diff mbox series

[1/3] dmaengine: sh: rz-dmac: handle configs where one address is zero

Message ID 20240930145955.4248-2-wsa+renesas@sang-engineering.com (mailing list archive)
State Superseded
Headers show
Series dmaengine: sh: rz-dmac: add r7s72100 support | expand

Commit Message

Wolfram Sang Sept. 30, 2024, 2:59 p.m. UTC
Configs like the ones coming from the MMC subsystem will have either
'src' or 'dst' zeroed, resulting in an unknown bus width. This will bail
out on the RZ DMA driver because of the sanity check for a valid bus
width. Reorder the code, so that the check will only be applied when the
corresponding address is non-zero.

Fixes: 5000d37042a6 ("dmaengine: sh: Add DMAC driver for RZ/G2L SoC")
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
---
 drivers/dma/sh/rz-dmac.c | 25 ++++++++++++++-----------
 1 file changed, 14 insertions(+), 11 deletions(-)

Comments

Biju Das Sept. 30, 2024, 4:40 p.m. UTC | #1
Hi Wolfram Sang,

Thanks for the patch.

> -----Original Message-----
> From: Wolfram Sang <wsa+renesas@sang-engineering.com>
> Sent: Monday, September 30, 2024 4:00 PM
> Subject: [PATCH 1/3] dmaengine: sh: rz-dmac: handle configs where one address is zero
> 
> Configs like the ones coming from the MMC subsystem will have either 'src' or 'dst' zeroed, resulting
> in an unknown bus width. This will bail out on the RZ DMA driver because of the sanity check for a
> valid bus width. Reorder the code, so that the check will only be applied when the corresponding
> address is non-zero.
> 
> Fixes: 5000d37042a6 ("dmaengine: sh: Add DMAC driver for RZ/G2L SoC")
> Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
> ---
>  drivers/dma/sh/rz-dmac.c | 25 ++++++++++++++-----------
>  1 file changed, 14 insertions(+), 11 deletions(-)
> 
> diff --git a/drivers/dma/sh/rz-dmac.c b/drivers/dma/sh/rz-dmac.c index 65a27c5a7bce..811389fc9cb8
> 100644
> --- a/drivers/dma/sh/rz-dmac.c
> +++ b/drivers/dma/sh/rz-dmac.c
> @@ -601,22 +601,25 @@ static int rz_dmac_config(struct dma_chan *chan,
>  	struct rz_dmac_chan *channel = to_rz_dmac_chan(chan);
>  	u32 val;
> 
> -	channel->src_per_address = config->src_addr;
>  	channel->dst_per_address = config->dst_addr;
> -
> -	val = rz_dmac_ds_to_val_mapping(config->dst_addr_width);
> -	if (val == CHCFG_DS_INVALID)
> -		return -EINVAL;
> -
>  	channel->chcfg &= ~CHCFG_FILL_DDS_MASK;
> -	channel->chcfg |= FIELD_PREP(CHCFG_FILL_DDS_MASK, val);
> +	if (channel->dst_per_address) {
> +		val = rz_dmac_ds_to_val_mapping(config->dst_addr_width);
> +		if (val == CHCFG_DS_INVALID)
> +			return -EINVAL;
> 
> -	val = rz_dmac_ds_to_val_mapping(config->src_addr_width);
> -	if (val == CHCFG_DS_INVALID)
> -		return -EINVAL;
> +		channel->chcfg |= FIELD_PREP(CHCFG_FILL_DDS_MASK, val);
> +	}
> 
> +	channel->src_per_address = config->src_addr;
>  	channel->chcfg &= ~CHCFG_FILL_SDS_MASK;
> -	channel->chcfg |= FIELD_PREP(CHCFG_FILL_SDS_MASK, val);
> +	if (channel->src_per_address) {
> +		val = rz_dmac_ds_to_val_mapping(config->src_addr_width);
> +		if (val == CHCFG_DS_INVALID)
> +			return -EINVAL;
> +
> +		channel->chcfg |= FIELD_PREP(CHCFG_FILL_SDS_MASK, val);
> +	}

Now both code paths are identical, not sure may be introducing a helper
by passing channel, CHCFG_FILL_*_MASK and *_addr_width
will save some code??

Anyway, current code LGTM. So,

Reviewed-by: Biju Das <biju.das.jz@bp.renesas.com>


Cheers,
Biju

> 
>  	return 0;
>  }
> --
> 2.45.2
Wolfram Sang Sept. 30, 2024, 7:25 p.m. UTC | #2
> Now both code paths are identical, not sure may be introducing a helper
> by passing channel, CHCFG_FILL_*_MASK and *_addr_width
> will save some code??

I had a look and I don't think so because we'd need to pass so many
parameters to the helper, that it doesn't really save anything, I'd say.

> Anyway, current code LGTM. So,
> 
> Reviewed-by: Biju Das <biju.das.jz@bp.renesas.com>

Thanks!
Claudiu Beznea Oct. 1, 2024, 6:29 a.m. UTC | #3
On 30.09.2024 17:59, Wolfram Sang wrote:
> Configs like the ones coming from the MMC subsystem will have either
> 'src' or 'dst' zeroed, resulting in an unknown bus width. This will bail
> out on the RZ DMA driver because of the sanity check for a valid bus
> width. Reorder the code, so that the check will only be applied when the
> corresponding address is non-zero.
> 
> Fixes: 5000d37042a6 ("dmaengine: sh: Add DMAC driver for RZ/G2L SoC")
> Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>

Tested-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>

> ---
>  drivers/dma/sh/rz-dmac.c | 25 ++++++++++++++-----------
>  1 file changed, 14 insertions(+), 11 deletions(-)
> 
> diff --git a/drivers/dma/sh/rz-dmac.c b/drivers/dma/sh/rz-dmac.c
> index 65a27c5a7bce..811389fc9cb8 100644
> --- a/drivers/dma/sh/rz-dmac.c
> +++ b/drivers/dma/sh/rz-dmac.c
> @@ -601,22 +601,25 @@ static int rz_dmac_config(struct dma_chan *chan,
>  	struct rz_dmac_chan *channel = to_rz_dmac_chan(chan);
>  	u32 val;
>  
> -	channel->src_per_address = config->src_addr;
>  	channel->dst_per_address = config->dst_addr;
> -
> -	val = rz_dmac_ds_to_val_mapping(config->dst_addr_width);
> -	if (val == CHCFG_DS_INVALID)
> -		return -EINVAL;
> -
>  	channel->chcfg &= ~CHCFG_FILL_DDS_MASK;
> -	channel->chcfg |= FIELD_PREP(CHCFG_FILL_DDS_MASK, val);
> +	if (channel->dst_per_address) {
> +		val = rz_dmac_ds_to_val_mapping(config->dst_addr_width);
> +		if (val == CHCFG_DS_INVALID)
> +			return -EINVAL;
>  
> -	val = rz_dmac_ds_to_val_mapping(config->src_addr_width);
> -	if (val == CHCFG_DS_INVALID)
> -		return -EINVAL;
> +		channel->chcfg |= FIELD_PREP(CHCFG_FILL_DDS_MASK, val);
> +	}
>  
> +	channel->src_per_address = config->src_addr;
>  	channel->chcfg &= ~CHCFG_FILL_SDS_MASK;
> -	channel->chcfg |= FIELD_PREP(CHCFG_FILL_SDS_MASK, val);
> +	if (channel->src_per_address) {
> +		val = rz_dmac_ds_to_val_mapping(config->src_addr_width);
> +		if (val == CHCFG_DS_INVALID)
> +			return -EINVAL;
> +
> +		channel->chcfg |= FIELD_PREP(CHCFG_FILL_SDS_MASK, val);
> +	}
>  
>  	return 0;
>  }
Biju Das Oct. 1, 2024, 8:56 a.m. UTC | #4
Hi Wolfram Sang,

> -----Original Message-----
> From: Wolfram Sang <wsa+renesas@sang-engineering.com>
> Sent: Monday, September 30, 2024 8:25 PM
> Subject: Re: [PATCH 1/3] dmaengine: sh: rz-dmac: handle configs where one address is zero
> 
> 
> > Now both code paths are identical, not sure may be introducing a
> > helper by passing channel, CHCFG_FILL_*_MASK and *_addr_width will
> > save some code??
> 
> I had a look and I don't think so because we'd need to pass so many parameters to the helper, that it
> doesn't really save anything, I'd say.

OK, fine by me. I just meant to avoid code duplication as it is identical blocks.

Current changes are tested with RZ/G2L on with RSPI and SSI interfaces.

So,
Tested-by: Biju Das <biju.das.jz@bp.renesas.com>

Cheers,
Biju
diff mbox series

Patch

diff --git a/drivers/dma/sh/rz-dmac.c b/drivers/dma/sh/rz-dmac.c
index 65a27c5a7bce..811389fc9cb8 100644
--- a/drivers/dma/sh/rz-dmac.c
+++ b/drivers/dma/sh/rz-dmac.c
@@ -601,22 +601,25 @@  static int rz_dmac_config(struct dma_chan *chan,
 	struct rz_dmac_chan *channel = to_rz_dmac_chan(chan);
 	u32 val;
 
-	channel->src_per_address = config->src_addr;
 	channel->dst_per_address = config->dst_addr;
-
-	val = rz_dmac_ds_to_val_mapping(config->dst_addr_width);
-	if (val == CHCFG_DS_INVALID)
-		return -EINVAL;
-
 	channel->chcfg &= ~CHCFG_FILL_DDS_MASK;
-	channel->chcfg |= FIELD_PREP(CHCFG_FILL_DDS_MASK, val);
+	if (channel->dst_per_address) {
+		val = rz_dmac_ds_to_val_mapping(config->dst_addr_width);
+		if (val == CHCFG_DS_INVALID)
+			return -EINVAL;
 
-	val = rz_dmac_ds_to_val_mapping(config->src_addr_width);
-	if (val == CHCFG_DS_INVALID)
-		return -EINVAL;
+		channel->chcfg |= FIELD_PREP(CHCFG_FILL_DDS_MASK, val);
+	}
 
+	channel->src_per_address = config->src_addr;
 	channel->chcfg &= ~CHCFG_FILL_SDS_MASK;
-	channel->chcfg |= FIELD_PREP(CHCFG_FILL_SDS_MASK, val);
+	if (channel->src_per_address) {
+		val = rz_dmac_ds_to_val_mapping(config->src_addr_width);
+		if (val == CHCFG_DS_INVALID)
+			return -EINVAL;
+
+		channel->chcfg |= FIELD_PREP(CHCFG_FILL_SDS_MASK, val);
+	}
 
 	return 0;
 }