Skip to contentSkip to conversation

mmc: host: mtk-sd: enable recheck_sdio_irq for MT8516 SoC

PATCH In review Fabien Parent 5 participants 6 messages Oct 23, 2020 – Oct 28, 2020

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 mmc: host: mtk-sd: enable recheck_sdio_irq for MT8516 SoC

MT8516 SoC suffers from sometimes losing SDIO IRQs, this makes SDIO devices sometimes unstable. Make use of the new property recheck_sdio_irq to fix the SDIO stability issues on MT8516.

---
 drivers/mmc/host/mtk-sd.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/mmc/host/mtk-sd.c +1 -1
index a704745e5882..3dc102eefe49 100644
@@ -524,7 +524,7 @@ static const struct mtk_mmc_compatible mt7622_compat = {
 
 static const struct mtk_mmc_compatible mt8516_compat = {
 	.clk_div_bits = 12,
-	.recheck_sdio_irq = false,
+	.recheck_sdio_irq = true,
 	.hs400_tune = false,
 	.pad_tune_reg = MSDC_PAD_TUNE0,
 	.async_fifo = true,
Signature · 1 line
2.28.0
maintainer Reviewed-by on v1 1/1
#1 lore raw
Quoted 5 lines · Fabien Parent

MT8516 SoC suffers from sometimes losing SDIO IRQs, this makes SDIO devices sometimes unstable. Make use of the new property recheck_sdio_irq to fix the SDIO stability issues on MT8516.

Quoted 18 lines
---
  drivers/mmc/host/mtk-sd.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
drivers/mmc/host/mtk-sd.c +1 -1
index a704745e5882..3dc102eefe49 100644
@@ -524,7 +524,7 @@ static const struct mtk_mmc_compatible mt7622_compat = {
 
  static const struct mtk_mmc_compatible mt8516_compat = {
  	.clk_div_bits = 12,
-	.recheck_sdio_irq = false,
+	.recheck_sdio_irq = true,
  	.hs400_tune = false,
  	.pad_tune_reg = MSDC_PAD_TUNE0,
  	.async_fifo = true,
#2 lore raw
In reply to Fabien Parent

+ Yong Mao, Chaotian Jing

Quoted 6 lines · Fabien Parent

MT8516 SoC suffers from sometimes losing SDIO IRQs, this makes SDIO devices sometimes unstable. Make use of the new property recheck_sdio_irq to fix the SDIO stability issues on MT8516.

Maybe this is a common problem, thus I am thinking that potentially we should enable the workaround for all variants?

I have looped in Yong Mao (who invented the workaround) and Chaotian Jing, to see if they can advise us how to proceed.

In any case, I think we should add a stable tag and a fixes tag.

Kind regards
Uffe

Quoted 20 lines
---
 drivers/mmc/host/mtk-sd.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/mmc/host/mtk-sd.c +1 -2
index a704745e5882..3dc102eefe49 100644
@@ -524,7 +524,7 @@ static const struct mtk_mmc_compatible mt7622_compat = {
 
 static const struct mtk_mmc_compatible mt8516_compat = {
        .clk_div_bits = 12,
-       .recheck_sdio_irq = false,
+       .recheck_sdio_irq = true,
        .hs400_tune = false,
        .pad_tune_reg = MSDC_PAD_TUNE0,
        .async_fifo = true,
--

2.28.0

#3 lore raw

Hi Ulf,

Quoted 4 lines · Ulf Hansson

+ Yong Mao, Chaotian Jing

Quoted 1 line · Fabien Parent
Quoted 8 lines
Quoted 5 lines

MT8516 SoC suffers from sometimes losing SDIO IRQs, this makes SDIO devices sometimes unstable. Make use of the new property recheck_sdio_irq to fix the SDIO stability issues on MT8516.

Maybe this is a common problem, thus I am thinking that potentially we should enable the workaround for all variants?

Not sure if this is of any help, but: we use the btmtksdio driver on a MT8183 soc with an Android kernel based on upstream.

With that kernel, we did not to apply this work-around in order to have a stable bluetooth experience (pairing with a remote controller)

However, on the MT8516 SoC, it's impossible for us to use btmtksdio without Fabien's fix.

Quoted 34 lines

I have looped in Yong Mao (who invented the workaround) and Chaotian Jing, to see if they can advise us how to proceed.

In any case, I think we should add a stable tag and a fixes tag.

Kind regards
Uffe

Quoted 18 lines
---
 drivers/mmc/host/mtk-sd.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/mmc/host/mtk-sd.c +1 -2
index a704745e5882..3dc102eefe49 100644
@@ -524,7 +524,7 @@ static const struct mtk_mmc_compatible mt7622_compat = {
 
 static const struct mtk_mmc_compatible mt8516_compat = {
        .clk_div_bits = 12,
-       .recheck_sdio_irq = false,
+       .recheck_sdio_irq = true,
        .hs400_tune = false,
        .pad_tune_reg = MSDC_PAD_TUNE0,
        .async_fifo = true,
--
Quoted 2 lines

2.28.0

_______________________________________________
Linux-mediatek mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-mediatek

#4 lore raw
Quoted 16 lines · Mattijs Korpershoek

Hi Ulf,

Quoted 4 lines · Ulf Hansson

+ Yong Mao, Chaotian Jing

Quoted 1 line · Fabien Parent
Quoted 8 lines
Quoted 5 lines

MT8516 SoC suffers from sometimes losing SDIO IRQs, this makes SDIO devices sometimes unstable. Make use of the new property recheck_sdio_irq to fix the SDIO stability issues on MT8516.

Maybe this is a common problem, thus I am thinking that potentially we should enable the workaround for all variants?

Quoted 9 lines

Not sure if this is of any help, but: we use the btmtksdio driver on a MT8183 soc with an Android kernel based on upstream.

With that kernel, we did not to apply this work-around in order to have a stable bluetooth experience (pairing with a remote controller)

However, on the MT8516 SoC, it's impossible for us to use btmtksdio without Fabien's fix.

Yes. For mt8516 SoC,recheck_sdio_irq should be set to true for avoiding SDIO dat1 irq lost issue. But for mt8183 SoC, it does not need recheck sdio irq mechanism.

Quoted 8 lines
Quoted 8 lines

I have looped in Yong Mao (who invented the workaround) and Chaotian Jing, to see if they can advise us how to proceed.

In any case, I think we should add a stable tag and a fixes tag.

Kind regards
Uffe

Hi Ulf,
Sorry. On the patch "mmc:mediatek:fix SDIO irq issue", I only consider fixing the issue on mt8173 SoC.But for the whole MTK upstream SoC, mt8183, mt2712, mt6779 and mt8192 does not need this mechanism, but the other upstream Soc such as mt8135, mt8173, mt2701, mt7622, mt8516 and mt7620 need this recheck mechanism. And all future SoC of our company does not need this mechanism.
If it's convenient for you, please help me to revise it.
Or I will submit a new patch to correct the value of setting. Thanks.

Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
---
 drivers/mmc/host/mtk-sd.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/mmc/host/mtk-sd.c +0 -0
index a704745e5882..3dc102eefe49 100644
@@ -524,7 +524,7 @@ static const struct mtk_mmc_compatible mt7622_compat = {
 
Quoted 15 lines
Quoted 11 lines
Quoted 10 lines
drivers/mmc/host/mtk-sd.c +1 -2
 static const struct mtk_mmc_compatible mt8516_compat = {
        .clk_div_bits = 12,
-       .recheck_sdio_irq = false,
+       .recheck_sdio_irq = true,
        .hs400_tune = false,
        .pad_tune_reg = MSDC_PAD_TUNE0,
        .async_fifo = true,
--

2.28.0

Quoted 4 lines

_______________________________________________
Linux-mediatek mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-mediatek

_______________________________________________
Linux-mediatek mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-mediatek

#5 lore raw
Quoted 24 lines · On Wed, 28 Oct 2020 at 08:22, [email protected] <[email protected]>…
Quoted 22 lines · Mattijs Korpershoek

Hi Ulf,

Quoted 4 lines · Ulf Hansson

+ Yong Mao, Chaotian Jing

Quoted 1 line · Fabien Parent
Quoted 8 lines
Quoted 5 lines

MT8516 SoC suffers from sometimes losing SDIO IRQs, this makes SDIO devices sometimes unstable. Make use of the new property recheck_sdio_irq to fix the SDIO stability issues on MT8516.

Maybe this is a common problem, thus I am thinking that potentially we should enable the workaround for all variants?

Not sure if this is of any help, but: we use the btmtksdio driver on a MT8183 soc with an Android kernel based on upstream.

With that kernel, we did not to apply this work-around in order to have a stable bluetooth experience (pairing with a remote controller)

Quoted 6 lines
Quoted 3 lines

However, on the MT8516 SoC, it's impossible for us to use btmtksdio without Fabien's fix.

Yes. For mt8516 SoC,recheck_sdio_irq should be set to true for avoiding SDIO dat1 irq lost issue. But for mt8183 SoC, it does not need recheck sdio irq mechanism.

Okay, thanks for confirming.

Quoted 10 lines
Quoted 9 lines
Quoted 9 lines

I have looped in Yong Mao (who invented the workaround) and Chaotian Jing, to see if they can advise us how to proceed.

In any case, I think we should add a stable tag and a fixes tag.

Kind regards
Uffe

Quoted 10 lines

Hi Ulf,
Sorry. On the patch "mmc:mediatek:fix SDIO irq issue", I only consider fixing the issue on mt8173 SoC.But for the whole MTK upstream SoC, mt8183, mt2712, mt6779 and mt8192 does not need this mechanism, but the other upstream Soc such as mt8135, mt8173, mt2701, mt7622, mt8516 and mt7620 need this recheck mechanism. And all future SoC of our company does not need this mechanism.
If it's convenient for you, please help me to revise it.
Or I will submit a new patch to correct the value of setting. Thanks.

Please submit a new patch to enable the "recheck_sdio_irq" flag for those variants that need it.

I would also add a fixes+stable tag and add reported-by tags from Fabien and Mattijs.

Kind regards
Uffe

Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
---
 drivers/mmc/host/mtk-sd.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/mmc/host/mtk-sd.c +0 -0
index a704745e5882..3dc102eefe49 100644
@@ -524,7 +524,7 @@ static const struct mtk_mmc_compatible mt7622_compat = {
 
Quoted 16 lines
Quoted 11 lines
Quoted 11 lines
Quoted 10 lines
drivers/mmc/host/mtk-sd.c +1 -2
 static const struct mtk_mmc_compatible mt8516_compat = {
        .clk_div_bits = 12,
-       .recheck_sdio_irq = false,
+       .recheck_sdio_irq = true,
        .hs400_tune = false,
        .pad_tune_reg = MSDC_PAD_TUNE0,
        .async_fifo = true,
--

2.28.0

Quoted 4 lines
Quoted 4 lines

_______________________________________________
Linux-mediatek mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-mediatek

_______________________________________________
Linux-mediatek mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-mediatek