pthread_t *thread;
int verbose;
int64_t n = 0;
if (argc < 2 ) {
printf(" Please provide 2 args Num of threads && Outer Loop\n");
exit (-1);
}
numThreads = atoi(argv[1]);
numOuterLoops = atoi (argv[2]);
if (1) {
printf("\tthreads: %d; outer loops: %d;\n",
numThreads, numOuterLoops);
}
thread = calloc(numThreads, sizeof(pthread_t));
if (thread == NULL) {
printf ("calloc error so exit\n");
exit (-1);
}
__atomic_store (&glob, &n, __ATOMIC_RELEASE);
for (j = 0; j < numThreads; j++) {
s = pthread_create(&thread[j], NULL, threadFunc, NULL);
if (s != 0) {
printf ("pthread_create failed so exit\n");
exit (-1);
}
}
for (j = 0; j < numThreads; j++) {
s = pthread_join(thread[j], NULL);
if (s != 0) {
printf ("pthread_join failed so exit\n");
exit (-1);
}
}
printf("glob value is %ld\n",__atomic_load_n (&glob,__ATOMIC_RELAXED));
exit(0);
}
time ./thr_count 800 800000
threads: 800; outer loops: 800000;
glob value is 640000000
real 1m10.288s
user 0m57.269s
sys 3m31.565s
time ./thr_count_atomic 800 800000
threads: 800; outer loops: 800000;
glob value is 640000000
real 0m20.313s
user 1m20.558s
sys 0m0.028
Change-Id: Ie5030a52ea264875e002e108dd4b207b15ab7cc7
Signed-off-by: Mohit Agrawal <moagrawa@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Problem: gf_rdma_device_t->all_mr is a __gf_rdma_arena_mr(includes MR content)
kind of list in the rdma rpc-transport. The rdma rpc-transport will
add/delete items to the list when MRs register, deregister, and free.
Because gf_rdma_device_t->all_mr is used by different threads
and it is not mutex protected, rdma transport maybe access obsolete
items in it.
Solution: Add a mutex protection for the gf_rdma_device_t->all_mr.
Change-Id: I2b7de0f7aa516b90bb6f3c6aae3aadd23b243900
BUG: 1522651
Signed-off-by: Yi Wang <wangyi@storswift.com>
|
|
|
|
|
|
| |
Change-Id: I22e622212f30defe6f2af1a67d7b48a88d37a097
BUG: 1520974
Signed-off-by: Shreyas Siravara <sshreyas@fb.com>
|
|
|
|
|
|
|
|
|
|
| |
* Introduce xlator methods to allow dumping of metrics
* Separate options to get the metrics dumped in a path
Updates #168
Change-Id: I7df80df33b71d6f449f03c2332665b4a45f6ddf2
Signed-off-by: Amar Tumballi <amarts@redhat.com>
|