2022-08-18 16:41:36

by Wolfram Sang

[permalink] [raw]
Subject: Re: [PATCH v1 4/7] i2c-mlxbf.c: incorrect base address passed during io write

On Tue, Aug 16, 2022 at 06:54:09PM -0400, Asmaa Mnebhi wrote:
> Correct base address used during io write.

So, the driver never ever worked? Or were we lucky that it worked
somehow?

>
> Reviewed-by: Khalil Blaiech <[email protected]>
> Signed-off-by: Asmaa Mnebhi <[email protected]>
> ---
> drivers/i2c/busses/i2c-mlxbf.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/i2c/busses/i2c-mlxbf.c b/drivers/i2c/busses/i2c-mlxbf.c
> index 39051c4a6db2..02ed6983c35c 100644
> --- a/drivers/i2c/busses/i2c-mlxbf.c
> +++ b/drivers/i2c/busses/i2c-mlxbf.c
> @@ -655,7 +655,7 @@ static int mlxbf_i2c_smbus_enable(struct mlxbf_i2c_priv *priv, u8 slave,
> /* Clear status bits. */
> writel(0x0, priv->smbus->io + MLXBF_I2C_SMBUS_MASTER_STATUS);
> /* Set the cause data. */
> - writel(~0x0, priv->smbus->io + MLXBF_I2C_CAUSE_OR_CLEAR);
> + writel(~0x0, priv->mst_cause->io + MLXBF_I2C_CAUSE_OR_CLEAR);
> /* Zero PEC byte. */
> writel(0x0, priv->smbus->io + MLXBF_I2C_SMBUS_MASTER_PEC);
> /* Zero byte count. */
> --
> 2.30.1
>


2022-08-18 17:03:07

by Asmaa Mnebhi

[permalink] [raw]
Subject: RE: [PATCH v1 4/7] i2c-mlxbf.c: incorrect base address passed during io write

It only worked on our first generation of BlueField (and we were lucky I think), not in BlueField-2 gen.

-----Original Message-----
From: Wolfram Sang <[email protected]>
Sent: Thursday, August 18, 2022 12:27 PM
To: Asmaa Mnebhi <[email protected]>
Cc: Wolfram Sang <[email protected]>; [email protected]; [email protected]; Khalil Blaiech <[email protected]>
Subject: Re: [PATCH v1 4/7] i2c-mlxbf.c: incorrect base address passed during io write

On Tue, Aug 16, 2022 at 06:54:09PM -0400, Asmaa Mnebhi wrote:
> Correct base address used during io write.

So, the driver never ever worked? Or were we lucky that it worked somehow?

>
> Reviewed-by: Khalil Blaiech <[email protected]>
> Signed-off-by: Asmaa Mnebhi <[email protected]>
> ---
> drivers/i2c/busses/i2c-mlxbf.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/i2c/busses/i2c-mlxbf.c
> b/drivers/i2c/busses/i2c-mlxbf.c index 39051c4a6db2..02ed6983c35c
> 100644
> --- a/drivers/i2c/busses/i2c-mlxbf.c
> +++ b/drivers/i2c/busses/i2c-mlxbf.c
> @@ -655,7 +655,7 @@ static int mlxbf_i2c_smbus_enable(struct mlxbf_i2c_priv *priv, u8 slave,
> /* Clear status bits. */
> writel(0x0, priv->smbus->io + MLXBF_I2C_SMBUS_MASTER_STATUS);
> /* Set the cause data. */
> - writel(~0x0, priv->smbus->io + MLXBF_I2C_CAUSE_OR_CLEAR);
> + writel(~0x0, priv->mst_cause->io + MLXBF_I2C_CAUSE_OR_CLEAR);
> /* Zero PEC byte. */
> writel(0x0, priv->smbus->io + MLXBF_I2C_SMBUS_MASTER_PEC);
> /* Zero byte count. */
> --
> 2.30.1
>

2022-08-18 17:46:05

by Asmaa Mnebhi

[permalink] [raw]
Subject: RE: [PATCH v1 4/7] i2c-mlxbf.c: incorrect base address passed during io write

Apologies, just want to clarify that I was talking about the frequency fix "[PATCH v1 1/7] i2c-mlxbf.c: Fix frequency calculation".
In regards to the below bug fix "[PATCH v1 4/7] i2c-mlxbf.c: incorrect base address passed during io write", we were lucky it works because this write:
writel(~0x0, priv->smbus->io + MLXBF_I2C_CAUSE_OR_CLEAR);
didn't have any impact on the overall functionality of the read/write transactions. Writing to smbus->io + 0x18 (aka MLXBF_I2C_CAUSE_OR_CLEAR) corresponds to sc_low_timeout register which sets the timeout value before a read/write transaction aborts.

-----Original Message-----
From: Asmaa Mnebhi
Sent: Thursday, August 18, 2022 12:41 PM
To: Wolfram Sang <[email protected]>
Cc: Wolfram Sang <[email protected]>; [email protected]; [email protected]; Khalil Blaiech <[email protected]>
Subject: RE: [PATCH v1 4/7] i2c-mlxbf.c: incorrect base address passed during io write

It only worked on our first generation of BlueField (and we were lucky I think), not in BlueField-2 gen.

-----Original Message-----
From: Wolfram Sang <[email protected]>
Sent: Thursday, August 18, 2022 12:27 PM
To: Asmaa Mnebhi <[email protected]>
Cc: Wolfram Sang <[email protected]>; [email protected]; [email protected]; Khalil Blaiech <[email protected]>
Subject: Re: [PATCH v1 4/7] i2c-mlxbf.c: incorrect base address passed during io write

On Tue, Aug 16, 2022 at 06:54:09PM -0400, Asmaa Mnebhi wrote:
> Correct base address used during io write.

So, the driver never ever worked? Or were we lucky that it worked somehow?

>
> Reviewed-by: Khalil Blaiech <[email protected]>
> Signed-off-by: Asmaa Mnebhi <[email protected]>
> ---
> drivers/i2c/busses/i2c-mlxbf.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/i2c/busses/i2c-mlxbf.c
> b/drivers/i2c/busses/i2c-mlxbf.c index 39051c4a6db2..02ed6983c35c
> 100644
> --- a/drivers/i2c/busses/i2c-mlxbf.c
> +++ b/drivers/i2c/busses/i2c-mlxbf.c
> @@ -655,7 +655,7 @@ static int mlxbf_i2c_smbus_enable(struct mlxbf_i2c_priv *priv, u8 slave,
> /* Clear status bits. */
> writel(0x0, priv->smbus->io + MLXBF_I2C_SMBUS_MASTER_STATUS);
> /* Set the cause data. */
> - writel(~0x0, priv->smbus->io + MLXBF_I2C_CAUSE_OR_CLEAR);
> + writel(~0x0, priv->mst_cause->io + MLXBF_I2C_CAUSE_OR_CLEAR);
> /* Zero PEC byte. */
> writel(0x0, priv->smbus->io + MLXBF_I2C_SMBUS_MASTER_PEC);
> /* Zero byte count. */
> --
> 2.30.1
>

2022-08-18 20:43:35

by Wolfram Sang

[permalink] [raw]
Subject: Re: [PATCH v1 4/7] i2c-mlxbf.c: incorrect base address passed during io write


Thanks for the heads up.

> didn't have any impact on the overall functionality of the read/write transactions. Writing to smbus->io + 0x18 (aka MLXBF_I2C_CAUSE_OR_CLEAR) corresponds to sc_low_timeout register which sets the timeout value before a read/write transaction aborts.

I understand. Could we have the above sentences (or something similar)
in the commit message then?


Attachments:
(No filename) (391.00 B)
signature.asc (849.00 B)
Download all attachments

2022-08-18 21:34:51

by Asmaa Mnebhi

[permalink] [raw]
Subject: RE: [PATCH v1 4/7] i2c-mlxbf.c: incorrect base address passed during io write

Sure!

-----Original Message-----
From: Wolfram Sang <[email protected]>
Sent: Thursday, August 18, 2022 4:24 PM
To: Asmaa Mnebhi <[email protected]>
Cc: [email protected]; [email protected]; Khalil Blaiech <[email protected]>
Subject: Re: [PATCH v1 4/7] i2c-mlxbf.c: incorrect base address passed during io write


Thanks for the heads up.

> didn't have any impact on the overall functionality of the read/write transactions. Writing to smbus->io + 0x18 (aka MLXBF_I2C_CAUSE_OR_CLEAR) corresponds to sc_low_timeout register which sets the timeout value before a read/write transaction aborts.

I understand. Could we have the above sentences (or something similar) in the commit message then?