Re: [PATCH] infiniband: hw: hns: Fix a sleep-in-atomic-context bug in hns_roce_set_hem()

From: Jason Gunthorpe
Date: Wed Sep 05 2018 - 17:54:55 EST


On Sat, Sep 01, 2018 at 07:39:48PM +0800, Jia-Ju Bai wrote:
> The driver may sleep with holding a spinlock.
>
> The function call paths (from bottom to top) in Linux-4.16 are:
>
> drivers/infiniband/hw/hns/hns_roce_hem.c, 345:
> msleep in hns_roce_set_hem
> drivers/infiniband/hw/hns/hns_roce_hem.c, 330:
> _raw_spin_lock_irqsave in hns_roce_set_hem
>
> To fix this bug, msleep() is replaced with mdelay().
>
> This bug is found by my static analysis tool DSAC.
>
> Signed-off-by: Jia-Ju Bai <baijiaju1990@xxxxxxxxx>
> ---
> drivers/infiniband/hw/hns/hns_roce_hem.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)

This was already fixed here:

commit 73b4e1f4c04e76b19fca296d09b97555bfb35202
Author: Lijun Ou <oulijun@xxxxxxxxxx>
Date: Wed Jul 25 15:29:38 2018 +0800

RDMA/hns: Use delay instead of usleep

In order to avoid using usleep function in lock function, we use delay
function instead of it. Besides, it also use brackets for standardized
the computed order.

Signed-off-by: Lijun Ou <oulijun@xxxxxxxxxx>
Signed-off-by: Jason Gunthorpe <jgg@xxxxxxxxxxxx>