diff mbox series

scsi: ufs-qcom: Fix scheduling while atomic issue

Message ID 20200525204125.46171-1-jeffrey.l.hugo@gmail.com (mailing list archive)
State Mainlined
Commit 3be60b564de49875e47974c37fabced893cd0931
Headers show
Series scsi: ufs-qcom: Fix scheduling while atomic issue | expand

Commit Message

Jeffrey Hugo May 25, 2020, 8:41 p.m. UTC
ufs_qcom_dump_dbg_regs() uses usleep_range, a sleeping function, but can
be called from atomic context in the following flow:

ufshcd_intr -> ufshcd_sl_intr -> ufshcd_check_errors ->
ufshcd_print_host_regs -> ufshcd_vops_dbg_register_dump ->
ufs_qcom_dump_dbg_regs

This causes a boot crash on the Lenovo Miix 630 when the interrupt is
handled on the idle thread.

Fix the issue by switching to udelay().

Fixes: 9c46b8676271 ("scsi: ufs-qcom: dump additional testbus registers")
Signed-off-by: Jeffrey Hugo <jeffrey.l.hugo@gmail.com>
---
 drivers/scsi/ufs/ufs-qcom.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

Comments

Avri Altman May 26, 2020, 6:25 a.m. UTC | #1
> ufs_qcom_dump_dbg_regs() uses usleep_range, a sleeping function, but can
> be called from atomic context in the following flow:
> 
> ufshcd_intr -> ufshcd_sl_intr -> ufshcd_check_errors ->
> ufshcd_print_host_regs -> ufshcd_vops_dbg_register_dump ->
> ufs_qcom_dump_dbg_regs
> 
> This causes a boot crash on the Lenovo Miix 630 when the interrupt is
> handled on the idle thread.
> 
> Fix the issue by switching to udelay().
> 
> Fixes: 9c46b8676271 ("scsi: ufs-qcom: dump additional testbus registers")
> Signed-off-by: Jeffrey Hugo <jeffrey.l.hugo@gmail.com>
Reviewed-by: Avri Altman <avri.altman@wdc.com>
Bean Huo May 26, 2020, 9:41 a.m. UTC | #2
On Tue, 2020-05-26 at 06:25 +0000, Avri Altman wrote:
>  
> > ufs_qcom_dump_dbg_regs() uses usleep_range, a sleeping function,
> > but can
> > be called from atomic context in the following flow:
> > 
> > ufshcd_intr -> ufshcd_sl_intr -> ufshcd_check_errors ->
> > ufshcd_print_host_regs -> ufshcd_vops_dbg_register_dump ->
> > ufs_qcom_dump_dbg_regs
> > 
> > This causes a boot crash on the Lenovo Miix 630 when the interrupt
> > is
> > handled on the idle thread.
> > 
> > Fix the issue by switching to udelay().
> > 
> > Fixes: 9c46b8676271 ("scsi: ufs-qcom: dump additional testbus
> > registers")
> > Signed-off-by: Jeffrey Hugo <jeffrey.l.hugo@gmail.com>
> 
> Reviewed-by: Avri Altman <avri.altman@wdc.com>
Reviewed-by: Bean Huo <beanhuo@micron.com>
Martin K. Petersen May 27, 2020, 2:12 a.m. UTC | #3
On Mon, 25 May 2020 13:41:25 -0700, Jeffrey Hugo wrote:

> ufs_qcom_dump_dbg_regs() uses usleep_range, a sleeping function, but can
> be called from atomic context in the following flow:
> 
> ufshcd_intr -> ufshcd_sl_intr -> ufshcd_check_errors ->
> ufshcd_print_host_regs -> ufshcd_vops_dbg_register_dump ->
> ufs_qcom_dump_dbg_regs
> 
> [...]

Applied to 5.8/scsi-queue, thanks!

[1/1] scsi: ufs-qcom: Fix scheduling while atomic issue
      https://git.kernel.org/mkp/scsi/c/3be60b564de4
diff mbox series

Patch

diff --git a/drivers/scsi/ufs/ufs-qcom.c b/drivers/scsi/ufs/ufs-qcom.c
index 19aa5c44e0da..f938867301a0 100644
--- a/drivers/scsi/ufs/ufs-qcom.c
+++ b/drivers/scsi/ufs/ufs-qcom.c
@@ -1658,11 +1658,11 @@  static void ufs_qcom_dump_dbg_regs(struct ufs_hba *hba)
 
 	/* sleep a bit intermittently as we are dumping too much data */
 	ufs_qcom_print_hw_debug_reg_all(hba, NULL, ufs_qcom_dump_regs_wrapper);
-	usleep_range(1000, 1100);
+	udelay(1000);
 	ufs_qcom_testbus_read(hba);
-	usleep_range(1000, 1100);
+	udelay(1000);
 	ufs_qcom_print_unipro_testbus(hba);
-	usleep_range(1000, 1100);
+	udelay(1000);
 }
 
 /**