Skip to content

Commit

Permalink
rapidio/rio_cm: avoid GFP_KERNEL in atomic context
Browse files Browse the repository at this point in the history
As reported by Alexey Khoroshilov (https://lkml.org/lkml/2016/9/9/737):
riocm_send_close() is called from rio_cm_shutdown() under
spin_lock_bh(idr_lock), but riocm_send_close() uses a GFP_KERNEL
allocation.

Fix by taking riocm_send_close() outside of spinlock protected code.

[akpm@linux-foundation.org: remove unneeded `if (!list_empty())']
Link: http://lkml.kernel.org/r/20160915175402.10122-1-alexandre.bounine@idt.com
Signed-off-by: Alexandre Bounine <alexandre.bounine@idt.com>
Reported-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
  • Loading branch information
Alexandre Bounine authored and torvalds committed Sep 19, 2016
1 parent 63b52c4 commit b92ae13
Showing 1 changed file with 16 additions and 3 deletions.
19 changes: 16 additions & 3 deletions drivers/rapidio/rio_cm.c
Original file line number Diff line number Diff line change
Expand Up @@ -2247,17 +2247,30 @@ static int rio_cm_shutdown(struct notifier_block *nb, unsigned long code,
{
struct rio_channel *ch;
unsigned int i;
LIST_HEAD(list);

riocm_debug(EXIT, ".");

/*
* If there are any channels left in connected state send
* close notification to the connection partner.
* First build a list of channels that require a closing
* notification because function riocm_send_close() should
* be called outside of spinlock protected code.
*/
spin_lock_bh(&idr_lock);
idr_for_each_entry(&ch_idr, ch, i) {
riocm_debug(EXIT, "close ch %d", ch->id);
if (ch->state == RIO_CM_CONNECTED)
riocm_send_close(ch);
if (ch->state == RIO_CM_CONNECTED) {
riocm_debug(EXIT, "close ch %d", ch->id);
idr_remove(&ch_idr, ch->id);
list_add(&ch->ch_node, &list);
}
}
spin_unlock_bh(&idr_lock);

list_for_each_entry(ch, &list, ch_node)
riocm_send_close(ch);

return NOTIFY_DONE;
}

Expand Down

0 comments on commit b92ae13

Please sign in to comment.