[PATCH] mmc: block: Support Host to control FUA.

[PATCH] mmc: block: Support Host to control FUA.

[Wenchao Chen]

From: Wenchao Chen <wenchao.chen@unisoc.com>

This patch introduces host->fua_disable for MMC host controller.
The host can turn off FUA to improve performance.

1. fua_disable = 1
/sys/block/mmcblk0/queue # cat fua 0
I tested 5 times for each case and output a average speed.

1) Sequential read:
Speed: 266.8MiB/s, 265.1MiB/s, 262.9MiB/s, 268.7MiB/s, 265.2MiB/s
Average speed: 265.74MiB/s

2) Random read:
Speed: 98.75MiB/s, 98.7MiB/s, 98.5MiB/s, 99.4MiB/s, 98.7MiB/s
Average speed: 98.81MiB/s

3) Sequential write:
Speed: 199.94MiB/s, 199.1MiB/s, 205.5MiB/s, 206.5MiB/s, 191.5MiB/s
Average speed: 200.5MiB/s

4) Random write:
Speed: 68.6MiB/s, 71.8MiB/s, 77.1MiB/s, 64.8MiB/s, 69.3MiB/s
Average speed: 70.32MiB/s

2. fua_disable = 0 (default 0)
/sys/block/mmcblk0/queue # cat fua 1
I tested 5 times for each case and output a average speed.

1) Sequential read:
Speed: 259.3MiB/s, 258.8MiB/s, 258.2MiB/s, 259.5MiB/s, 253.5MiB/s
Average speed: 257.86MiB/s

2) Random read:
Speed: 98.9MiB/s, 101MiB/s, 101MiB/s, 99MiB/s, 101.1MiB/s
Average speed: 100.2MiB/s

3) Sequential write:
Speed: 153.7MiB/s, 146.2MiB/s, 151.2MiB/s, 148.8MiB/s, 147.5MiB/s
Average speed: 149.48MiB/s

4) Random write:
Speed: 12.9MiB/s, 12.3MiB/s, 12.6MiB/s, 12.8MiB/s, 12.8MiB/s
Average speed: 12.68MiB/s

According to the above data, disable FUA (fua_disable = 1) improves the
performance. Therefore, it is recommended to support the host to control
FUA.

Signed-off-by: Wenchao Chen <wenchao.chen@unisoc.com>
---
 drivers/mmc/core/block.c | 3 ++-
 include/linux/mmc/host.h | 3 +++
 2 files changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/mmc/core/block.c b/drivers/mmc/core/block.c
index 54cd009aee50..333e819e077a 100644
--- a/drivers/mmc/core/block.c
+++ b/drivers/mmc/core/block.c
@@ -2490,7 +2490,8 @@ static struct mmc_blk_data *mmc_blk_alloc_req(struct mmc_card *card,
 	    ((card->ext_csd.rel_param & EXT_CSD_WR_REL_PARAM_EN) ||
 	     card->ext_csd.rel_sectors)) {
 		md->flags |= MMC_BLK_REL_WR;
-		fua_enabled = true;
+		if (!card->host->fua_disable)
+			fua_enabled = true;
 		cache_enabled = true;
 	}
 	if (mmc_cache_enabled(card->host))
diff --git a/include/linux/mmc/host.h b/include/linux/mmc/host.h
index 8fdd3cf971a3..16a5bee3eeae 100644
--- a/include/linux/mmc/host.h
+++ b/include/linux/mmc/host.h
@@ -517,6 +517,9 @@ struct mmc_host {
 	struct blk_crypto_profile crypto_profile;
 #endif
 
+	/* Host FUA support */
+	bool			fua_disable;
+
 	/* Host Software Queue support */
 	bool			hsq_enabled;
 
-- 
2.17.1

Re: [PATCH] mmc: block: Support Host to control FUA.

[Chaitanya Kulkarni]

On 10/19/22 19:45, Wenchao Chen wrote:
> From: Wenchao Chen <wenchao.chen@unisoc.com>
> 
> This patch introduces host->fua_disable for MMC host controller.
> The host can turn off FUA to improve performance.
> 
> 1. fua_disable = 1
> /sys/block/mmcblk0/queue # cat fua 0
> I tested 5 times for each case and output a average speed.
> 
> 1) Sequential read:
> Speed: 266.8MiB/s, 265.1MiB/s, 262.9MiB/s, 268.7MiB/s, 265.2MiB/s
> Average speed: 265.74MiB/s
> 
> 2) Random read:
> Speed: 98.75MiB/s, 98.7MiB/s, 98.5MiB/s, 99.4MiB/s, 98.7MiB/s
> Average speed: 98.81MiB/s
> 
> 3) Sequential write:
> Speed: 199.94MiB/s, 199.1MiB/s, 205.5MiB/s, 206.5MiB/s, 191.5MiB/s
> Average speed: 200.5MiB/s
> 
> 4) Random write:
> Speed: 68.6MiB/s, 71.8MiB/s, 77.1MiB/s, 64.8MiB/s, 69.3MiB/s
> Average speed: 70.32MiB/s
> 
> 2. fua_disable = 0 (default 0)
> /sys/block/mmcblk0/queue # cat fua 1
> I tested 5 times for each case and output a average speed.
> 
> 1) Sequential read:
> Speed: 259.3MiB/s, 258.8MiB/s, 258.2MiB/s, 259.5MiB/s, 253.5MiB/s
> Average speed: 257.86MiB/s
> 
> 2) Random read:
> Speed: 98.9MiB/s, 101MiB/s, 101MiB/s, 99MiB/s, 101.1MiB/s
> Average speed: 100.2MiB/s
> 
> 3) Sequential write:
> Speed: 153.7MiB/s, 146.2MiB/s, 151.2MiB/s, 148.8MiB/s, 147.5MiB/s
> Average speed: 149.48MiB/s
> 
> 4) Random write:
> Speed: 12.9MiB/s, 12.3MiB/s, 12.6MiB/s, 12.8MiB/s, 12.8MiB/s
> Average speed: 12.68MiB/s
> 
> According to the above data, disable FUA (fua_disable = 1) improves the
> performance. Therefore, it is recommended to support the host to control
> FUA.
> 
> Signed-off-by: Wenchao Chen <wenchao.chen@unisoc.com>

Based on the data provided seems it does increase the performance.
Perhaps it is worth mentioning the reason behind the performance
increase and other side effects of doing so ?

Also, it is usually helpful to mention performance increase in the
% in the commit log so you can move verbose log (like above) into
the cover-letter for documentation purpose, this is based on
the feedback I got, unless maintainer(s) prefers is this way :).

-ck

RE: [PATCH] mmc: block: Support Host to control FUA.

[Avri Altman]

> According to the above data, disable FUA (fua_disable = 1) improves the
> performance. Therefore, it is recommended to support the host to control
> FUA.
> 
> Signed-off-by: Wenchao Chen <wenchao.chen@unisoc.com>
> ---
>  drivers/mmc/core/block.c | 3 ++-
>  include/linux/mmc/host.h | 3 +++
>  2 files changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/mmc/core/block.c b/drivers/mmc/core/block.c index
> 54cd009aee50..333e819e077a 100644
> --- a/drivers/mmc/core/block.c
> +++ b/drivers/mmc/core/block.c
> @@ -2490,7 +2490,8 @@ static struct mmc_blk_data
> *mmc_blk_alloc_req(struct mmc_card *card,
>             ((card->ext_csd.rel_param & EXT_CSD_WR_REL_PARAM_EN) ||
>              card->ext_csd.rel_sectors)) {
>                 md->flags |= MMC_BLK_REL_WR;
> -               fua_enabled = true;
> +               if (!card->host->fua_disable)
> +                       fua_enabled = true;
Where are you allowing to set fua_disable?
And then I would expect you to call blk_queue_flag_set to set QUEUE_FLAG_FUA.

Thanks,
Avri

>                 cache_enabled = true;
>         }
>         if (mmc_cache_enabled(card->host)) diff --git
> a/include/linux/mmc/host.h b/include/linux/mmc/host.h index
> 8fdd3cf971a3..16a5bee3eeae 100644
> --- a/include/linux/mmc/host.h
> +++ b/include/linux/mmc/host.h
> @@ -517,6 +517,9 @@ struct mmc_host {
>         struct blk_crypto_profile crypto_profile;  #endif
> 
> +       /* Host FUA support */
> +       bool                    fua_disable;
> +
>         /* Host Software Queue support */
>         bool                    hsq_enabled;
> 
> --
> 2.17.1

Re: [PATCH] mmc: block: Support Host to control FUA.

[Wenchao Chen]

On Thu, Oct 20, 2022 at 3:44 PM Avri Altman <Avri.Altman@wdc.com> wrote:
>
> > According to the above data, disable FUA (fua_disable = 1) improves the
> > performance. Therefore, it is recommended to support the host to control
> > FUA.
> >
> > Signed-off-by: Wenchao Chen <wenchao.chen@unisoc.com>
> > ---
> >  drivers/mmc/core/block.c | 3 ++-
> >  include/linux/mmc/host.h | 3 +++
> >  2 files changed, 5 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/mmc/core/block.c b/drivers/mmc/core/block.c index
> > 54cd009aee50..333e819e077a 100644
> > --- a/drivers/mmc/core/block.c
> > +++ b/drivers/mmc/core/block.c
> > @@ -2490,7 +2490,8 @@ static struct mmc_blk_data
> > *mmc_blk_alloc_req(struct mmc_card *card,
> >             ((card->ext_csd.rel_param & EXT_CSD_WR_REL_PARAM_EN) ||
> >              card->ext_csd.rel_sectors)) {
> >                 md->flags |= MMC_BLK_REL_WR;
> > -               fua_enabled = true;
> > +               if (!card->host->fua_disable)
> > +                       fua_enabled = true;
> Where are you allowing to set fua_disable?
> And then I would expect you to call blk_queue_flag_set to set QUEUE_FLAG_FUA.
>
> Thanks,
> Avri
Hi Avri
We will implement sdhci_sprd_probe in sdhci-sprd.c host->fua_disable = true;
Because sdhci_sprd_probe is before mmc_blk_alloc_req, calling blk_queue_flag_set
directly will also be updated.

Thanks,
Wenchao
>
> >                 cache_enabled = true;
> >         }
> >         if (mmc_cache_enabled(card->host)) diff --git
> > a/include/linux/mmc/host.h b/include/linux/mmc/host.h index
> > 8fdd3cf971a3..16a5bee3eeae 100644
> > --- a/include/linux/mmc/host.h
> > +++ b/include/linux/mmc/host.h
> > @@ -517,6 +517,9 @@ struct mmc_host {
> >         struct blk_crypto_profile crypto_profile;  #endif
> >
> > +       /* Host FUA support */
> > +       bool                    fua_disable;
> > +
> >         /* Host Software Queue support */
> >         bool                    hsq_enabled;
> >
> > --
> > 2.17.1
>

Re: [PATCH] mmc: block: Support Host to control FUA.

[Wenchao Chen]

On Thu, Oct 20, 2022 at 11:07 AM Chaitanya Kulkarni
<chaitanyak@nvidia.com> wrote:
>
> On 10/19/22 19:45, Wenchao Chen wrote:
> > From: Wenchao Chen <wenchao.chen@unisoc.com>
> >
> > This patch introduces host->fua_disable for MMC host controller.
> > The host can turn off FUA to improve performance.
> >
> > 1. fua_disable = 1
> > /sys/block/mmcblk0/queue # cat fua 0
> > I tested 5 times for each case and output a average speed.
> >
> > 1) Sequential read:
> > Speed: 266.8MiB/s, 265.1MiB/s, 262.9MiB/s, 268.7MiB/s, 265.2MiB/s
> > Average speed: 265.74MiB/s
> >
> > 2) Random read:
> > Speed: 98.75MiB/s, 98.7MiB/s, 98.5MiB/s, 99.4MiB/s, 98.7MiB/s
> > Average speed: 98.81MiB/s
> >
> > 3) Sequential write:
> > Speed: 199.94MiB/s, 199.1MiB/s, 205.5MiB/s, 206.5MiB/s, 191.5MiB/s
> > Average speed: 200.5MiB/s
> >
> > 4) Random write:
> > Speed: 68.6MiB/s, 71.8MiB/s, 77.1MiB/s, 64.8MiB/s, 69.3MiB/s
> > Average speed: 70.32MiB/s
> >
> > 2. fua_disable = 0 (default 0)
> > /sys/block/mmcblk0/queue # cat fua 1
> > I tested 5 times for each case and output a average speed.
> >
> > 1) Sequential read:
> > Speed: 259.3MiB/s, 258.8MiB/s, 258.2MiB/s, 259.5MiB/s, 253.5MiB/s
> > Average speed: 257.86MiB/s
> >
> > 2) Random read:
> > Speed: 98.9MiB/s, 101MiB/s, 101MiB/s, 99MiB/s, 101.1MiB/s
> > Average speed: 100.2MiB/s
> >
> > 3) Sequential write:
> > Speed: 153.7MiB/s, 146.2MiB/s, 151.2MiB/s, 148.8MiB/s, 147.5MiB/s
> > Average speed: 149.48MiB/s
> >
> > 4) Random write:
> > Speed: 12.9MiB/s, 12.3MiB/s, 12.6MiB/s, 12.8MiB/s, 12.8MiB/s
> > Average speed: 12.68MiB/s
> >
> > According to the above data, disable FUA (fua_disable = 1) improves the
> > performance. Therefore, it is recommended to support the host to control
> > FUA.
> >
> > Signed-off-by: Wenchao Chen <wenchao.chen@unisoc.com>
>
> Based on the data provided seems it does increase the performance.
> Perhaps it is worth mentioning the reason behind the performance
> increase and other side effects of doing so ?

We suspect it will affect the progressive power-down test, but after a week
of testing, turning off FUA did not affect this test.

>
> Also, it is usually helpful to mention performance increase in the
> % in the commit log so you can move verbose log (like above) into
> the cover-letter for documentation purpose, this is based on
> the feedback I got, unless maintainer(s) prefers is this way :).
>
> -ck
>

Thank you very much for your suggestion, I will update it in the next version.