| Message ID | 20191218140552.12249-1-baijiaju1990@gmail.com |
|---|---|
| State | New, archived |
| Headers | show |
| Series | mtd: maps: pcmciamtd: fix possible sleep-in-atomic-context bugs in pcmciamtd_set_vpp() | expand |
On Wed, Dec 18, 2019 at 10:05:52PM +0800, Jia-Ju Bai wrote: > The driver may sleep while holding a spinlock. > The function call path (from bottom to top) in Linux 4.19 is: > > drivers/pcmcia/pcmcia_resource.c, 312: > mutex_lock in pcmcia_fixup_vpp > drivers/mtd/maps/pcmciamtd.c, 309: > pcmcia_fixup_vpp in pcmciamtd_set_vpp > drivers/mtd/maps/pcmciamtd.c, 306: > _raw_spin_lock_irqsave in pcmciamtd_set_vpp > > drivers/pcmcia/pcmcia_resource.c, 312: > mutex_lock in pcmcia_fixup_vpp > drivers/mtd/maps/pcmciamtd.c, 312: > pcmcia_fixup_vpp in pcmciamtd_set_vpp > drivers/mtd/maps/pcmciamtd.c, 306: > _raw_spin_lock_irqsave in pcmciamtd_set_vp > > mutex_lock() may sleep at runtime. Thanks for noticing this issue. > To fix these bugs, pcmcia_fixup_vpp() is called without holding the > spinlock. I don't think that this is the right approach here -- we lose the protection against races in calls to pcmcia_fixup_vpp(). Instead, we should change the spinlock to a mutex, which seems to be sufficient here. Could you prepare such a patch, please? Thanks, Dominik
On 2019/12/19 1:28, Dominik Brodowski wrote: > On Wed, Dec 18, 2019 at 10:05:52PM +0800, Jia-Ju Bai wrote: >> The driver may sleep while holding a spinlock. >> The function call path (from bottom to top) in Linux 4.19 is: >> >> drivers/pcmcia/pcmcia_resource.c, 312: >> mutex_lock in pcmcia_fixup_vpp >> drivers/mtd/maps/pcmciamtd.c, 309: >> pcmcia_fixup_vpp in pcmciamtd_set_vpp >> drivers/mtd/maps/pcmciamtd.c, 306: >> _raw_spin_lock_irqsave in pcmciamtd_set_vpp >> >> drivers/pcmcia/pcmcia_resource.c, 312: >> mutex_lock in pcmcia_fixup_vpp >> drivers/mtd/maps/pcmciamtd.c, 312: >> pcmcia_fixup_vpp in pcmciamtd_set_vpp >> drivers/mtd/maps/pcmciamtd.c, 306: >> _raw_spin_lock_irqsave in pcmciamtd_set_vp >> >> mutex_lock() may sleep at runtime. > Thanks for noticing this issue. > >> To fix these bugs, pcmcia_fixup_vpp() is called without holding the >> spinlock. > I don't think that this is the right approach here -- we lose the protection > against races in calls to pcmcia_fixup_vpp(). Instead, we should change the > spinlock to a mutex, which seems to be sufficient here. Could you prepare > such a patch, please? Okay, thanks for the advice :) I will send a new patch. Best wishes, Jia-Ju Bai
diff --git a/drivers/mtd/maps/pcmciamtd.c b/drivers/mtd/maps/pcmciamtd.c index 70bb403f69f7..d2cd1708aa49 100644 --- a/drivers/mtd/maps/pcmciamtd.c +++ b/drivers/mtd/maps/pcmciamtd.c @@ -301,17 +301,23 @@ static void pcmciamtd_set_vpp(struct map_info *map, int on) struct pcmciamtd_dev *dev = (struct pcmciamtd_dev *)map->map_priv_1; struct pcmcia_device *link = dev->p_dev; unsigned long flags; + int fixup_flag = 0; pr_debug("dev = %p on = %d vpp = %d\n\n", dev, on, dev->vpp); spin_lock_irqsave(&pcmcia_vpp_lock, flags); if (on) { if (++pcmcia_vpp_refcnt == 1) /* first nested 'on' */ - pcmcia_fixup_vpp(link, dev->vpp); + fixup_flag = 1; } else { if (--pcmcia_vpp_refcnt == 0) /* last nested 'off' */ - pcmcia_fixup_vpp(link, 0); + fixup_flag = 2; } spin_unlock_irqrestore(&pcmcia_vpp_lock, flags); + + if (fixup_flag == 1) + pcmcia_fixup_vpp(link, dev->vpp); + else if (fixup_flag == 2) + pcmcia_fixup_vpp(link, 0); }
The driver may sleep while holding a spinlock. The function call path (from bottom to top) in Linux 4.19 is: drivers/pcmcia/pcmcia_resource.c, 312: mutex_lock in pcmcia_fixup_vpp drivers/mtd/maps/pcmciamtd.c, 309: pcmcia_fixup_vpp in pcmciamtd_set_vpp drivers/mtd/maps/pcmciamtd.c, 306: _raw_spin_lock_irqsave in pcmciamtd_set_vpp drivers/pcmcia/pcmcia_resource.c, 312: mutex_lock in pcmcia_fixup_vpp drivers/mtd/maps/pcmciamtd.c, 312: pcmcia_fixup_vpp in pcmciamtd_set_vpp drivers/mtd/maps/pcmciamtd.c, 306: _raw_spin_lock_irqsave in pcmciamtd_set_vp mutex_lock() may sleep at runtime. To fix these bugs, pcmcia_fixup_vpp() is called without holding the spinlock. These bugs are found by a static analysis tool STCheck written by myself. Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com> --- drivers/mtd/maps/pcmciamtd.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-)