Skip to contentSkip to conversation

mmc: renesas_sdhi_sys_dmac: use DMA_SLAVE_BUSWIDTH_UNDEFINED

PATCH In review Sergey Shtylyov 3 participants 3 messages Jan 21, 2022 – Feb 2, 2022

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 mmc: renesas_sdhi_sys_dmac: use DMA_SLAVE_BUSWIDTH_UNDEFINED

The 'dma_slave_config::{src|dst}_addr_width' fields have the *enum* type which isn't isomorphic with the *bool* type, however is used as a boolean expression. Use the *enum* dma_slave_buswidth's value corresponding to 0 instead.

Found by Linux Verification Center (linuxtesting.org) with the SVACE static analysis tool.

---
This patch is against the 'next' branch of Ulf Hansson's 'mmc.git' repo.

drivers/mmc/host/renesas_sdhi_sys_dmac.c | 4 ++--

 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/mmc/host/renesas_sdhi_sys_dmac.c +2 -2
Index: mmc/drivers/mmc/host/renesas_sdhi_sys_dmac.c
===================================================================
@@ -370,7 +370,7 @@ static void renesas_sdhi_sys_dmac_reques
 		cfg.dst_addr = res->start +
 			(CTL_SD_DATA_PORT << host->bus_shift);
 		cfg.dst_addr_width = priv->dma_priv.dma_buswidth;
-		if (!cfg.dst_addr_width)
+		if (cfg.dst_addr_width == DMA_SLAVE_BUSWIDTH_UNDEFINED)
 			cfg.dst_addr_width = DMA_SLAVE_BUSWIDTH_2_BYTES;
 		cfg.src_addr = 0;
 		ret = dmaengine_slave_config(host->chan_tx, &cfg);
@@ -389,7 +389,7 @@ static void renesas_sdhi_sys_dmac_reques
 		cfg.direction = DMA_DEV_TO_MEM;
 		cfg.src_addr = cfg.dst_addr + host->pdata->dma_rx_offset;
 		cfg.src_addr_width = priv->dma_priv.dma_buswidth;
-		if (!cfg.src_addr_width)
+		if (cfg.src_addr_width == DMA_SLAVE_BUSWIDTH_UNDEFINED)
 			cfg.src_addr_width = DMA_SLAVE_BUSWIDTH_2_BYTES;
 		cfg.dst_addr = 0;
 		ret = dmaengine_slave_config(host->chan_rx, &cfg);
#1 lore raw

+ Vinod

Quoted 8 lines · Sergey Shtylyov

The 'dma_slave_config::{src|dst}_addr_width' fields have the *enum* type which isn't isomorphic with the *bool* type, however is used as a boolean expression. Use the *enum* dma_slave_buswidth's value corresponding to 0 instead.

Found by Linux Verification Center (linuxtesting.org) with the SVACE static analysis tool.

I have no strong opinion, but just wanted to make sure this is inline with what Vinod wants.

There are not so many users checking DMA_SLAVE_BUSWIDTH_UNDEFINED, perhaps we should simply drop it and instead leave the checks below as is?

Quoted 2 lines

Kind regards
Uffe

Quoted 29 lines

---
This patch is against the 'next' branch of Ulf Hansson's 'mmc.git' repo.

drivers/mmc/host/renesas_sdhi_sys_dmac.c | 4 ++--

 1 file changed, 2 insertions(+), 2 deletions(-)

Index: mmc/drivers/mmc/host/renesas_sdhi_sys_dmac.c
===================================================================

drivers/mmc/host/renesas_sdhi_sys_dmac.c +2 -2
@@ -370,7 +370,7 @@ static void renesas_sdhi_sys_dmac_reques
                cfg.dst_addr = res->start +
                        (CTL_SD_DATA_PORT << host->bus_shift);
                cfg.dst_addr_width = priv->dma_priv.dma_buswidth;
-               if (!cfg.dst_addr_width)
+               if (cfg.dst_addr_width == DMA_SLAVE_BUSWIDTH_UNDEFINED)
                        cfg.dst_addr_width = DMA_SLAVE_BUSWIDTH_2_BYTES;
                cfg.src_addr = 0;
                ret = dmaengine_slave_config(host->chan_tx, &cfg);
@@ -389,7 +389,7 @@ static void renesas_sdhi_sys_dmac_reques
                cfg.direction = DMA_DEV_TO_MEM;
                cfg.src_addr = cfg.dst_addr + host->pdata->dma_rx_offset;
                cfg.src_addr_width = priv->dma_priv.dma_buswidth;
-               if (!cfg.src_addr_width)
+               if (cfg.src_addr_width == DMA_SLAVE_BUSWIDTH_UNDEFINED)
                        cfg.src_addr_width = DMA_SLAVE_BUSWIDTH_2_BYTES;
                cfg.dst_addr = 0;
                ret = dmaengine_slave_config(host->chan_rx, &cfg);
#2 lore raw
Quoted 12 lines · Ulf Hansson

+ Vinod

Quoted 8 lines · Sergey Shtylyov

The 'dma_slave_config::{src|dst}_addr_width' fields have the *enum* type which isn't isomorphic with the *bool* type, however is used as a boolean expression. Use the *enum* dma_slave_buswidth's value corresponding to 0 instead.

Found by Linux Verification Center (linuxtesting.org) with the SVACE static analysis tool.

Quoted 6 lines

I have no strong opinion, but just wanted to make sure this is inline with what Vinod wants.

There are not so many users checking DMA_SLAVE_BUSWIDTH_UNDEFINED, perhaps we should simply drop it and instead leave the checks below as is?

DMA_SLAVE_BUSWIDTH_UNDEFINED was defined as 0.
I think it might be overkill and should be fine to keep the check as is..

Quoted 36 lines
Quoted 3 lines

Kind regards
Uffe

Quoted 29 lines

---
This patch is against the 'next' branch of Ulf Hansson's 'mmc.git' repo.

drivers/mmc/host/renesas_sdhi_sys_dmac.c | 4 ++--

 1 file changed, 2 insertions(+), 2 deletions(-)

Index: mmc/drivers/mmc/host/renesas_sdhi_sys_dmac.c
===================================================================

drivers/mmc/host/renesas_sdhi_sys_dmac.c +2 -2
@@ -370,7 +370,7 @@ static void renesas_sdhi_sys_dmac_reques
                cfg.dst_addr = res->start +
                        (CTL_SD_DATA_PORT << host->bus_shift);
                cfg.dst_addr_width = priv->dma_priv.dma_buswidth;
-               if (!cfg.dst_addr_width)
+               if (cfg.dst_addr_width == DMA_SLAVE_BUSWIDTH_UNDEFINED)
                        cfg.dst_addr_width = DMA_SLAVE_BUSWIDTH_2_BYTES;
                cfg.src_addr = 0;
                ret = dmaengine_slave_config(host->chan_tx, &cfg);
@@ -389,7 +389,7 @@ static void renesas_sdhi_sys_dmac_reques
                cfg.direction = DMA_DEV_TO_MEM;
                cfg.src_addr = cfg.dst_addr + host->pdata->dma_rx_offset;
                cfg.src_addr_width = priv->dma_priv.dma_buswidth;
-               if (!cfg.src_addr_width)
+               if (cfg.src_addr_width == DMA_SLAVE_BUSWIDTH_UNDEFINED)
                        cfg.src_addr_width = DMA_SLAVE_BUSWIDTH_2_BYTES;
                cfg.dst_addr = 0;
                ret = dmaengine_slave_config(host->chan_rx, &cfg);
Signature · 1 line
~Vinod