On 09/29/2015 03:48 AM, Hannes Reinecke wrote:
+static void alua_rtpg_work(struct work_struct *work)
+{
+ struct alua_port_group *pg =
+ container_of(work, struct alua_port_group, rtpg_work.work);
+ struct scsi_device *sdev;
+ LIST_HEAD(qdata_list);
+ int err = SCSI_DH_OK;
+ struct alua_queue_data *qdata, *tmp;
+ unsigned long flags;
+
+ spin_lock_irqsave(&pg->lock, flags);
+ sdev = pg->rtpg_sdev;
+ if (!sdev) {
+ WARN_ON(pg->flags & ALUA_PG_RUN_RTPG ||
+ pg->flags & ALUA_PG_RUN_STPG);
+ spin_unlock_irqrestore(&pg->lock, flags);
+ return;
+ }
+ pg->flags |= ALUA_PG_RUNNING;
+ if (pg->flags & ALUA_PG_RUN_RTPG) {
+ spin_unlock_irqrestore(&pg->lock, flags);
+ err = alua_rtpg(sdev, pg);
+ spin_lock_irqsave(&pg->lock, flags);
+ if (err == SCSI_DH_RETRY) {
+ pg->flags &= ~ALUA_PG_RUNNING;
+ spin_unlock_irqrestore(&pg->lock, flags);
+ queue_delayed_work(kaluad_wq, &pg->rtpg_work,
+ pg->interval * HZ);
+ return;
+ }
+ pg->flags &= ~ALUA_PG_RUN_RTPG;
+ if (err != SCSI_DH_OK)
+ pg->flags &= ~ALUA_PG_RUN_STPG;
+ }
+ if (pg->flags & ALUA_PG_RUN_STPG) {
+ spin_unlock_irqrestore(&pg->lock, flags);
+ err = alua_stpg(sdev, pg);
+ spin_lock_irqsave(&pg->lock, flags);
+ pg->flags &= ~ALUA_PG_RUN_STPG;
+ if (err == SCSI_DH_RETRY) {
+ pg->flags |= ALUA_PG_RUN_RTPG;
+ pg->interval = 0;
+ pg->flags &= ~ALUA_PG_RUNNING;
+ spin_unlock_irqrestore(&pg->lock, flags);
+ queue_delayed_work(kaluad_wq, &pg->rtpg_work,
+ pg->interval * HZ);
+ return;
+ }
+ }
+
+ list_splice_init(&pg->rtpg_list, &qdata_list);
+ pg->rtpg_sdev = NULL;
+ spin_unlock_irqrestore(&pg->lock, flags);
+
+ list_for_each_entry_safe(qdata, tmp, &qdata_list, entry) {
+ list_del(&qdata->entry);
+ if (qdata->callback_fn)
+ qdata->callback_fn(qdata->callback_data, err);
+ kfree(qdata);
+ }
+ spin_lock_irqsave(&pg->lock, flags);
+ pg->flags &= ~ALUA_PG_RUNNING;
+ spin_unlock_irqrestore(&pg->lock, flags);
+ scsi_device_put(sdev);
+ kref_put(&pg->kref, release_port_group);
+}
With this patch series applied kmemleak reports several leaks that were
not reported without this patch. Is scsi_device_put() + kref_put()
always called before this function returns without requeueing the work
item ? Shouldn't the return value of queue_delayed_work() be checked ?
The leaks reported by kmemleak are:
unreferenced object 0xffff880423d31728 (size 128):
comm "kworker/2:3", pid 3589, jiffies 4294946634 (age 501.720s)
hex dump (first 32 bytes):
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
backtrace:
[<ffffffff814ed75e>] kmemleak_alloc+0x4e/0xb0
[<ffffffff811a9128>] kmem_cache_alloc_trace+0xc8/0x210
[<ffffffffa0269d78>] alua_rtpg_work+0x1b8/0xc10 [scsi_dh_alua]
[<ffffffff8108d9a8>] process_one_work+0x1d8/0x610
[<ffffffff8108def4>] worker_thread+0x114/0x460
[<ffffffff810941e8>] kthread+0xf8/0x110
[<ffffffff814f5c2f>] ret_from_fork+0x3f/0x70
[<ffffffffffffffff>] 0xffffffffffffffff
unreferenced object 0xffff88042ba00150 (size 8):
comm "srp_daemon", pid 608, jiffies 4294946454 (age 503.510s)
hex dump (first 8 bytes):
68 6f 73 74 31 35 00 a5 host15..
backtrace:
[<ffffffff814ed75e>] kmemleak_alloc+0x4e/0xb0
[<ffffffff811abd53>] __kmalloc_track_caller+0xe3/0x240
[<ffffffff81295ce2>] kvasprintf+0x52/0x80
[<ffffffff8128adbf>] kobject_set_name_vargs+0x1f/0x60
[<ffffffff81382217>] dev_set_name+0x47/0x50
[<ffffffffa000dcaa>] scsi_host_alloc+0x32a/0x4b0 [scsi_mod]
[<ffffffffa0279524>] srp_create_target+0x54/0x1410 [ib_srp]
[<ffffffff81381bf8>] dev_attr_store+0x18/0x30
[<ffffffff812300f4>] sysfs_kf_write+0x44/0x60
[<ffffffff8122f724>] kernfs_fop_write+0x144/0x190
[<ffffffff811b8788>] __vfs_write+0x28/0xe0
[<ffffffff811b8e19>] vfs_write+0xa9/0x190
[<ffffffff811b9b19>] SyS_write+0x49/0xa0
[<ffffffff814f5876>] entry_SYSCALL_64_fastpath+0x16/0x7a
[<ffffffffffffffff>] 0xffffffffffffffff
Bart.
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
