Hi,
On Wed 26-02-14 13:48:35, Dan Carpenter wrote:
> [ False positive, this is not a bug, it's just that the second IRQ save
> is unnecessary ].
> FYI, there are new smatch warnings show up in
>
> tree: git://git.cmpxchg.org/linux-mmotm.git master
> head: a6a1126d3535f0bd8d7c56810061541a4f5595af
> commit: c8acdd9e9cf2dd5a5f62034dfaf93c721b3f405d [64/350] ocfs2: avoid blocking in ocfs2_mark_lockres_freeing() in downconvert thread
>
> fs/ocfs2/dlmglue.c:3184 ocfs2_mark_lockres_freeing() error: double lock 'irqsave:flags'
> fs/ocfs2/dlmglue.c:3204 ocfs2_mark_lockres_freeing() error: double unlock 'irqsave:flags'
Yeah. I actually have an updated version of a patch which doesn't have
these two nested irqsafe calls. I'll send it to Andrew. Thanks for letting
me know.
Honza
> git remote add mmotm git://git.cmpxchg.org/linux-mmotm.git
> git remote update mmotm
> git checkout c8acdd9e9cf2dd5a5f62034dfaf93c721b3f405d
> vim +3184 fs/ocfs2/dlmglue.c
>
> c8acdd9e Jan Kara 2014-02-25 3178 * ocfs2_clear_inode()
> c8acdd9e Jan Kara 2014-02-25 3179 * ocfs2_mark_lockres_freeing()
> c8acdd9e Jan Kara 2014-02-25 3180 * ... blocks waiting for OCFS2_LOCK_QUEUED
> c8acdd9e Jan Kara 2014-02-25 3181 * since we are the downconvert thread which
> c8acdd9e Jan Kara 2014-02-25 3182 * should clear the flag.
> c8acdd9e Jan Kara 2014-02-25 3183 */
> c8acdd9e Jan Kara 2014-02-25 @3184 spin_lock_irqsave(&osb->dc_task_lock, flags);
> c8acdd9e Jan Kara 2014-02-25 3185 list_del_init(&lockres->l_blocked_list);
> c8acdd9e Jan Kara 2014-02-25 3186 osb->blocked_lock_count--;
> c8acdd9e Jan Kara 2014-02-25 3187 spin_unlock_irqrestore(&osb->dc_task_lock, flags);
> c8acdd9e Jan Kara 2014-02-25 3188 lockres_clear_flags(lockres, OCFS2_LOCK_QUEUED);
> c8acdd9e Jan Kara 2014-02-25 3189 goto out_unlock;
> c8acdd9e Jan Kara 2014-02-25 3190 }
> ccd979bd Mark Fasheh 2005-12-15 3191 while (lockres->l_flags & OCFS2_LOCK_QUEUED) {
> ccd979bd Mark Fasheh 2005-12-15 3192 lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_QUEUED, 0);
> ccd979bd Mark Fasheh 2005-12-15 3193 spin_unlock_irqrestore(&lockres->l_lock, flags);
> ccd979bd Mark Fasheh 2005-12-15 3194
> ccd979bd Mark Fasheh 2005-12-15 3195 mlog(0, "Waiting on lockres %s\n", lockres->l_name);
> ccd979bd Mark Fasheh 2005-12-15 3196
> ccd979bd Mark Fasheh 2005-12-15 3197 status = ocfs2_wait_for_mask(&mw);
> ccd979bd Mark Fasheh 2005-12-15 3198 if (status)
> ccd979bd Mark Fasheh 2005-12-15 3199 mlog_errno(status);
> ccd979bd Mark Fasheh 2005-12-15 3200
> ccd979bd Mark Fasheh 2005-12-15 3201 spin_lock_irqsave(&lockres->l_lock, flags);
> ccd979bd Mark Fasheh 2005-12-15 3202 }
> c8acdd9e Jan Kara 2014-02-25 3203 out_unlock:
> ccd979bd Mark Fasheh 2005-12-15 @3204 spin_unlock_irqrestore(&lockres->l_lock, flags);
> ccd979bd Mark Fasheh 2005-12-15 3205 }
> ccd979bd Mark Fasheh 2005-12-15 3206
> d680efe9 Mark Fasheh 2006-09-08 3207 void ocfs2_simple_drop_lockres(struct ocfs2_super *osb,
>
> ---
> 0-DAY kernel build testing backend Open Source Technology Center
> http://lists.01.org/mailman/listinfo/kbuild Intel Corporation
--
Jan Kara <jack@xxxxxxx>
SUSE Labs, CR
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>