* [PATCH] pvcalls-front: Use GFP_ATOMIC under spin_lock
@ 2018-11-29 12:01 Wen Yang
2018-11-29 17:01 ` Juergen Gross
0 siblings, 1 reply; 2+ messages in thread
From: Wen Yang @ 2018-11-29 12:01 UTC (permalink / raw)
To: boris.ostrovsky, jgross, sstabellini
Cc: xen-devel, linux-kernel, zhong.weidong, Wen Yang, Julia Lawall
The problem is that we call this with a spin lock held.
The call tree is:
pvcalls_front_accept() holds bedata->socket_lock.
-> create_active()
-> __get_free_pages() uses GFP_KERNEL
The create_active() function is only called from pvcalls_front_accept()
with a spin_lock held, The allocation is not allowed to sleep and
GFP_KERNEL is not sufficient, it has to be ATOMIC.
This issue was detected by using the Coccinelle software.
Signed-off-by: Wen Yang <wen.yang99@zte.com.cn>
CC: Julia Lawall <julia.lawall@lip6.fr>
CC: Boris Ostrovsky <boris.ostrovsky@oracle.com>
CC: Juergen Gross <jgross@suse.com>
CC: Stefano Stabellini <sstabellini@kernel.org>
CC: xen-devel@lists.xenproject.org
CC: linux-kernel@vger.kernel.org
---
drivers/xen/pvcalls-front.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/xen/pvcalls-front.c b/drivers/xen/pvcalls-front.c
index 2f11ca72a281..f2bbc06a0f7f 100644
--- a/drivers/xen/pvcalls-front.c
+++ b/drivers/xen/pvcalls-front.c
@@ -344,11 +344,11 @@ static int create_active(struct sock_mapping *map, int *evtchn)
init_waitqueue_head(&map->active.inflight_conn_req);
map->active.ring = (struct pvcalls_data_intf *)
- __get_free_page(GFP_KERNEL | __GFP_ZERO);
+ __get_free_page(GFP_ATOMIC | __GFP_ZERO);
if (map->active.ring == NULL)
goto out_error;
map->active.ring->ring_order = PVCALLS_RING_ORDER;
- bytes = (void *)__get_free_pages(GFP_KERNEL | __GFP_ZERO,
+ bytes = (void *)__get_free_pages(GFP_ATOMIC | __GFP_ZERO,
PVCALLS_RING_ORDER);
if (bytes == NULL)
goto out_error;
--
2.19.1
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH] pvcalls-front: Use GFP_ATOMIC under spin_lock
2018-11-29 12:01 [PATCH] pvcalls-front: Use GFP_ATOMIC under spin_lock Wen Yang
@ 2018-11-29 17:01 ` Juergen Gross
0 siblings, 0 replies; 2+ messages in thread
From: Juergen Gross @ 2018-11-29 17:01 UTC (permalink / raw)
To: Wen Yang, boris.ostrovsky, sstabellini
Cc: xen-devel, linux-kernel, zhong.weidong, Julia Lawall
On 29/11/2018 13:01, Wen Yang wrote:
> The problem is that we call this with a spin lock held.
> The call tree is:
> pvcalls_front_accept() holds bedata->socket_lock.
> -> create_active()
> -> __get_free_pages() uses GFP_KERNEL
>
> The create_active() function is only called from pvcalls_front_accept()
> with a spin_lock held, The allocation is not allowed to sleep and
> GFP_KERNEL is not sufficient, it has to be ATOMIC.
I'd rather have a function doing the allocations which is called
outside the lock and either passing the allocated data to
create_active() or hook it into map in the allocation function.
Juergen
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2018-11-29 17:01 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-11-29 12:01 [PATCH] pvcalls-front: Use GFP_ATOMIC under spin_lock Wen Yang
2018-11-29 17:01 ` Juergen Gross
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).