[<prev] [next>] [day] [month] [year] [list]
Message-Id: <20210402085549.77050-1-bernard@vivo.com>
Date: Fri, 2 Apr 2021 01:55:47 -0700
From: Bernard Zhao <bernard@...o.com>
To: Ben Skeggs <bskeggs@...hat.com>, David Airlie <airlied@...ux.ie>,
Daniel Vetter <daniel@...ll.ch>, Lyude Paul <lyude@...hat.com>,
dri-devel@...ts.freedesktop.org, nouveau@...ts.freedesktop.org,
linux-kernel@...r.kernel.org
Cc: Bernard Zhao <bernard@...o.com>
Subject: [PATCH] drm/nouveau: fix potential abnormal lock/unlock
Fix coccicheck warning:
drivers/gpu/drm/nouveau/nvkm/engine/fifo/base.c:115:3-9: preceding lock on line 109
drivers/gpu/drm/nouveau/nvkm/engine/fifo/base.c:98:2-8: preceding lock on line 95
As we see, function nvkm_fifo_chan_inst & nvkm_fifo_chan_chid both
use spin_lock_irqsave, but no spin_unlock_irqrestore in if/return
branch, seems like a potential bug?
Signed-off-by: Bernard Zhao <bernard@...o.com>
---
drivers/gpu/drm/nouveau/nvkm/engine/fifo/base.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/base.c
index 2ed4ff05d207..e3f624d97644 100644
--- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/base.c
@@ -95,6 +95,7 @@ nvkm_fifo_chan_inst(struct nvkm_fifo *fifo, u64 inst, unsigned long *rflags)
spin_lock_irqsave(&fifo->lock, flags);
if ((chan = nvkm_fifo_chan_inst_locked(fifo, inst))) {
*rflags = flags;
+ spin_unlock_irqrestore(&fifo->lock, flags);
return chan;
}
spin_unlock_irqrestore(&fifo->lock, flags);
@@ -112,6 +113,7 @@ nvkm_fifo_chan_chid(struct nvkm_fifo *fifo, int chid, unsigned long *rflags)
list_del(&chan->head);
list_add(&chan->head, &fifo->chan);
*rflags = flags;
+ spin_unlock_irqrestore(&fifo->lock, flags);
return chan;
}
}
--
2.31.0
Powered by blists - more mailing lists