diff mbox series

mmc: core: Use GFP_NOIO in ACMD22

Message ID 20241014114458.360538-1-avri.altman@wdc.com
State New
Headers show
Series mmc: core: Use GFP_NOIO in ACMD22 | expand

Commit Message

Avri Altman Oct. 14, 2024, 11:44 a.m. UTC
While reviewing the SDUC series, Adrian made a comment concerning the
memory allocation code in mmc_sd_num_wr_blocks() - see [1].
Prevent memory allocations from triggering I/O operations while ACMD22
is in progress.

[1] https://www.spinics.net/lists/linux-mmc/msg82199.html

Suggested-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Avri Altman <avri.altman@wdc.com>
Cc: stable@vger.kernel.org
---
 drivers/mmc/core/block.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

Comments

Adrian Hunter Oct. 15, 2024, 9:44 a.m. UTC | #1
On 14/10/24 14:44, Avri Altman wrote:
> While reviewing the SDUC series, Adrian made a comment concerning the
> memory allocation code in mmc_sd_num_wr_blocks() - see [1].
> Prevent memory allocations from triggering I/O operations while ACMD22
> is in progress.
> 
> [1] https://www.spinics.net/lists/linux-mmc/msg82199.html
> 
> Suggested-by: Adrian Hunter <adrian.hunter@intel.com>
> Signed-off-by: Avri Altman <avri.altman@wdc.com>
> Cc: stable@vger.kernel.org
> ---
>  drivers/mmc/core/block.c | 10 +++++++++-
>  1 file changed, 9 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/mmc/core/block.c b/drivers/mmc/core/block.c
> index 04f3165cf9ae..042b0147d47e 100644
> --- a/drivers/mmc/core/block.c
> +++ b/drivers/mmc/core/block.c
> @@ -995,6 +995,8 @@ static int mmc_sd_num_wr_blocks(struct mmc_card *card, u32 *written_blocks)
>  	u32 result;
>  	__be32 *blocks;
>  	u8 resp_sz = mmc_card_ult_capacity(card) ? 8 : 4;
> +	unsigned int noio_flag;
> +
>  	struct mmc_request mrq = {};
>  	struct mmc_command cmd = {};
>  	struct mmc_data data = {};
> @@ -1018,9 +1020,13 @@ static int mmc_sd_num_wr_blocks(struct mmc_card *card, u32 *written_blocks)
>  	mrq.cmd = &cmd;
>  	mrq.data = &data;
>  
> +	noio_flag = memalloc_noio_save();
> +
>  	blocks = kmalloc(resp_sz, GFP_KERNEL);

Could have memalloc_noio_restore() here:

	memalloc_noio_restore(noio_flag);

but I feel maybe adding something like:

	u64 __aligned(8)	tiny_io_buf;

to either struct mmc_card or struct mmc_host is better?
Ulf, any thoughts?

> -	if (!blocks)
> +	if (!blocks) {
> +		memalloc_noio_restore(noio_flag);
>  		return -ENOMEM;
> +	}
>  
>  	sg_init_one(&sg, blocks, resp_sz);
>  
> @@ -1041,6 +1047,8 @@ static int mmc_sd_num_wr_blocks(struct mmc_card *card, u32 *written_blocks)
>  	}
>  	kfree(blocks);
>  
> +	memalloc_noio_restore(noio_flag);
> +
>  	if (cmd.error || data.error)
>  		return -EIO;
>
Ulf Hansson Oct. 16, 2024, 2:44 p.m. UTC | #2
On Tue, 15 Oct 2024 at 11:44, Adrian Hunter <adrian.hunter@intel.com> wrote:
>
> On 14/10/24 14:44, Avri Altman wrote:
> > While reviewing the SDUC series, Adrian made a comment concerning the
> > memory allocation code in mmc_sd_num_wr_blocks() - see [1].
> > Prevent memory allocations from triggering I/O operations while ACMD22
> > is in progress.
> >
> > [1] https://www.spinics.net/lists/linux-mmc/msg82199.html
> >
> > Suggested-by: Adrian Hunter <adrian.hunter@intel.com>
> > Signed-off-by: Avri Altman <avri.altman@wdc.com>
> > Cc: stable@vger.kernel.org
> > ---
> >  drivers/mmc/core/block.c | 10 +++++++++-
> >  1 file changed, 9 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/mmc/core/block.c b/drivers/mmc/core/block.c
> > index 04f3165cf9ae..042b0147d47e 100644
> > --- a/drivers/mmc/core/block.c
> > +++ b/drivers/mmc/core/block.c
> > @@ -995,6 +995,8 @@ static int mmc_sd_num_wr_blocks(struct mmc_card *card, u32 *written_blocks)
> >       u32 result;
> >       __be32 *blocks;
> >       u8 resp_sz = mmc_card_ult_capacity(card) ? 8 : 4;
> > +     unsigned int noio_flag;
> > +
> >       struct mmc_request mrq = {};
> >       struct mmc_command cmd = {};
> >       struct mmc_data data = {};
> > @@ -1018,9 +1020,13 @@ static int mmc_sd_num_wr_blocks(struct mmc_card *card, u32 *written_blocks)
> >       mrq.cmd = &cmd;
> >       mrq.data = &data;
> >
> > +     noio_flag = memalloc_noio_save();
> > +
> >       blocks = kmalloc(resp_sz, GFP_KERNEL);
>
> Could have memalloc_noio_restore() here:
>
>         memalloc_noio_restore(noio_flag);
>
> but I feel maybe adding something like:
>
>         u64 __aligned(8)        tiny_io_buf;
>
> to either struct mmc_card or struct mmc_host is better?
> Ulf, any thoughts?
>

I have no strong opinion.

A third option could be to allocate the buffer dynamically in the
struct mmc_card when probing the mmc block device driver, based on
mmc_card_sd() returning true.

if (mmc_card_sd())
   card->io_buf = devm_kmalloc(&card->dev, 4, GFP_KERNEL);

Kind regards
Uffe
Avri Altman Oct. 16, 2024, 3:20 p.m. UTC | #3
> On Tue, 15 Oct 2024 at 11:44, Adrian Hunter <adrian.hunter@intel.com>
> wrote:
> >
> > On 14/10/24 14:44, Avri Altman wrote:
> > > While reviewing the SDUC series, Adrian made a comment concerning
> > > the memory allocation code in mmc_sd_num_wr_blocks() - see [1].
> > > Prevent memory allocations from triggering I/O operations while
> > > ACMD22 is in progress.
> > >
> > > [1] https://www.spinics.net/lists/linux-mmc/msg82199.html
> > >
> > > Suggested-by: Adrian Hunter <adrian.hunter@intel.com>
> > > Signed-off-by: Avri Altman <avri.altman@wdc.com>
> > > Cc: stable@vger.kernel.org
> > > ---
> > >  drivers/mmc/core/block.c | 10 +++++++++-
> > >  1 file changed, 9 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/mmc/core/block.c b/drivers/mmc/core/block.c
> > > index 04f3165cf9ae..042b0147d47e 100644
> > > --- a/drivers/mmc/core/block.c
> > > +++ b/drivers/mmc/core/block.c
> > > @@ -995,6 +995,8 @@ static int mmc_sd_num_wr_blocks(struct
> mmc_card *card, u32 *written_blocks)
> > >       u32 result;
> > >       __be32 *blocks;
> > >       u8 resp_sz = mmc_card_ult_capacity(card) ? 8 : 4;
> > > +     unsigned int noio_flag;
> > > +
> > >       struct mmc_request mrq = {};
> > >       struct mmc_command cmd = {};
> > >       struct mmc_data data = {};
> > > @@ -1018,9 +1020,13 @@ static int mmc_sd_num_wr_blocks(struct
> mmc_card *card, u32 *written_blocks)
> > >       mrq.cmd = &cmd;
> > >       mrq.data = &data;
> > >
> > > +     noio_flag = memalloc_noio_save();
> > > +
> > >       blocks = kmalloc(resp_sz, GFP_KERNEL);
> >
> > Could have memalloc_noio_restore() here:
> >
> >         memalloc_noio_restore(noio_flag);
> >
> > but I feel maybe adding something like:
> >
> >         u64 __aligned(8)        tiny_io_buf;
> >
> > to either struct mmc_card or struct mmc_host is better?
> > Ulf, any thoughts?
> >
> 
> I have no strong opinion.
Then I would vote to stay with Adrian's original NOIO suggestion, because:
1) My testing shows that mmc_sd_num_wr_blocks() is hardly being hit, and
2) that allocation is within the write timeout anyway

So unless you want it otherwise, will remove the redundant macro call and re-spin.

Thanks,
Avri
> 
> A third option could be to allocate the buffer dynamically in the struct
> mmc_card when probing the mmc block device driver, based on
> mmc_card_sd() returning true.
> 
> if (mmc_card_sd())
>    card->io_buf = devm_kmalloc(&card->dev, 4, GFP_KERNEL);
> 
> Kind regards
> Uffe
Ulf Hansson Oct. 17, 2024, 9:13 a.m. UTC | #4
On Wed, 16 Oct 2024 at 17:21, Avri Altman <Avri.Altman@wdc.com> wrote:
>
> > On Tue, 15 Oct 2024 at 11:44, Adrian Hunter <adrian.hunter@intel.com>
> > wrote:
> > >
> > > On 14/10/24 14:44, Avri Altman wrote:
> > > > While reviewing the SDUC series, Adrian made a comment concerning
> > > > the memory allocation code in mmc_sd_num_wr_blocks() - see [1].
> > > > Prevent memory allocations from triggering I/O operations while
> > > > ACMD22 is in progress.
> > > >
> > > > [1] https://www.spinics.net/lists/linux-mmc/msg82199.html
> > > >
> > > > Suggested-by: Adrian Hunter <adrian.hunter@intel.com>
> > > > Signed-off-by: Avri Altman <avri.altman@wdc.com>
> > > > Cc: stable@vger.kernel.org
> > > > ---
> > > >  drivers/mmc/core/block.c | 10 +++++++++-
> > > >  1 file changed, 9 insertions(+), 1 deletion(-)
> > > >
> > > > diff --git a/drivers/mmc/core/block.c b/drivers/mmc/core/block.c
> > > > index 04f3165cf9ae..042b0147d47e 100644
> > > > --- a/drivers/mmc/core/block.c
> > > > +++ b/drivers/mmc/core/block.c
> > > > @@ -995,6 +995,8 @@ static int mmc_sd_num_wr_blocks(struct
> > mmc_card *card, u32 *written_blocks)
> > > >       u32 result;
> > > >       __be32 *blocks;
> > > >       u8 resp_sz = mmc_card_ult_capacity(card) ? 8 : 4;
> > > > +     unsigned int noio_flag;
> > > > +
> > > >       struct mmc_request mrq = {};
> > > >       struct mmc_command cmd = {};
> > > >       struct mmc_data data = {};
> > > > @@ -1018,9 +1020,13 @@ static int mmc_sd_num_wr_blocks(struct
> > mmc_card *card, u32 *written_blocks)
> > > >       mrq.cmd = &cmd;
> > > >       mrq.data = &data;
> > > >
> > > > +     noio_flag = memalloc_noio_save();
> > > > +
> > > >       blocks = kmalloc(resp_sz, GFP_KERNEL);
> > >
> > > Could have memalloc_noio_restore() here:
> > >
> > >         memalloc_noio_restore(noio_flag);
> > >
> > > but I feel maybe adding something like:
> > >
> > >         u64 __aligned(8)        tiny_io_buf;
> > >
> > > to either struct mmc_card or struct mmc_host is better?
> > > Ulf, any thoughts?
> > >
> >
> > I have no strong opinion.
> Then I would vote to stay with Adrian's original NOIO suggestion, because:
> 1) My testing shows that mmc_sd_num_wr_blocks() is hardly being hit, and
> 2) that allocation is within the write timeout anyway
>
> So unless you want it otherwise, will remove the redundant macro call and re-spin.

Sounds good to me!

[...]

Kind regards
Uffe
diff mbox series

Patch

diff --git a/drivers/mmc/core/block.c b/drivers/mmc/core/block.c
index 04f3165cf9ae..042b0147d47e 100644
--- a/drivers/mmc/core/block.c
+++ b/drivers/mmc/core/block.c
@@ -995,6 +995,8 @@  static int mmc_sd_num_wr_blocks(struct mmc_card *card, u32 *written_blocks)
 	u32 result;
 	__be32 *blocks;
 	u8 resp_sz = mmc_card_ult_capacity(card) ? 8 : 4;
+	unsigned int noio_flag;
+
 	struct mmc_request mrq = {};
 	struct mmc_command cmd = {};
 	struct mmc_data data = {};
@@ -1018,9 +1020,13 @@  static int mmc_sd_num_wr_blocks(struct mmc_card *card, u32 *written_blocks)
 	mrq.cmd = &cmd;
 	mrq.data = &data;
 
+	noio_flag = memalloc_noio_save();
+
 	blocks = kmalloc(resp_sz, GFP_KERNEL);
-	if (!blocks)
+	if (!blocks) {
+		memalloc_noio_restore(noio_flag);
 		return -ENOMEM;
+	}
 
 	sg_init_one(&sg, blocks, resp_sz);
 
@@ -1041,6 +1047,8 @@  static int mmc_sd_num_wr_blocks(struct mmc_card *card, u32 *written_blocks)
 	}
 	kfree(blocks);
 
+	memalloc_noio_restore(noio_flag);
+
 	if (cmd.error || data.error)
 		return -EIO;