* [PATCH] mtd: spi-nor: Use DMA-safe buffer for JEDEC ID in spi_nor_read_id()
@ 2018-09-14 10:49 Jarkko Nikula
2018-09-15 8:28 ` Boris Brezillon
0 siblings, 1 reply; 2+ messages in thread
From: Jarkko Nikula @ 2018-09-14 10:49 UTC (permalink / raw)
To: linux-mtd
Cc: Marek Vasut, David Woodhouse, Brian Norris, Boris Brezillon,
Richard Weinberger, Jarkko Nikula
After commit 4120f8d158ef ("mtd: spi-nor: Use the spi_mem_xx() API")
there is no allocation for DMA-safe buffer when transmitting data bytes
over SPI bus in m25p80 driver.
JEDEC ID reading in spi_nor_read_id() has the buffer in stack. This is
not safe with the m25p80 driver anymore after commit 4120f8d158ef if
underlying SPI controller is using DMA for transfers.
Therefore allocate a temporary DMA-safe buffer for JEDEC ID reading.
Signed-off-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
---
I'm not an spi-nor expert at all but noticed this "WARNING: CPU: 3 PID: 154
at kernel/dma/debug.c:1191 check_for_stack+0xc2/0x1a0" after
4120f8d158ef since my test setup has DMA debugging enabled and using DMA
for SPI.
I don't know are there other places that may transfer from stack but it
looked potentially better to have the buffer allocated in
spi_nor_read_id() instead of doing allocation & copy all the places
that 4120f8d158ef touches.
---
drivers/mtd/spi-nor/spi-nor.c | 24 ++++++++++++++++++------
1 file changed, 18 insertions(+), 6 deletions(-)
diff --git a/drivers/mtd/spi-nor/spi-nor.c b/drivers/mtd/spi-nor/spi-nor.c
index f028277fb1ce..f2294ed14d8f 100644
--- a/drivers/mtd/spi-nor/spi-nor.c
+++ b/drivers/mtd/spi-nor/spi-nor.c
@@ -1269,25 +1269,37 @@ static const struct flash_info spi_nor_ids[] = {
static const struct flash_info *spi_nor_read_id(struct spi_nor *nor)
{
int tmp;
- u8 id[SPI_NOR_MAX_ID_LEN];
- const struct flash_info *info;
+ u8 *id;
+ const struct flash_info *info, *ret;
+
+ id = kzalloc(SPI_NOR_MAX_ID_LEN, GFP_KERNEL | GFP_DMA);
+ if (!id) {
+ ret = ERR_PTR(ENOMEM);
+ goto out;
+ }
tmp = nor->read_reg(nor, SPINOR_OP_RDID, id, SPI_NOR_MAX_ID_LEN);
if (tmp < 0) {
dev_dbg(nor->dev, "error %d reading JEDEC ID\n", tmp);
- return ERR_PTR(tmp);
+ ret = ERR_PTR(tmp);
+ goto out;
}
for (tmp = 0; tmp < ARRAY_SIZE(spi_nor_ids) - 1; tmp++) {
info = &spi_nor_ids[tmp];
if (info->id_len) {
- if (!memcmp(info->id, id, info->id_len))
- return &spi_nor_ids[tmp];
+ if (!memcmp(info->id, id, info->id_len)) {
+ ret = &spi_nor_ids[tmp];
+ goto out;
+ }
}
}
dev_err(nor->dev, "unrecognized JEDEC id bytes: %02x, %02x, %02x\n",
id[0], id[1], id[2]);
- return ERR_PTR(-ENODEV);
+ ret = ERR_PTR(-ENODEV);
+out:
+ kfree(id);
+ return ret;
}
static int spi_nor_read(struct mtd_info *mtd, loff_t from, size_t len,
--
2.18.0
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH] mtd: spi-nor: Use DMA-safe buffer for JEDEC ID in spi_nor_read_id()
2018-09-14 10:49 [PATCH] mtd: spi-nor: Use DMA-safe buffer for JEDEC ID in spi_nor_read_id() Jarkko Nikula
@ 2018-09-15 8:28 ` Boris Brezillon
0 siblings, 0 replies; 2+ messages in thread
From: Boris Brezillon @ 2018-09-15 8:28 UTC (permalink / raw)
To: Jarkko Nikula
Cc: linux-mtd, Marek Vasut, David Woodhouse, Brian Norris,
Richard Weinberger
On Fri, 14 Sep 2018 13:49:01 +0300
Jarkko Nikula <jarkko.nikula@linux.intel.com> wrote:
> After commit 4120f8d158ef ("mtd: spi-nor: Use the spi_mem_xx() API")
> there is no allocation for DMA-safe buffer when transmitting data bytes
> over SPI bus in m25p80 driver.
>
> JEDEC ID reading in spi_nor_read_id() has the buffer in stack. This is
> not safe with the m25p80 driver anymore after commit 4120f8d158ef if
> underlying SPI controller is using DMA for transfers.
>
> Therefore allocate a temporary DMA-safe buffer for JEDEC ID reading.
>
> Signed-off-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
> ---
> I'm not an spi-nor expert at all but noticed this "WARNING: CPU: 3 PID: 154
> at kernel/dma/debug.c:1191 check_for_stack+0xc2/0x1a0" after
> 4120f8d158ef since my test setup has DMA debugging enabled and using DMA
> for SPI.
> I don't know are there other places that may transfer from stack but it
> looked potentially better to have the buffer allocated in
> spi_nor_read_id() instead of doing allocation & copy all the places
> that 4120f8d158ef touches.
> ---
> drivers/mtd/spi-nor/spi-nor.c | 24 ++++++++++++++++++------
> 1 file changed, 18 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/mtd/spi-nor/spi-nor.c b/drivers/mtd/spi-nor/spi-nor.c
> index f028277fb1ce..f2294ed14d8f 100644
> --- a/drivers/mtd/spi-nor/spi-nor.c
> +++ b/drivers/mtd/spi-nor/spi-nor.c
> @@ -1269,25 +1269,37 @@ static const struct flash_info spi_nor_ids[] = {
> static const struct flash_info *spi_nor_read_id(struct spi_nor *nor)
> {
> int tmp;
> - u8 id[SPI_NOR_MAX_ID_LEN];
> - const struct flash_info *info;
> + u8 *id;
> + const struct flash_info *info, *ret;
> +
> + id = kzalloc(SPI_NOR_MAX_ID_LEN, GFP_KERNEL | GFP_DMA);
You can drop the GFP_DMA flag here, it's not needed (see the GFP_DMA
description [1] for more details).
the patch looks good otherwise.
> + if (!id) {
> + ret = ERR_PTR(ENOMEM);
> + goto out;
> + }
>
> tmp = nor->read_reg(nor, SPINOR_OP_RDID, id, SPI_NOR_MAX_ID_LEN);
> if (tmp < 0) {
> dev_dbg(nor->dev, "error %d reading JEDEC ID\n", tmp);
> - return ERR_PTR(tmp);
> + ret = ERR_PTR(tmp);
> + goto out;
> }
>
> for (tmp = 0; tmp < ARRAY_SIZE(spi_nor_ids) - 1; tmp++) {
> info = &spi_nor_ids[tmp];
> if (info->id_len) {
> - if (!memcmp(info->id, id, info->id_len))
> - return &spi_nor_ids[tmp];
> + if (!memcmp(info->id, id, info->id_len)) {
> + ret = &spi_nor_ids[tmp];
> + goto out;
> + }
> }
> }
> dev_err(nor->dev, "unrecognized JEDEC id bytes: %02x, %02x, %02x\n",
> id[0], id[1], id[2]);
> - return ERR_PTR(-ENODEV);
> + ret = ERR_PTR(-ENODEV);
> +out:
> + kfree(id);
> + return ret;
> }
>
> static int spi_nor_read(struct mtd_info *mtd, loff_t from, size_t len,
[1]https://elixir.bootlin.com/linux/v4.19-rc3/source/Documentation/kernel-hacking/hacking.rst#L320
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2018-09-15 8:28 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-09-14 10:49 [PATCH] mtd: spi-nor: Use DMA-safe buffer for JEDEC ID in spi_nor_read_id() Jarkko Nikula
2018-09-15 8:28 ` Boris Brezillon
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).