- 23 Jul, 2013 40 commits
-
-
Andriy Skulysh authored
Deadlock isn't detected if 2 threads are trying to grant 2 locks which deadlock on each other. They call ldlm_flock_deadlock() simultaneously and deadlock ins't detected. The soulition is to add lock to blocking list before calling ldlm_flock_deadlock() Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-1602 Lustre-change: http://review.whamcloud.com/3277Signed-off-by:
Andriy Skulysh <Andriy_Skulysh@xyratex.com> Reviewed-by:
Vitaly Fertman <vitaly_fertman@xyratex.com> Reviewed-by:
Bruce Korb <bruce_korb@xyratex.com> Reviewed-by:
Keith Mannthey <keith.mannthey@intel.com> Reviewed-by:
Oleg Drokin <oleg.drokin@intel.com> Signed-off-by:
Peng Tao <tao.peng@emc.com> Signed-off-by:
Andreas Dilger <andreas.dilger@intel.com> Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
John L. Hammond authored
This partially reverts the change from "LU-2482 layout: introduce new layout for released files" by calling simple_setattr() from ll_md_setattr() without ATTR_SIZE set. Doing so avoids failed assertions in osc_page_delete(). Disable truncates on released files and modify sanity 229 accordingly. Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3448 Lustre-change: http://review.whamcloud.com/6643Signed-off-by:
John L. Hammond <john.hammond@intel.com> Reviewed-by:
jacques-Charles Lafoucriere <jacques-charles.lafoucriere@cea.fr> Reviewed-by:
Jinshan Xiong <jinshan.xiong@intel.com> Reviewed-by:
-
Andreas Dilger authored
When creating a hard link to a file, the MDT/MDD/OSD code does not verify whether the target link name already exists in the directory. The ZFS ZAP code checks for duplicate entries. The add_dirent_to_buf() function in ldiskfs only checks entries for duplicates while it is traversing the leaf block looking for free space. Even if it scanned the whole leaf block, this would not work for non-htree directories since there is no guarantee that the name is being inserted into the same leaf block. To fix this, link should check target object doesn't exist as other creat operations. Add sanity.sh test_31o with multiple threads racing to link a new name into the directory, while ensuring that there is a free entry in the leaf block that is large enough to hold the duplicate name. This needs to be racy, because otherwise the client VFS will see the existing name and not send the RPC to the MDS, hiding the bug. Add DLDLMRES/PLDLMRES macros for printing the whole lock resource name (including the name hash) in LDLM_DEBUG() messages in a format similar to DFID/PFID so they can be found in debug logs more easily. The patch pickes client side change of the original patch, which only contains the DLM printk part. Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2901 Lustre-change: http://review.whamcloud.com/6591Signed-off-by:
Lai Siyao <lai.siyao@intel.com> Reviewed-by:
Alex Zhuravlev <alexey.zhuravlev@intel.com> Reviewed-by:
Mike Pershin <mike.pershin@intel.com> Signed-off-by:
-
Alexander.Boyko authored
This patch add automatically module loading for crc32c when libcfs is starting. Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2212 Lustre-change: http://review.whamcloud.com/4372Signed-off-by:
Alexander Boyko <alexander_boyko@xyratex.com> Reviewed-by:
Shuichi Ihara <sihara@ddn.com> Reviewed-by:
James Simmons <uja.ornl@gmail.com> Signed-off-by:
-
Hiroya Nozaki authored
When ptlrpc_start_thread fails to create a new thread, it will finalize and free a struct ptlrpc_thread created and used here. Considering this, it can be a problem when ptlrpc_svcpt_stop_thread is driven and handles the struct ptlrpc_thread right after or right before failure of cfs_create_thread. Because this situation let the both of ptlrpc_start_thread and ptlrpc_svcpt_stop_threads access the freed ptlrpc_thread and cause OS panic. Or, it may happen that ptlrpc_svcpt_stop_threads waits forever holding an already-freed waitq. This patch adds an error handling into ptlrpc_start_thread to fix this problem. Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2889 Lustre-change: http://review.whamcloud.com/5552Signed-off-by:
Hiroya Nozaki <nozaki.hiroya@jp.fujitsu.com> Reviewed-by:
Liang Zhen <liang.zhen@intel.com> Reviewed-by:
Nikitas Angelinas <nikitas_angelinas@xyratex.com> Reviewed-by:
-
Keith Mannthey authored
lu_kmem_init can fail and returns has a return code. Check for this return code in lu_kmem_init. This issue was found during 2gb VM Racer testing Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3063 Lustre-change: http://review.whamcloud.com/6514Signed-off-by:
Nathaniel Clark <nathaniel.l.clark@intel.com> Reviewed-by:
Bob Glossman <bob.glossman@intel.com> Reviewed-by:
-
Fan Yong authored
Original patch resolves some LFSCK 1.5 technical debts, including: 1) Check and remove repeated linkea entries. 2) Merge some "goto" branches to make the code more readable. 3) Some comments about object's nlink inconsistency processing. This patch picks the obd flags change. Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2915 Lustre-change: http://review.whamcloud.com/6344Signed-off-by:
Fan Yong <fan.yong@intel.com> Reviewed-by:
-
jcl authored
When a file is restored the layout lock is first associated with the released layout and after restore it has to be assocaited with the new layout. This patch forces lvb_data update in ll_layout_fetch() even if one is present (case for released->normal state change) Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3299 Lustre-change: http://review.whamcloud.com/6291Signed-off-by:
Johann Lombardi <johann.lombardi@intel.com> Signed-off-by:
-
Liu Xuezhao authored
Split FID server from client, fid_{handler,store,lib}.c are not compliled unless server support is enabled. Generally cleanup includes in lustre/fid/ and reduce the need for client code to directly or indirectly include {dt,md}_object.h. Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-1330 Lustre-change: http://review.whamcloud.com/2673Signed-off-by:Liu Xuezhao <xuezhao.liu@emc.com> Signed-off-by:
-
Sebastien Buisson authored
Fix 'code maintainability' issues found by Coverity version 6.5.1: Unused pointer value (UNUSED_VALUE) Pointer returned by function is never used. Missing varargs init or cleanup (VARARGS) va_end was not called for variable. Signed-off-by:
Sebastien Buisson <sebastien.buisson@bull.net> Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3107 Lustre-change: http://review.whamcloud.com/5944Reviewed-by:
Dmitry Eremin <dmitry.eremin@intel.com> Reviewed-by:
John Hammond <johnlockwoodhammond@gmail.com> Reviewed-by:
-
jcl authored
move lov_mds_md_size from obd_lov.h to lustre_idl.h to have it close to lov_mds_md definition. add lov_user_md_size() to compute lum size so llapi and user space utils do not use kernel internal definitions/methods Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3345 Lustre-change: http://review.whamcloud.com/6345Signed-off-by:
-
jcl authored
add 2 macros for the length of a fid string 0xSEQ:0xOID:0xVER and it's brace version (FID_NOBRACE_LEN, and FID_LEN) Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2782 Lustre-change: http://review.whamcloud.com/5299Signed-off-by:
-
Jinshan Xiong authored
Released files now have a standard layout (with generation, pool, ...) and a stripe count 0 and lmm_pattern flag LOV_PATTERN_F_RELEASED. Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2482 Lustre-change: http://review.whamcloud.com/4816Signed-off-by:
-
John L. Hammond authored
In preparing Ie3a3cd99 (LU-1330 obdclass: splits server-side object stack from client) the lu_ucred infrastructure was put in its own file. Fixup the boilerplate of this file to give the proper path, short description, and authors. Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-1330 Lustre-change: http://review.whamcloud.com/5910Signed-off-by:
-
Bruce Korb authored
* lustre/include/lustre_dlm.h: Remove all bit fields and the unused weighing callback procedure. respell LDLM_AST_DISCARD_DATA as LDLM_FL_AST_DISCARD_DATA to match other flags. * .gitignore: ignore emacs temporary files * autogen.sh: rebuild the lock bits, if autogen is available. * contrib/bit-masks/lustre_dlm_flags.def: define the ldlm_lock flags * contrib/bit-masks/lustre_dlm_flags.tpl: template for emitting text * contrib/bit-masks/Makefile: construct the .c and .h files The .c file is for constructing a crash extension and is not preserved. * contrib/bit-masks/.gitignore: ignore built products * lustre/contrib/wireshark/packet-lustre.c: use built files instead of local versions of the defines. In the rest of the modified sources, replace flag field references with bit mask references. * lustre/osc/osc_lock.c: removed osc_lock_weigh, too Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2771 Lustre-change: http://review.whamcloud.com/5312Signed-off-by:
-
Oleg Drokin authored
It seems quota namespace is needlessly referenced on connect, but that's not necessary as it could not go away until entire obd goes away. On the other hand this extra reference disturbs other logic depending on empty namespace having zero refcount, so this patch drops such extra referencing. This picks client side change of the original patch. Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2924 Lustre-change: http://review.whamcloud.com/6234Signed-off-by:
Niu Yawei <yawei.niu@intel.com> Reviewed-by:
-
Oleg Drokin authored
We used to wake up ldlm poold every second, but that's overkill, we should just see how much time is left until next closest recalc interval hits and sleep this much. This will make "per-second" client grant statistic not actually per-second, but I don't think we need any precision in that code Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2924 Lustre-change: http://review.whamcloud.com/5793Signed-off-by:
-
Oleg Drokin authored
The main reason behind this is ldlm_poold walks all namespaces currently no matter if there are any locks or not. On large systems this could take quite a bit of time, esp. since ldlm_poold is currently woken up once per second. Now every time a client namespace loses it's last resource it is placed into an inactive list that is not touched by ldlm_poold as pointless. On creation of a first resource in a namespace it is placed back into the active list. Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2924 Lustre-change: http://review.whamcloud.com/5624Signed-off-by:
-
jcl authored
When a client issue an RPC to get a layout lock, it must not hold rpc_lock because in case of a restore the rpc can be blocking for a long time Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3200 Lustre-change: http://review.whamcloud.com/6115Signed-off-by:
-
Li Wei authored
Lustre puts system errors (e.g., ENOTCONN) on wire as numbers essentially specific to senders' architectures. While this is fine for x86-only sites, where receivers share the same error number definition with senders, problems will arise, however, for sites involving multiple architectures with different error number definitions. For instance, an ENOTCONN reply from a sparc server will be put on wire as -57, which, for an x86 client, means EBADSLT instead. To solve the problem, this patch defines a set of network errors for on-wire or on-disk uses. These errors correspond to a subset of the x86 system errors and share the same number definition, maintaining compatibility with existing x86 clients and servers. Then, either error numbers could be translated at run time, or all host errors going on wire could be replaced with network errors in the code. This patch does the former by introducing both generic and field-specific translation routines and calling them at proper places, so that translations for existing fields are transparent. (Personally, I tend to think the latter way might be worthwhile, as it is more straightforward conceptually. Do we really need so many different errors? Should errors returned by kernel routines really be passed up and eventually put on wire? There could even be security implications in that.) Thank Fujitsu for the original idea and their contributions that make this available upstream. Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2743 Lustre-change: http://review.whamcloud.com/5577Signed-off-by:
Li Wei <wei.g.li@intel.com> Reviewed-by:
-
Dmitry Eremin authored
The race is result of use-after-free situation: ~ ptlrpc_stop_pinger() ~ ptlrpc_pinger_main() --------------------------------------------------------------- thread_set_flags(SVC_STOPPING) cfs_waitq_signal(pinger_thread) ... ... thread_set_flags(SVC_STOPPED) l_wait_event(thread_is_stopped) OBD_FREE_PTR(pinger_thread) ... cfs_waitq_signal(pinger_thread) --------------------------------------------------------------- The memory used by pinger_thread might have been freed and reallocated to something else, when ptlrpc_pinger_main() used it in cvs_waitq_signal(). Signed-off-by:
Faccini Bruno <bruno.faccini@intel.com> Reviewed-by:
-
Andreas Dilger authored
Print the namespace and OBD device name, as well as the first two lock resource fields (typically the FID) if there is an error with loading the object from disk. This will be more important with FID-on-OST and also the MDS. Using fid_extract_from_res_name() isn't possible in the LDLM code, since the lock resource may not be a FID. Make fid_extract_quota_resid() argument order and name consistent with other fid_*_res() functions, with FID first and resource second. Fix a bug in ofd_lvbo_init() where NULL lvb is accessed on error. Print FID in ofd_lvbo_update() CDEBUG() and CERROR() messages. Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2193 Lustre-change: http://review.whamcloud.com/4501Signed-off-by:
Bobi Jam <bobijam@gmail.com> Reviewed-by:
-
John L. Hammond authored
In ll_file_ioctl() and ll_swap_layouts() check the result of ll_prep_md_op_data() using IS_ERR(). Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3283 Lustre-change: http://review.whamcloud.com/6275Signed-off-by:
-
Dmitry Eremin authored
In case of memory pressure a not locked mutex can be unlocked in function ll_file_open(). This is not allowed and subsequent behavior is not defined. Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3157 Lustre-change: http://review.whamcloud.com/6028Signed-off-by:
-
John L. Hammond authored
In ll_file_data_get() and ll_dir_ioctl() return error on failed allocations. Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2753 Lustre-change: http://review.whamcloud.com/5845Signed-off-by:
-
Sebastien Buisson authored
Fix 'program hangs' defects found by Coverity version 6.5.1: Missing unlock (LOCK) Returning without unlocking. Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3054 Lustre-change: http://review.whamcloud.com/5870Signed-off-by:
-
wang di authored
Missing the last bit during INODELOCK check in ll_have_md_lock. Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3385 Lustre-change: http://review.whamcloud.com/6438Signed-off-by:
wang di <di.wang@intel.com> Reviewed-by:
-
John L. Hammond authored
In vvp_io_write_start() the stats function ll_rw_stats_tally() was incorrectly called with a rw argument of 0. Correct this and use the macros READ and WRITE in and around ll_rw_stats_tally() for clarity. Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3384 Lustre-change: http://review.whamcloud.com/6447Signed-off-by:
-
Peng Tao authored
Signed-off-by:
-
Paul Bolle authored
Commit 4b5b4c72 ("staging/lustre/libcfs: restore LINVRNT") added "default false" to this Kconfig file. It was obviously meant to use "default n" here. But we might as well drop this line, as a Kconfig bool defaults to 'n' anyway. Signed-off-by:
Paul Bolle <pebolle@tiscali.nl> Signed-off-by:
-
Dragos Foianu authored
Confirmed by cscope that the functions are not used anymore. A fresh compilation does not yield any errors. Signed-off-by:
Dragos Foianu <dragos.foianu@gmail.com> Reviewed-by:
-
Zhao Hongjiang authored
kmap_atomic allows only one argument now, just remove the second. Signed-off-by:
Zhao Hongjiang <zhaohongjiang@huawei.com> Signed-off-by:
-
Peng Tao authored
This reverts commit 0ad1ea69 I didn't use git revert because it can not be done cleanly. Hopefully it will be the last time we do it... Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by:
Peng Tao <bergwolf@gmail.com> Signed-off-by:
-
Peng Tao authored
Signed-off-by:
-
Peng Tao authored
Signed-off-by:
-
Stephen Rothwell authored
somehow this got dropped during merge window... Signed-off-by:
Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by:
-
Peng Tao authored
Building on 32bit system, I got warnings like below: drivers/staging/lustre/lustre/llite/../include/lprocfs_status.h:666:7: note: expected ‘long unsigned int *’ but argument is of type ‘size_t *’ char *lprocfs_find_named_value(const char *buffer, const char *name, drivers/staging/lustre/lustre/lov/lov_io.c: In function ‘lov_io_rw_iter_init’: include/asm-generic/div64.h:43:28: warning: comparison of distinct pointer types lacks a cast [enabled by default] (void)(((typeof((n)) *)0) == ((uint64_t *)0)); \ Signed-off-by:
-
Peng Tao authored
dump_trace() is only available on X86. Without it, Lustre's own watchdog is broken. We can only dump current task's stack. The client-side this code is much less likely to hit deadlocks and it's probably OK to drop this altogether, since we hardly have any ptlrpc threads on clients, most notable ones are ldlm cb threads that should not really be blocking on the client anyway. Remove libcfs watchdog for now, until the upstream kernel watchdog can detect distributed deadlocks and dump other kernel threads. Cc: Oleg Drokin <oleg.drokin@intel.com> Signed-off-by:
-
Peng Tao authored
kuid_t/kgid_t are wrappered when CONFIG_UIDGID_STRICT_TYPE_CHECKS is on. Lustre build is broken because we always treat them as plain __u32. The patch fixes it. Internally, Lustre always use __u32 uid/gid, and convert to kuid_t/kgid_t when necessary. Signed-off-by:
-
Peng Tao authored
Signed-off-by:
-
