diff mbox

Applied "spi: sh-msiof: Fix timeout failures for TX-only DMA transfers" to the spi tree

Message ID E1eWn56-0005rw-N5@debutante (mailing list archive)
State Not Applicable
Delegated to: Geert Uytterhoeven
Headers show

Commit Message

Mark Brown Jan. 3, 2018, 5:44 p.m. UTC
The patch

   spi: sh-msiof: Fix timeout failures for TX-only DMA transfers

has been applied to the spi tree at

   https://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi.git 

All being well this means that it will be integrated into the linux-next
tree (usually sometime in the next 24 hours) and sent to Linus during
the next merge window (or sooner if it is a bug fix), however if
problems are discovered then the patch may be dropped or reverted.  

You may get further e-mails resulting from automated or manual testing
and review of the tree, please engage with people reporting problems and
send followup patches addressing any issues that are reported if needed.

If any updates are required or you are submitting further changes they
should be sent as incremental updates against current git, existing
patches will not be replaced.

Please add any relevant lists and maintainers to the CCs when replying
to this mail.

Thanks,
Mark

From 89434c3c35081439627baa2225622d5bd12242fe Mon Sep 17 00:00:00 2001
From: Geert Uytterhoeven <geert+renesas@glider.be>
Date: Wed, 3 Jan 2018 18:11:14 +0100
Subject: [PATCH] spi: sh-msiof: Fix timeout failures for TX-only DMA transfers

When using RX (with or without TX), the DMA interrupt triggers
completion when the RX FIFO has been emptied, i.e. after the full
transfer has finished.

However, when using TX without RX, the DMA interrupt triggers completion
as soon as the DMA engine has filled the TX FIFO, i.e. before the full
transfer has finished.  Then sh_msiof_modify_ctr_wait() will spin until
the transfer has really finished and the TFSE bit is cleared, for at
most 1 ms.  For slow speeds and/or large transfers, this may cause
timeouts and transfer failures:

    spi_sh_msiof e6e10000.spi: failed to shut down hardware
    74x164 spi2.0: SPI transfer failed: -110
    spi_master spi2: failed to transfer one message from queue
    74x164 spi2.0: Failed writing: -110

Fix this by waiting explicitly until the TX FIFO has been emptied.

Based on a patch in the BSP by Hiromitsu Yamasaki.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Mark Brown <broonie@kernel.org>
---
 drivers/spi/spi-sh-msiof.c | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)

Comments

Dirk Behme Jan. 4, 2018, 7:05 a.m. UTC | #1
On 03.01.2018 18:44, Mark Brown wrote:
> The patch
> 
>     spi: sh-msiof: Fix timeout failures for TX-only DMA transfers
> 
> has been applied to the spi tree at
> 
>     https://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi.git


Would it make sense to forward this to -stable, too?


Best regards

Dirk


> All being well this means that it will be integrated into the linux-next
> tree (usually sometime in the next 24 hours) and sent to Linus during
> the next merge window (or sooner if it is a bug fix), however if
> problems are discovered then the patch may be dropped or reverted.
> 
> You may get further e-mails resulting from automated or manual testing
> and review of the tree, please engage with people reporting problems and
> send followup patches addressing any issues that are reported if needed.
> 
> If any updates are required or you are submitting further changes they
> should be sent as incremental updates against current git, existing
> patches will not be replaced.
> 
> Please add any relevant lists and maintainers to the CCs when replying
> to this mail.
> 
> Thanks,
> Mark
> 
>>From 89434c3c35081439627baa2225622d5bd12242fe Mon Sep 17 00:00:00 2001
> From: Geert Uytterhoeven <geert+renesas@glider.be>
> Date: Wed, 3 Jan 2018 18:11:14 +0100
> Subject: [PATCH] spi: sh-msiof: Fix timeout failures for TX-only DMA transfers
> 
> When using RX (with or without TX), the DMA interrupt triggers
> completion when the RX FIFO has been emptied, i.e. after the full
> transfer has finished.
> 
> However, when using TX without RX, the DMA interrupt triggers completion
> as soon as the DMA engine has filled the TX FIFO, i.e. before the full
> transfer has finished.  Then sh_msiof_modify_ctr_wait() will spin until
> the transfer has really finished and the TFSE bit is cleared, for at
> most 1 ms.  For slow speeds and/or large transfers, this may cause
> timeouts and transfer failures:
> 
>      spi_sh_msiof e6e10000.spi: failed to shut down hardware
>      74x164 spi2.0: SPI transfer failed: -110
>      spi_master spi2: failed to transfer one message from queue
>      74x164 spi2.0: Failed writing: -110
> 
> Fix this by waiting explicitly until the TX FIFO has been emptied.
> 
> Based on a patch in the BSP by Hiromitsu Yamasaki.
> 
> Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
> Signed-off-by: Mark Brown <broonie@kernel.org>
> ---
>   drivers/spi/spi-sh-msiof.c | 12 +++++++++++-
>   1 file changed, 11 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/spi/spi-sh-msiof.c b/drivers/spi/spi-sh-msiof.c
> index fcd261f98b9f..06bc4b170c47 100644
> --- a/drivers/spi/spi-sh-msiof.c
> +++ b/drivers/spi/spi-sh-msiof.c
> @@ -784,11 +784,21 @@ static int sh_msiof_dma_once(struct sh_msiof_spi_priv *p, const void *tx,
>   		goto stop_dma;
>   	}
>   
> -	/* wait for tx fifo to be emptied / rx fifo to be filled */
> +	/* wait for tx/rx DMA completion */
>   	ret = sh_msiof_wait_for_completion(p);
>   	if (ret)
>   		goto stop_reset;
>   
> +	if (!rx) {
> +		reinit_completion(&p->done);
> +		sh_msiof_write(p, IER, IER_TEOFE);
> +
> +		/* wait for tx fifo to be emptied */
> +		ret = sh_msiof_wait_for_completion(p);
> +		if (ret)
> +			goto stop_reset;
> +	}
> +
>   	/* clear status bits */
>   	sh_msiof_reset_str(p);
>
Geert Uytterhoeven Jan. 4, 2018, 8:26 a.m. UTC | #2
Hi Dirk,

On Thu, Jan 4, 2018 at 8:05 AM, Dirk Behme <dirk.behme@de.bosch.com> wrote:
> On 03.01.2018 18:44, Mark Brown wrote:
>> The patch
>>
>>     spi: sh-msiof: Fix timeout failures for TX-only DMA transfers
>>
>> has been applied to the spi tree at
>>
>>     https://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi.git
>
> Would it make sense to forward this to -stable, too?

Yes, _after_ it has landed in Linus' tree.
I didn't add a Fixes tag, as the fix needs changes to apply to any pre-4.13
kernel.

Most probably it will be picked up by stable automatically, as it has the
magic word "Fix" in the one-line summary.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds
Mark Brown Jan. 4, 2018, noon UTC | #3
On Thu, Jan 04, 2018 at 09:26:31AM +0100, Geert Uytterhoeven wrote:
> On Thu, Jan 4, 2018 at 8:05 AM, Dirk Behme <dirk.behme@de.bosch.com> wrote:
> > On 03.01.2018 18:44, Mark Brown wrote:

> > Would it make sense to forward this to -stable, too?

> Yes, _after_ it has landed in Linus' tree.
> I didn't add a Fixes tag, as the fix needs changes to apply to any pre-4.13
> kernel.

> Most probably it will be picked up by stable automatically, as it has the
> magic word "Fix" in the one-line summary.

I'm also always a bit worried about just sending things on when nobody's
been complaining about problems without lots of testing.
diff mbox

Patch

diff --git a/drivers/spi/spi-sh-msiof.c b/drivers/spi/spi-sh-msiof.c
index fcd261f98b9f..06bc4b170c47 100644
--- a/drivers/spi/spi-sh-msiof.c
+++ b/drivers/spi/spi-sh-msiof.c
@@ -784,11 +784,21 @@  static int sh_msiof_dma_once(struct sh_msiof_spi_priv *p, const void *tx,
 		goto stop_dma;
 	}
 
-	/* wait for tx fifo to be emptied / rx fifo to be filled */
+	/* wait for tx/rx DMA completion */
 	ret = sh_msiof_wait_for_completion(p);
 	if (ret)
 		goto stop_reset;
 
+	if (!rx) {
+		reinit_completion(&p->done);
+		sh_msiof_write(p, IER, IER_TEOFE);
+
+		/* wait for tx fifo to be emptied */
+		ret = sh_msiof_wait_for_completion(p);
+		if (ret)
+			goto stop_reset;
+	}
+
 	/* clear status bits */
 	sh_msiof_reset_str(p);