mbox series

[v2,0/2] Try to fix the sg_mitet bugs in SH MMCIF

Message ID 20240221-fix-sh-mmcif-v2-0-5e521eb25ae4@linaro.org
Headers show
Series Try to fix the sg_mitet bugs in SH MMCIF | expand

Message

Linus Walleij Feb. 21, 2024, 9:22 p.m. UTC 
This adresses some bugs found after the introduction of the
memory iterator sg_miter to the sh mmcif driver.

This was first just one patch for fixing the atomic bug, but
now also a second patch is needed to fix a semantic issue.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
Changes in v2:
- Collect Geerts test tag on patch 1
- Add a second patch fixing the problem with advancing to
  the next sglist entry before reading/writing the first block and
  after reading/writing each block in a multiblock operation.
- Link to v1: https://lore.kernel.org/r/20240220-fix-sh-mmcif-v1-1-b9d08a787c1f@linaro.org

---
Linus Walleij (2):
      mmc: sh_mmcif: sg_miter does not need to be atomic
      mmc: sh_mmcif: Advance sg_miter before reading blocks

 drivers/mmc/host/sh_mmcif.c | 46 +++++++++++++++++++++++++++++----------------
 1 file changed, 30 insertions(+), 16 deletions(-)
---
base-commit: 2d5c7b7eb345249cb34d42cbc2b97b4c57ea944e
change-id: 20240220-fix-sh-mmcif-49c1de70c5b7

Best regards,

Comments

Geert Uytterhoeven Feb. 22, 2024, 9:19 a.m. UTC | #1 
Hi Linus,

On Wed, Feb 21, 2024 at 10:23 PM Linus Walleij <linus.walleij@linaro.org> wrote:
> All the sglist iterations happen in the *threaded* interrupt handler
> and that context is not atomic, so don't request an atomic
> sglist miter. Using an atomic miter results in "BUG: scheduling while
> atomic" splats.
>
> Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
> Fixes: 27b57277d9ba ("mmc: sh_mmcif: Use sg_miter for PIO")
> Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
> ---
> Hi Geert, it'd be great if you could test this!

Done before (see the Tb above ;-)

You probably still want to s/does not need to/must not/?
(sorry, I miswrote that previously)

Gr{oetje,eeting}s,

                        Geert
Ulf Hansson Feb. 28, 2024, 1:03 p.m. UTC | #2 
On Wed, 21 Feb 2024 at 22:23, Linus Walleij <linus.walleij@linaro.org> wrote:
>
> This adresses some bugs found after the introduction of the
> memory iterator sg_miter to the sh mmcif driver.
>
> This was first just one patch for fixing the atomic bug, but
> now also a second patch is needed to fix a semantic issue.
>
> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
> ---
> Changes in v2:
> - Collect Geerts test tag on patch 1
> - Add a second patch fixing the problem with advancing to
>   the next sglist entry before reading/writing the first block and
>   after reading/writing each block in a multiblock operation.
> - Link to v1: https://lore.kernel.org/r/20240220-fix-sh-mmcif-v1-1-b9d08a787c1f@linaro.org
>
> ---
> Linus Walleij (2):
>       mmc: sh_mmcif: sg_miter does not need to be atomic
>       mmc: sh_mmcif: Advance sg_miter before reading blocks
>
>  drivers/mmc/host/sh_mmcif.c | 46 +++++++++++++++++++++++++++++----------------
>  1 file changed, 30 insertions(+), 16 deletions(-)
> ---

Applied for next and by amending the commit message header as pointed
out by Geert, thanks!

Kind regards
Uffe