[PATCH] mtd: rawnand: cadence: fix address space mixup

[PATCH] mtd: rawnand: cadence: fix address space mixup

[Arnd Bergmann]

dma_addr_t and pointers can are not interchangeable, and can
be different sizes:

drivers/mtd/nand/raw/cadence-nand-controller.c: In function 'cadence_nand_cdma_transfer':
drivers/mtd/nand/raw/cadence-nand-controller.c:1283:12: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
            (void *)dma_buf, (void *)dma_ctrl_dat,
            ^
drivers/mtd/nand/raw/cadence-nand-controller.c:1283:29: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
            (void *)dma_buf, (void *)dma_ctrl_dat,
                             ^

Use dma_addr_t consistently here, which cleans up a couple of casts
as a side-effect.

Fixes: ec4ba01e894d ("mtd: rawnand: Add new Cadence NAND driver to MTD subsystem")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
 .../mtd/nand/raw/cadence-nand-controller.c    | 21 ++++++++-----------
 1 file changed, 9 insertions(+), 12 deletions(-)

diff --git a/drivers/mtd/nand/raw/cadence-nand-controller.c b/drivers/mtd/nand/raw/cadence-nand-controller.c
index 3a36285a8d8a..5a2d7e7ffaee 100644
--- a/drivers/mtd/nand/raw/cadence-nand-controller.c
+++ b/drivers/mtd/nand/raw/cadence-nand-controller.c
@@ -402,7 +402,7 @@ struct cadence_nand_cdma_desc {
 	u16 rsvd2;
 
 	/* System/host memory address required for data DMA commands. */
-	u64 memory_pointer;
+	u64 memory_addr;
 
 	/* Status of operation. */
 	u32 status;
@@ -416,7 +416,7 @@ struct cadence_nand_cdma_desc {
 	u32 rsvd4;
 
 	/* Control data pointer. */
-	u64 ctrl_data_ptr;
+	u64 ctrl_data_addr;
 };
 
 /* Interrupt status. */
@@ -914,8 +914,8 @@ static void cadence_nand_get_caps(struct cdns_nand_ctrl *cdns_ctrl)
 /* Prepare CDMA descriptor. */
 static void
 cadence_nand_cdma_desc_prepare(struct cdns_nand_ctrl *cdns_ctrl,
-			       char nf_mem, u32 flash_ptr, char *mem_ptr,
-			       char *ctrl_data_ptr, u16 ctype)
+			       char nf_mem, u32 flash_ptr, dma_addr_t mem_addr,
+			       dma_addr_t ctrl_data_addr, u16 ctype)
 {
 	struct cadence_nand_cdma_desc *cdma_desc = cdns_ctrl->cdma_desc;
 
@@ -931,13 +931,13 @@ cadence_nand_cdma_desc_prepare(struct cdns_nand_ctrl *cdns_ctrl,
 	cdma_desc->command_flags |= CDMA_CF_DMA_MASTER;
 	cdma_desc->command_flags  |= CDMA_CF_INT;
 
-	cdma_desc->memory_pointer = (uintptr_t)mem_ptr;
+	cdma_desc->memory_addr = mem_addr;
 	cdma_desc->status = 0;
 	cdma_desc->sync_flag_pointer = 0;
 	cdma_desc->sync_arguments = 0;
 
 	cdma_desc->command_type = ctype;
-	cdma_desc->ctrl_data_ptr = (uintptr_t)ctrl_data_ptr;
+	cdma_desc->ctrl_data_addr = ctrl_data_addr;
 }
 
 static u8 cadence_nand_check_desc_error(struct cdns_nand_ctrl *cdns_ctrl,
@@ -1280,8 +1280,7 @@ cadence_nand_cdma_transfer(struct cdns_nand_ctrl *cdns_ctrl, u8 chip_nr,
 	}
 
 	cadence_nand_cdma_desc_prepare(cdns_ctrl, chip_nr, page,
-				       (void *)dma_buf, (void *)dma_ctrl_dat,
-				       ctype);
+				       dma_buf, dma_ctrl_dat, ctype);
 
 	status = cadence_nand_cdma_send_and_wait(cdns_ctrl, thread_nr);
 
@@ -1358,10 +1357,8 @@ static int cadence_nand_erase(struct nand_chip *chip, u32 page)
 	int status;
 	u8 thread_nr = cdns_chip->cs[chip->cur_cs];
 
-	cadence_nand_cdma_desc_prepare(cdns_ctrl,
-				       cdns_chip->cs[chip->cur_cs],
-				       page, NULL, NULL,
-				       CDMA_CT_ERASE);
+	cadence_nand_cdma_desc_prepare(cdns_ctrl, cdns_chip->cs[chip->cur_cs],
+				       page, 0, 0, CDMA_CT_ERASE);
 	status = cadence_nand_cdma_send_and_wait(cdns_ctrl, thread_nr);
 	if (status) {
 		dev_err(cdns_ctrl->dev, "erase operation failed\n");
-- 
2.20.0

Re: [PATCH] mtd: rawnand: cadence: fix address space mixup

[Miquel Raynal]

Hi Arnd,

Arnd Bergmann <arnd@arndb.de> wrote on Tue, 10 Dec 2019 20:59:55 +0100:

> dma_addr_t and pointers can are not interchangeable, and can
> be different sizes:
> 
> drivers/mtd/nand/raw/cadence-nand-controller.c: In function 'cadence_nand_cdma_transfer':
> drivers/mtd/nand/raw/cadence-nand-controller.c:1283:12: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
>             (void *)dma_buf, (void *)dma_ctrl_dat,
>             ^
> drivers/mtd/nand/raw/cadence-nand-controller.c:1283:29: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
>             (void *)dma_buf, (void *)dma_ctrl_dat,
>                              ^
> 
> Use dma_addr_t consistently here, which cleans up a couple of casts
> as a side-effect.
> 
> Fixes: ec4ba01e894d ("mtd: rawnand: Add new Cadence NAND driver to MTD subsystem")
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>

I just realized that I received three patches for the same issue in a
very tight timeframe about a month ago, yours was of course entirely
valid but I choose to apply the one from someone not contributing a lot
to encourage him, hope you don't mind :)


Cheers,
Miquèl

Re: [PATCH] mtd: rawnand: cadence: fix address space mixup

[Arnd Bergmann]

On Thu, Jan 9, 2020 at 8:21 PM Miquel Raynal <miquel.raynal@bootlin.com> wrote:
>
> Hi Arnd,
>
> Arnd Bergmann <arnd@arndb.de> wrote on Tue, 10 Dec 2019 20:59:55 +0100:
>
> > dma_addr_t and pointers can are not interchangeable, and can
> > be different sizes:
> >
> > drivers/mtd/nand/raw/cadence-nand-controller.c: In function 'cadence_nand_cdma_transfer':
> > drivers/mtd/nand/raw/cadence-nand-controller.c:1283:12: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
> >             (void *)dma_buf, (void *)dma_ctrl_dat,
> >             ^
> > drivers/mtd/nand/raw/cadence-nand-controller.c:1283:29: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
> >             (void *)dma_buf, (void *)dma_ctrl_dat,
> >                              ^
> >
> > Use dma_addr_t consistently here, which cleans up a couple of casts
> > as a side-effect.
> >
> > Fixes: ec4ba01e894d ("mtd: rawnand: Add new Cadence NAND driver to MTD subsystem")
> > Signed-off-by: Arnd Bergmann <arnd@arndb.de>
>
> I just realized that I received three patches for the same issue in a
> very tight timeframe about a month ago, yours was of course entirely
> valid but I choose to apply the one from someone not contributing a lot
> to encourage him, hope you don't mind :)

Sounds good to me, thanks for getting it fixed

      Arnd