| Message ID | 20220816225412.9095-5-asmaa@nvidia.com |
|---|---|
| State | Superseded |
| Headers | show |
| Series | i2c-mlxbf.c: bug fixes and new feature support | expand |
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 <kblaiech@nvidia.com> > Signed-off-by: Asmaa Mnebhi <asmaa@nvidia.com> > --- > 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 >
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 <wsa@kernel.org> Sent: Thursday, August 18, 2022 12:27 PM To: Asmaa Mnebhi <asmaa@nvidia.com> Cc: Wolfram Sang <wsa+renesas@sang-engineering.com>; linux-i2c@vger.kernel.org; linux-kernel@vger.kernel.org; Khalil Blaiech <kblaiech@nvidia.com> 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 <kblaiech@nvidia.com> > Signed-off-by: Asmaa Mnebhi <asmaa@nvidia.com> > --- > 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 >
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 <wsa@kernel.org> Cc: Wolfram Sang <wsa+renesas@sang-engineering.com>; linux-i2c@vger.kernel.org; linux-kernel@vger.kernel.org; Khalil Blaiech <kblaiech@nvidia.com> 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 <wsa@kernel.org> Sent: Thursday, August 18, 2022 12:27 PM To: Asmaa Mnebhi <asmaa@nvidia.com> Cc: Wolfram Sang <wsa+renesas@sang-engineering.com>; linux-i2c@vger.kernel.org; linux-kernel@vger.kernel.org; Khalil Blaiech <kblaiech@nvidia.com> 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 <kblaiech@nvidia.com> > Signed-off-by: Asmaa Mnebhi <asmaa@nvidia.com> > --- > 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 >
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?
Sure! -----Original Message----- From: Wolfram Sang <wsa@kernel.org> Sent: Thursday, August 18, 2022 4:24 PM To: Asmaa Mnebhi <asmaa@nvidia.com> Cc: linux-i2c@vger.kernel.org; linux-kernel@vger.kernel.org; Khalil Blaiech <kblaiech@nvidia.com> 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?
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. */