2023-07-24 10:19:57

by Minjie Du

[permalink] [raw]
Subject: [PATCH v4] ata: pata_arasan_cf: Use dev_err_probe() instead dev_err() in data_xfer()

It is possible for dma_request_chan() to return EPROBE_DEFER, which means
acdev->host->dev is not ready yet.
At this point dev_err() will have no output.

Signed-off-by: Minjie Du <[email protected]>
---
V1 - V4:
Fix code format.
---
drivers/ata/pata_arasan_cf.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/ata/pata_arasan_cf.c b/drivers/ata/pata_arasan_cf.c
index 6ab294322e79..b32d47112c0a 100644
--- a/drivers/ata/pata_arasan_cf.c
+++ b/drivers/ata/pata_arasan_cf.c
@@ -529,7 +529,8 @@ static void data_xfer(struct work_struct *work)
/* dma_request_channel may sleep, so calling from process context */
acdev->dma_chan = dma_request_chan(acdev->host->dev, "data");
if (IS_ERR(acdev->dma_chan)) {
- dev_err(acdev->host->dev, "Unable to get dma_chan\n");
+ dev_err_probe(acdev->host->dev, PTR_ERR(acdev->dma_chan),
+ "Unable to get dma_chan\n");
acdev->dma_chan = NULL;
goto chan_request_fail;
}
--
2.39.0



2023-07-24 10:38:14

by Sergei Shtylyov

[permalink] [raw]
Subject: Re: [PATCH v4] ata: pata_arasan_cf: Use dev_err_probe() instead dev_err() in data_xfer()

On 7/24/23 12:57 PM, Minjie Du wrote:

> It is possible for dma_request_chan() to return EPROBE_DEFER, which means
> acdev->host->dev is not ready yet.
> At this point dev_err() will have no output.
>
> Signed-off-by: Minjie Du <[email protected]>
> ---
> V1 - V4:
> Fix code format.

I'm not seeing any fixing in v4

> ---
> drivers/ata/pata_arasan_cf.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/ata/pata_arasan_cf.c b/drivers/ata/pata_arasan_cf.c
> index 6ab294322e79..b32d47112c0a 100644
> --- a/drivers/ata/pata_arasan_cf.c
> +++ b/drivers/ata/pata_arasan_cf.c
> @@ -529,7 +529,8 @@ static void data_xfer(struct work_struct *work)
> /* dma_request_channel may sleep, so calling from process context */
> acdev->dma_chan = dma_request_chan(acdev->host->dev, "data");
> if (IS_ERR(acdev->dma_chan)) {
> - dev_err(acdev->host->dev, "Unable to get dma_chan\n");
> + dev_err_probe(acdev->host->dev, PTR_ERR(acdev->dma_chan),

"Unable to get dma_chan\n");

> acdev->dma_chan = NULL;
> goto chan_request_fail;
> }

That's how the dev_err_probe() continuation line shouuld be
formatted...

MBR, Sergey

2023-07-24 10:52:02

by Damien Le Moal

[permalink] [raw]
Subject: Re: [PATCH v4] ata: pata_arasan_cf: Use dev_err_probe() instead dev_err() in data_xfer()

On 7/24/23 18:57, Minjie Du wrote:
> It is possible for dma_request_chan() to return EPROBE_DEFER, which means
> acdev->host->dev is not ready yet.
> At this point dev_err() will have no output.
>
> Signed-off-by: Minjie Du <[email protected]>
> ---
> V1 - V4:
> Fix code format.

You are not fixing anything. You keep resending the same, which is missing the
dev_err_probe() message alignment like I (and Sergey) keep telling you. Can you
please take our comments seriously and act on them ? If you do not understand,
then say so and I will explain in more details.

> ---
> drivers/ata/pata_arasan_cf.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/ata/pata_arasan_cf.c b/drivers/ata/pata_arasan_cf.c
> index 6ab294322e79..b32d47112c0a 100644
> --- a/drivers/ata/pata_arasan_cf.c
> +++ b/drivers/ata/pata_arasan_cf.c
> @@ -529,7 +529,8 @@ static void data_xfer(struct work_struct *work)
> /* dma_request_channel may sleep, so calling from process context */
> acdev->dma_chan = dma_request_chan(acdev->host->dev, "data");
> if (IS_ERR(acdev->dma_chan)) {
> - dev_err(acdev->host->dev, "Unable to get dma_chan\n");
> + dev_err_probe(acdev->host->dev, PTR_ERR(acdev->dma_chan),
> + "Unable to get dma_chan\n");

Just use this please, formatted *exactly* like this.

dev_err_probe(acdev->host->dev, PTR_ERR(acdev->dma_chan),
"Unable to get dma_chan\n");

> acdev->dma_chan = NULL;
> goto chan_request_fail;
> }

--
Damien Le Moal
Western Digital Research


2023-07-24 11:16:36

by Viresh Kumar

[permalink] [raw]
Subject: Re: [PATCH v4] ata: pata_arasan_cf: Use dev_err_probe() instead dev_err() in data_xfer()

On 24-07-23, 17:57, Minjie Du wrote:
> It is possible for dma_request_chan() to return EPROBE_DEFER, which means
> acdev->host->dev is not ready yet.
> At this point dev_err() will have no output.
>
> Signed-off-by: Minjie Du <[email protected]>
> ---
> V1 - V4:
> Fix code format.
> ---
> drivers/ata/pata_arasan_cf.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/ata/pata_arasan_cf.c b/drivers/ata/pata_arasan_cf.c
> index 6ab294322e79..b32d47112c0a 100644
> --- a/drivers/ata/pata_arasan_cf.c
> +++ b/drivers/ata/pata_arasan_cf.c
> @@ -529,7 +529,8 @@ static void data_xfer(struct work_struct *work)
> /* dma_request_channel may sleep, so calling from process context */
> acdev->dma_chan = dma_request_chan(acdev->host->dev, "data");
> if (IS_ERR(acdev->dma_chan)) {
> - dev_err(acdev->host->dev, "Unable to get dma_chan\n");
> + dev_err_probe(acdev->host->dev, PTR_ERR(acdev->dma_chan),
> + "Unable to get dma_chan\n");
> acdev->dma_chan = NULL;
> goto chan_request_fail;
> }

Acked-by: Viresh Kumar <[email protected]>

--
viresh