RHEL6 x86_64

Lai

Could you please look at this one

Thanks

Peter

There are several places where in memory struct may be released or un-initialized before proc entry creation/removal for mgs/mds/obdfilter.

I tested with the below scripts, and it could pass:

while true; do
    umount -d /mnt/ost1
    mount -t lustre <OST_DEVICE> /mnt/ost1
done

while true; do
    cat /proc/fs/lustre/obdfilter/lustre-OST0000/*
    cat /proc/fs/lustre/obdfilter/lustre-OST0000/export/*/*
done

These two scripts are running concurrently for mds/obdfilter.

Patch is at http://review.whamcloud.com/#change,325.

Integrated in reviews-centos5 #442
LU-106 unable to handle kernel pagig request in lprocfs_stats_collect()

Lai Siyao : 28030d340504ce9a9e938d7a5a7b65205a82cc7b
Files :

• lustre/mds/handler.c
• lustre/mds/lproc_mds.c
• lustre/obdclass/lprocfs_status.c
• lustre/obdfilter/filter.c
• lustre/mgs/lproc_mgs.c

Integrated in reviews-centos5 #444
LU-106 unable to handle kernel paging request in lprocfs_stats_collect()

Lai Siyao : 6b85a6cc4348274e29cc6d26bd958cc977c16207
Files :

• lustre/mgs/lproc_mgs.c
• lustre/obdclass/lprocfs_status.c
• lustre/mdt/mdt_lproc.c
• lustre/mdt/mdt_handler.c
• lustre/obdfilter/filter.c

Integrated in reviews-centos5 #458
LU-106 unable to handle kernel pagig request in lprocfs_stats_collect()

Lai Siyao : e7146c54ba1ae58e796f66c9f1619d42bc90c025
Files :

• lustre/obdclass/lprocfs_status.c
• lustre/mgs/lproc_mgs.c
• lustre/obdfilter/filter.c
• lustre/mds/handler.c
• lustre/include/obd_class.h
• lustre/mds/lproc_mds.c

Integrated in reviews-centos5 #459
LU-106 unable to handle kernel paging request in lprocfs_stats_collect()

Lai Siyao : 95a83d156c0b6e63d4b6ef197393839c6f9869c7
Files :

• lustre/mdt/mdt_handler.c
• lustre/mgs/lproc_mgs.c
• lustre/obdfilter/filter.c
• lustre/include/obd_class.h
• lustre/mdt/mdt_lproc.c
• lustre/obdclass/lprocfs_status.c

Integrated in reviews-centos5 #460
LU-106 unable to handle kernel paging request in lprocfs_stats_collect()

Lai Siyao : 95a83d156c0b6e63d4b6ef197393839c6f9869c7
Files :

• lustre/mdt/mdt_lproc.c
• lustre/mdt/mdt_handler.c
• lustre/obdclass/lprocfs_status.c
• lustre/obdfilter/filter.c
• lustre/include/obd_class.h
• lustre/mgs/lproc_mgs.c

Integrated in reviews-centos5 #463
LU-106 unable to handle kernel pagig request in lprocfs_stats_collect()

Lai Siyao : e7146c54ba1ae58e796f66c9f1619d42bc90c025
Files :

• lustre/obdfilter/filter.c
• lustre/mgs/lproc_mgs.c
• lustre/obdclass/lprocfs_status.c
• lustre/mds/lproc_mds.c
• lustre/include/obd_class.h
• lustre/mds/handler.c

Integrated in reviews-centos5 #468
LU-106 unable to handle kernel paging request in lprocfs_stats_collect()

Lai Siyao : 496d18c30da3998711bb3ab3379a486a4739b324
Files :

• lustre/mdt/mdt_handler.c
• lustre/mdt/mdt_lproc.c
• lustre/obdclass/lprocfs_status.c
• lustre/mgs/lproc_mgs.c
• lustre/obdfilter/filter.c
• lustre/include/obd_class.h

Integrated in reviews-centos5 #533
LU-106 unable to handle kernel paging request in lprocfs_stats_collect()

Lai Siyao : 8a21280779de8135e0255e13582053e6195cc3e1
Files :

• lustre/obdclass/lprocfs_status.c
• lustre/mdt/mdt_lproc.c
• lustre/mgs/lproc_mgs.c
• lustre/include/obd_class.h
• lustre/obdfilter/filter.c
• lustre/mdt/mdt_handler.c

Review is on http://review.whamcloud.com/#change,326.

Maloo result is on https://maloo.whamcloud.com/test_sessions/a3e23b96-6588-11e0-b32b-52540025f9af.

maloo result is on https://maloo.whamcloud.com/test_sessions/c057691e-70db-11e0-80b5-52540025f9af, and looks okay.
Oleg, do you think it's okay to land this?

Single node test w/ patch: https://maloo.whamcloud.com/test_sessions/9a08f0b6-750e-11e0-a1b3-52540025f9af

Prakash\Ned

When do you think LLNL might be able to test this at scale? I have discussed this patch with Oleg and he would prefer to get some scale testing of this patch (preferably using your monitoring tool) before this is landed. It would be great if we were able to land this patch in time for 2.1 but we probably would not hold the release just for this so I am dropping the priority to reflect that

Regards

Peter

Hi Peter,

We are currently waiting on getting our test server cluster updated to RHEL6 and Lustre 2.x. I need to check with our system administrators on the timeframe, but my guess is that this should happen within the next two weeks or so. They've been heavily focused on other system integration projects which have priority over our test infrastructure. Once that work is done we can start at scale testing.

Thanks,
Ned

What testing have you guys done at small scale?

I honestly don't see much to gain from scale testing except that we'll have more simultaneous small scale tests...

The test for this should look like this:

• Have a script walk and cat files in lustre /proc tree will repeatedly starting and stopping the OSTs. Run for days.

We can help with that, but I don't really think scale is an issue here.

Chris

I think that our concern is more around unexpected side-effects from this patch not just whether the patch corrects the reported issue.

Lai

Have you tried running tests as Chris suggests in order to validate this patch? If not, could you please look into doing so?

Thanks

Peter

Peter, in the second comment I described how I tested locally: cat files under /proc/fs/lustre/[mdt|obdfilter], and it's tested for tens of minutes. To follow Chris' suggest, I need walk and cat more files under /proc/fs/lustre and /proc/sys/[lnet|lustre] and test up to 1 day. I will test it later and post result.

Unable to handle kernel NULL pointer dereference at 0000000000000028 RIP:
[<ffffffff80023bb2>] filp_close+0x7/0x64
PGD 2d0057067 PUD 2d004d067 PMD 0
Oops: 0000 [1] SMP
last sysfs file: /devices/pci0000:00/0000:00:01.0/0000:01:00.1/irq
CPU 3
Modules linked in: llite_lloop(U) lustre(U) obdfilter(U) ost(U) osd_ldiskfs(U) cmm(U) fsfilt_ldiskfs(U) exportfs(U) mdt(U) mdd(U) mds(U) mgs(U) ldiskfs(U) jbd2(U) crc16(U) mgc(U) lov(U) osc(U) mdc(U) lmv(U) fid(U) fld(U) lquota(U) ptlrpc(U) obdclass(U) lvfs(U) ksocklnd(U) lnet(U) libcfs(U) nfs(U) fscache(U) nfs_acl(U) autofs4(U) hidp(U) rfcomm(U) l2cap(U) bluetooth(U) lockd(U) sunrpc(U) cpufreq_ondemand(U) acpi_cpufreq(U) freq_table(U) be2iscsi(U) iscsi_tcp(U) bnx2i(U) cnic(U) uio(U) cxgb3i(U) iw_cxgb3(U) cxgb3(U) libiscsi_tcp(U) ib_iser(U) libiscsi2(U) scsi_transport_iscsi2(U) scsi_transport_iscsi(U) ib_srp(U) rds(U) ib_sdp(U) ib_ipoib(U) ipoib_helper(U) ipv6(U) xfrm_nalgo(U) crypto_api(U) rdma_ucm(U) rdma_cm(U) ib_ucm(U) ib_uverbs(U) ib_umad(U) ib_cm(U) iw_cm(U) ib_addr(U) ib_sa(U) loop(U) dm_mirror(U) dm_multipath(U) scsi_dh(U) video(U) backlight(U) sbs(U) power_meter(U) hwmon(U) i2c_ec(U) dell_wmi(U) wmi(U) button(U) battery(U) asus_acpi(U) acpi_memhotplug(U) ac(U) parport_pc(U) lp(U) parport(U) mlx4_ib(U) ib_mad(U) ib_core(U) mlx4_en(U) joydev(U) sg(U) igb(U) 8021q(U) serio_raw(U) i2c_i801(U) i2c_core(U) mlx4_core(U) shpchp(U) dca(U) pcspkr(U) dm_raid45(U) dm_message(U) dm_region_hash(U) dm_log(U) dm_mod(U) dm_mem_cache(U) ahci(U) libata(U) sd_mod(U) scsi_mod(U) ext3(U) jbd(U) uhci_hcd(U) ohci_hcd(U) ehci_hcd(U)
Pid: 32369, comm: cat Tainted: G 2.6.18-194.17.1.el5_lustre.g4a053f7 #1
RIP: 0010:[<ffffffff80023bb2>] [<ffffffff80023bb2>] filp_close+0x7/0x64
RSP: 0018:ffff8102d17afc98 EFLAGS: 00010246
RAX: 0000000000000001 RBX: 0000000000000001 RCX: ffff8102d17afb40
RDX: 0000000000000003 RSI: 0000000000000000 RDI: 0000000000000000
RBP: 0000000000000000 R08: ffff8102d17ae000 R09: 000000000000003d
R10: ffff81010b7e4068 R11: 0000000000000080 R12: ffff8102db2d8928
R13: ffff8102db2d8670 R14: ffff8102d17afce8 R15: 0000000019844000
FS: 00002abeea526260(0000) GS:ffff81010b7636c0(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 0000000000000028 CR3: 00000002d165b000 CR4: 00000000000006e0
Process cat (pid: 32369, threadinfo ffff8102d17ae000, task ffff8102da4fd860)
Stack: 0000000000000001 ffff8102db2d83f8 ffff8102db2d8928 ffffffff88e43dbd
0000003000000030 ffff8102d17afdf0 ffff8102d17afcd8 0000000000000004
0000000000000000 0000000000000000 ffff81033f5cd580 ffff81030d8019c0
Call Trace:
[<ffffffff88e43dbd>] :obdfilter:filter_post+0x11d/0x240
[<ffffffff88e5629f>] :obdfilter:filter_cleanup+0x15f/0x290
[<ffffffff888bb31f>] :obdclass:class_decref+0x43f/0x5b0
[<ffffffff888a2215>] :obdclass:obd_devlist_next+0x65/0x80
[<ffffffff8888a124>] :obdclass:obd_device_list_seq_start+0x34/0x50
[<ffffffff80063b05>] mutex_lock+0xd/0x1d
[<ffffffff8003f649>] seq_read+0xe5/0x28c
[<ffffffff8000b729>] vfs_read+0xcb/0x171
[<ffffffff80011c15>] sys_read+0x45/0x6e
[<ffffffff8005d28d>] tracesys+0xd5/0xe0

Peter, during stress test I met one crash, this looks to be another issue related with procfs, should I fix it along in this ticket?

crash> bt 22333
PID: 22333 TASK: ffff8102e3a46860 CPU: 3 COMMAND: "umount"
#0 [ffff8102e48e9c58] schedule at ffffffff80062f96
#1 [ffff8102e48e9d30] schedule_timeout at ffffffff800638ab
#2 [ffff8102e48e9d80] server_put_super at ffffffff88861a01
#3 [ffff8102e48e9e60] generic_shutdown_super at ffffffff800e3f3f
#4 [ffff8102e48e9e80] kill_anon_super at ffffffff800e400f
#5 [ffff8102e48e9e90] deactivate_super at ffffffff800e40c0
#6 [ffff8102e48e9eb0] sys_umount at ffffffff800edf11

server_put_super() is at server_wait_finished(), it means this obd has been cleaned up. While the crash backtrace shows the obd is been cleaned up again. After checking obd_device_list_seq related code, I found `cat /proc/fs/lustre/devices` will class_getref() on obd, and then class_putref() later, but doing this on a cleaned up obd may cause obd_cleanup() called twice, and finally cause this crash.

I will add obd->obd_stopping check in obd_devlist_first/next() to fix this.

> Peter, during stress test I met one crash, this looks to be another issue related with procfs, should I fix it along in this ticket?

Yes but perhaps you should track it under a new ticket if it is a distinct issue

Ah it seems that you have already corrected this existing patch to address the new issue. That is fine

Yes, because it blocked the stress test.

I've run the stress test without failure on two setup, one for 8 hours, and another 24 hours. The test script is as below:

• one shell running walk.sh:

  while true; do
          cat /proc/fs/lustre/*
          cat /proc/fs/lustre/*/*
          cat /proc/fs/lustre/*/*/*
          cat /proc/fs/lustre/*/*/*/*
  
          cat /proc/sys/lnet/*
          cat /proc/sys/lustre/*
  
          usleep 20
  done
  

• the other one running failover_ost.sh:

  while true; do
          umount -d /mnt/ost1
          sleep 1
          mount -t lustre -o loop /tmp/lustre-ost1 /mnt/ost1
          sleep 9
  done
  

We have found one more race condition in lprocfs. In osd_device_alloc(), if a /proc read happens in between dt_device_init() and osd_device initialization, a kernel crash is triggered by functions like osd_statfs().

A reproducer:

1. patch lustre with the following

diff --git a/lustre/osd-ldiskfs/osd_handler.c b/lustre/osd-ldiskfs/osd_handler.c 
index 29764c9..728c371 100644 
--- a/lustre/osd-ldiskfs/osd_handler.c 
+++ b/lustre/osd-ldiskfs/osd_handler.c 
@@ -3874,6 +3874,9 @@ static int osd_device_init(const struct lu_env *env, struct lu_device *d, 
         rc = lu_context_init(ctx, LCT_MD_THREAD|LCT_REMEMBER|LCT_NOREF); 
         if (rc == 0) { 
                 rc = osd_procfs_init(osd_dev(d), name); 
+ 
+ cfs_schedule_timeout_and_set_state(CFS_TASK_UNINT, cfs_time_seconds(10)); 
+ 
                 ctx->lc_cookie = 0x3; 
         } 
         return rc; 

2. run llmount.sh
3. wait a few seconds and run "find /proc/fs/lustre -name 'filesfree' -exec cat {} \;"

We are cooking a fix in the context of MRP-275.

The following patch (applied to other pieces of osd-lprocfs as well) fixes the issue for us:

@@ -200,7 +250,17 @@ int lprocfs_osd_rd_filesfree(char *page, char **start, off_t off, int count,
                              int *eof, void *data)
 {
         struct osd_device *osd = data;
-        int rc = osd_statfs(NULL, &osd->od_dt_dev, &osd->od_kstatfs);
+        int rc;
+
+        /*
+         * We possibly haven't mounted yet. When od_mount is assigned, it
+         * should be consistent and should not deinit until lprocfs is
+         * fully unregistered.
+         */
+        if (unlikely(osd->od_mount == NULL))
+                return -EINPROGRESS;
+
+        rc = osd_statfs(NULL, &osd->od_dt_dev, &osd->od_kstatfs);
         if (!rc) {
                 *eof = 1;
                 rc = snprintf(page, count, LPU64"\n", osd->od_kstatfs.f_ffree);

The race described above will work just like that:

[root@panda ~]# find /proc/fs/lustre -name 'filesfree' -exec cat {} \;
cat: /proc/fs/lustre/osd-ldiskfs/lustre-MDT0000/filesfree: Operation now in progress
cat: /proc/fs/lustre/lov/lustre-MDT0000-mdtlov/filesfree: Input/output error

Nice catch, I'll go through the code and see whether other places needs this check too, and then I'll include this fix. Thanks!

Fix included in latest patch: http://review.whamcloud.com/#change,326.

Landed for 2.2

Integrated in lustre-master » x86_64,client,sles11,inkernel #378
LU-106 procfs: many proc entries are not accessed safely (Revision ad538dba7f5259df31fc45e90835edc12756db40)

Result = SUCCESS
Oleg Drokin : ad538dba7f5259df31fc45e90835edc12756db40
Files :

• lustre/obdclass/obd_config.c
• lustre/osd-ldiskfs/osd_lproc.c
• lustre/include/lustre_fld.h
• lustre/include/lprocfs_status.h
• lustre/mdt/mdt_handler.c
• lustre/mdt/mdt_lproc.c
• lustre/lov/lov_obd.c
• lustre/obdfilter/filter.c
• lustre/cmm/cmm_device.c
• lustre/lmv/lmv_obd.c
• lustre/obdclass/lprocfs_status.c
• lustre/fld/fld_request.c
• lustre/mgs/mgs_handler.c
• lustre/osc/osc_request.c
• lustre/mdc/mdc_request.c

Integrated in lustre-master » x86_64,client,el6,inkernel #378
LU-106 procfs: many proc entries are not accessed safely (Revision ad538dba7f5259df31fc45e90835edc12756db40)

Result = SUCCESS
Oleg Drokin : ad538dba7f5259df31fc45e90835edc12756db40
Files :

• lustre/osd-ldiskfs/osd_lproc.c
• lustre/fld/fld_request.c
• lustre/lmv/lmv_obd.c
• lustre/obdclass/obd_config.c
• lustre/mdt/mdt_handler.c
• lustre/cmm/cmm_device.c
• lustre/obdfilter/filter.c
• lustre/mdc/mdc_request.c
• lustre/include/lustre_fld.h
• lustre/lov/lov_obd.c
• lustre/osc/osc_request.c
• lustre/include/lprocfs_status.h
• lustre/mgs/mgs_handler.c
• lustre/mdt/mdt_lproc.c
• lustre/obdclass/lprocfs_status.c

Integrated in lustre-master » x86_64,server,el6,inkernel #378
LU-106 procfs: many proc entries are not accessed safely (Revision ad538dba7f5259df31fc45e90835edc12756db40)

Result = SUCCESS
Oleg Drokin : ad538dba7f5259df31fc45e90835edc12756db40
Files :

• lustre/mdt/mdt_lproc.c
• lustre/include/lustre_fld.h
• lustre/mgs/mgs_handler.c
• lustre/cmm/cmm_device.c
• lustre/fld/fld_request.c
• lustre/lmv/lmv_obd.c
• lustre/obdclass/lprocfs_status.c
• lustre/lov/lov_obd.c
• lustre/obdclass/obd_config.c
• lustre/osc/osc_request.c
• lustre/mdc/mdc_request.c
• lustre/mdt/mdt_handler.c
• lustre/obdfilter/filter.c
• lustre/osd-ldiskfs/osd_lproc.c
• lustre/include/lprocfs_status.h

Integrated in lustre-master » x86_64,client,el5,ofa #378
LU-106 procfs: many proc entries are not accessed safely (Revision ad538dba7f5259df31fc45e90835edc12756db40)

Result = SUCCESS
Oleg Drokin : ad538dba7f5259df31fc45e90835edc12756db40
Files :

• lustre/include/lprocfs_status.h
• lustre/cmm/cmm_device.c
• lustre/fld/fld_request.c
• lustre/lov/lov_obd.c
• lustre/obdclass/obd_config.c
• lustre/mdc/mdc_request.c
• lustre/mdt/mdt_lproc.c
• lustre/osc/osc_request.c
• lustre/mgs/mgs_handler.c
• lustre/mdt/mdt_handler.c
• lustre/obdclass/lprocfs_status.c
• lustre/lmv/lmv_obd.c
• lustre/osd-ldiskfs/osd_lproc.c
• lustre/obdfilter/filter.c
• lustre/include/lustre_fld.h

Integrated in lustre-master » i686,client,el6,inkernel #378
LU-106 procfs: many proc entries are not accessed safely (Revision ad538dba7f5259df31fc45e90835edc12756db40)

Result = SUCCESS
Oleg Drokin : ad538dba7f5259df31fc45e90835edc12756db40
Files :

• lustre/mdt/mdt_handler.c
• lustre/osd-ldiskfs/osd_lproc.c
• lustre/obdclass/lprocfs_status.c
• lustre/mdt/mdt_lproc.c
• lustre/osc/osc_request.c
• lustre/lmv/lmv_obd.c
• lustre/mgs/mgs_handler.c
• lustre/include/lustre_fld.h
• lustre/mdc/mdc_request.c
• lustre/obdclass/obd_config.c
• lustre/cmm/cmm_device.c
• lustre/obdfilter/filter.c
• lustre/fld/fld_request.c
• lustre/include/lprocfs_status.h
• lustre/lov/lov_obd.c

Integrated in lustre-master » x86_64,client,el5,inkernel #378
LU-106 procfs: many proc entries are not accessed safely (Revision ad538dba7f5259df31fc45e90835edc12756db40)

Result = SUCCESS
Oleg Drokin : ad538dba7f5259df31fc45e90835edc12756db40
Files :

• lustre/obdclass/obd_config.c
• lustre/cmm/cmm_device.c
• lustre/mdt/mdt_lproc.c
• lustre/osd-ldiskfs/osd_lproc.c
• lustre/include/lustre_fld.h
• lustre/lmv/lmv_obd.c
• lustre/obdclass/lprocfs_status.c
• lustre/lov/lov_obd.c
• lustre/include/lprocfs_status.h
• lustre/fld/fld_request.c
• lustre/mdc/mdc_request.c
• lustre/osc/osc_request.c
• lustre/mdt/mdt_handler.c
• lustre/mgs/mgs_handler.c
• lustre/obdfilter/filter.c

Integrated in lustre-master » x86_64,client,ubuntu1004,inkernel #378
LU-106 procfs: many proc entries are not accessed safely (Revision ad538dba7f5259df31fc45e90835edc12756db40)

Result = SUCCESS
Oleg Drokin : ad538dba7f5259df31fc45e90835edc12756db40
Files :

• lustre/osc/osc_request.c
• lustre/lov/lov_obd.c
• lustre/obdclass/lprocfs_status.c
• lustre/mdt/mdt_handler.c
• lustre/cmm/cmm_device.c
• lustre/obdclass/obd_config.c
• lustre/lmv/lmv_obd.c
• lustre/fld/fld_request.c
• lustre/osd-ldiskfs/osd_lproc.c
• lustre/mgs/mgs_handler.c
• lustre/include/lprocfs_status.h
• lustre/mdc/mdc_request.c
• lustre/obdfilter/filter.c
• lustre/mdt/mdt_lproc.c
• lustre/include/lustre_fld.h

Integrated in lustre-master » x86_64,server,el5,inkernel #378
LU-106 procfs: many proc entries are not accessed safely (Revision ad538dba7f5259df31fc45e90835edc12756db40)

Result = SUCCESS
Oleg Drokin : ad538dba7f5259df31fc45e90835edc12756db40
Files :

• lustre/mdt/mdt_handler.c
• lustre/lmv/lmv_obd.c
• lustre/cmm/cmm_device.c
• lustre/mgs/mgs_handler.c
• lustre/osc/osc_request.c
• lustre/include/lustre_fld.h
• lustre/obdclass/lprocfs_status.c
• lustre/include/lprocfs_status.h
• lustre/lov/lov_obd.c
• lustre/osd-ldiskfs/osd_lproc.c
• lustre/mdt/mdt_lproc.c
• lustre/mdc/mdc_request.c
• lustre/obdclass/obd_config.c
• lustre/obdfilter/filter.c
• lustre/fld/fld_request.c

Integrated in lustre-master » i686,client,el5,ofa #378
LU-106 procfs: many proc entries are not accessed safely (Revision ad538dba7f5259df31fc45e90835edc12756db40)

Result = SUCCESS
Oleg Drokin : ad538dba7f5259df31fc45e90835edc12756db40
Files :

• lustre/mdc/mdc_request.c
• lustre/obdfilter/filter.c
• lustre/mgs/mgs_handler.c
• lustre/mdt/mdt_lproc.c
• lustre/include/lprocfs_status.h
• lustre/osd-ldiskfs/osd_lproc.c
• lustre/cmm/cmm_device.c
• lustre/mdt/mdt_handler.c
• lustre/obdclass/lprocfs_status.c
• lustre/fld/fld_request.c
• lustre/obdclass/obd_config.c
• lustre/lmv/lmv_obd.c
• lustre/osc/osc_request.c
• lustre/include/lustre_fld.h
• lustre/lov/lov_obd.c

Integrated in lustre-master » i686,client,el5,inkernel #378
LU-106 procfs: many proc entries are not accessed safely (Revision ad538dba7f5259df31fc45e90835edc12756db40)

Result = SUCCESS
Oleg Drokin : ad538dba7f5259df31fc45e90835edc12756db40
Files :

• lustre/obdclass/lprocfs_status.c
• lustre/obdfilter/filter.c
• lustre/include/lprocfs_status.h
• lustre/fld/fld_request.c
• lustre/mdc/mdc_request.c
• lustre/osd-ldiskfs/osd_lproc.c
• lustre/osc/osc_request.c
• lustre/lmv/lmv_obd.c
• lustre/mdt/mdt_lproc.c
• lustre/obdclass/obd_config.c
• lustre/mdt/mdt_handler.c
• lustre/include/lustre_fld.h
• lustre/mgs/mgs_handler.c
• lustre/cmm/cmm_device.c
• lustre/lov/lov_obd.c

Integrated in lustre-master » x86_64,server,el5,ofa #378
LU-106 procfs: many proc entries are not accessed safely (Revision ad538dba7f5259df31fc45e90835edc12756db40)

Result = FAILURE
Oleg Drokin : ad538dba7f5259df31fc45e90835edc12756db40
Files :

• lustre/osc/osc_request.c
• lustre/include/lprocfs_status.h
• lustre/lmv/lmv_obd.c
• lustre/obdclass/lprocfs_status.c
• lustre/cmm/cmm_device.c
• lustre/obdfilter/filter.c
• lustre/obdclass/obd_config.c
• lustre/fld/fld_request.c
• lustre/lov/lov_obd.c
• lustre/mdc/mdc_request.c
• lustre/mdt/mdt_lproc.c
• lustre/mdt/mdt_handler.c
• lustre/mgs/mgs_handler.c
• lustre/include/lustre_fld.h
• lustre/osd-ldiskfs/osd_lproc.c

Integrated in lustre-master » i686,server,el5,inkernel #378
LU-106 procfs: many proc entries are not accessed safely (Revision ad538dba7f5259df31fc45e90835edc12756db40)

Result = SUCCESS
Oleg Drokin : ad538dba7f5259df31fc45e90835edc12756db40
Files :

• lustre/mgs/mgs_handler.c
• lustre/osc/osc_request.c
• lustre/include/lustre_fld.h
• lustre/osd-ldiskfs/osd_lproc.c
• lustre/lmv/lmv_obd.c
• lustre/mdt/mdt_handler.c
• lustre/obdfilter/filter.c
• lustre/mdc/mdc_request.c
• lustre/cmm/cmm_device.c
• lustre/include/lprocfs_status.h
• lustre/lov/lov_obd.c
• lustre/obdclass/obd_config.c
• lustre/obdclass/lprocfs_status.c
• lustre/fld/fld_request.c
• lustre/mdt/mdt_lproc.c

Integrated in lustre-master » i686,server,el6,inkernel #378
LU-106 procfs: many proc entries are not accessed safely (Revision ad538dba7f5259df31fc45e90835edc12756db40)

Result = SUCCESS
Oleg Drokin : ad538dba7f5259df31fc45e90835edc12756db40
Files :

• lustre/obdclass/obd_config.c
• lustre/mdt/mdt_handler.c
• lustre/mdt/mdt_lproc.c
• lustre/osd-ldiskfs/osd_lproc.c
• lustre/obdclass/lprocfs_status.c
• lustre/fld/fld_request.c
• lustre/lov/lov_obd.c
• lustre/include/lprocfs_status.h
• lustre/lmv/lmv_obd.c
• lustre/obdfilter/filter.c
• lustre/include/lustre_fld.h
• lustre/mgs/mgs_handler.c
• lustre/osc/osc_request.c
• lustre/mdc/mdc_request.c
• lustre/cmm/cmm_device.c

Integrated in lustre-master » i686,server,el5,ofa #378
LU-106 procfs: many proc entries are not accessed safely (Revision ad538dba7f5259df31fc45e90835edc12756db40)

Result = SUCCESS
Oleg Drokin : ad538dba7f5259df31fc45e90835edc12756db40
Files :

• lustre/cmm/cmm_device.c
• lustre/mgs/mgs_handler.c
• lustre/obdclass/obd_config.c
• lustre/include/lustre_fld.h
• lustre/lov/lov_obd.c
• lustre/mdt/mdt_lproc.c
• lustre/obdclass/lprocfs_status.c
• lustre/lmv/lmv_obd.c
• lustre/obdfilter/filter.c
• lustre/include/lprocfs_status.h
• lustre/osc/osc_request.c
• lustre/fld/fld_request.c
• lustre/osd-ldiskfs/osd_lproc.c
• lustre/mdc/mdc_request.c
• lustre/mdt/mdt_handler.c