* [PATCH] i2c: pasemi: Wait for write xfers to finish
@ 2022-03-29 18:38 ` Martin Povišer
0 siblings, 0 replies; 9+ messages in thread
From: Martin Povišer @ 2022-03-29 18:38 UTC (permalink / raw)
To: Hector Martin, Sven Peter, Michael Ellerman
Cc: Martin Povišer, linux-kernel, Olof Johansson,
Paul Mackerras, Alyssa Rosenzweig, Jean Delvare, Janne Grunau,
linuxppc-dev, linux-arm-kernel, linux-i2c
Wait for completion of write transfers before returning from the driver.
At first sight it may seem advantageous to leave write transfers queued
for the controller to carry out on its own time, but there's a couple of
issues with it:
* Driver doesn't check for FIFO space.
* The queued writes can complete while the driver is in its I2C read
transfer path which means it will get confused by the raising of
XEN (the 'transaction ended' signal). This can cause a spurious
ENODATA error due to premature reading of the MRXFIFO register.
Adding the wait fixes some unreliability issues with the driver. There's
some efficiency cost to it (especially with pasemi_smb_waitready doing
its polling), but that will be alleviated once the driver receives
interrupt support.
Fixes: beb58aa39e6e ("i2c: PA Semi SMBus driver")
Signed-off-by: Martin Povišer <povik+lin@cutebit.org>
---
Tested on Apple's t8103 chip. To my knowledge the PA Semi controller
in its pre-Apple occurences behaves the same as far as this patch is
concerned.
drivers/i2c/busses/i2c-pasemi-core.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/drivers/i2c/busses/i2c-pasemi-core.c b/drivers/i2c/busses/i2c-pasemi-core.c
index 7728c8460dc0..9028ffb58cc0 100644
--- a/drivers/i2c/busses/i2c-pasemi-core.c
+++ b/drivers/i2c/busses/i2c-pasemi-core.c
@@ -137,6 +137,12 @@ static int pasemi_i2c_xfer_msg(struct i2c_adapter *adapter,
TXFIFO_WR(smbus, msg->buf[msg->len-1] |
(stop ? MTXFIFO_STOP : 0));
+
+ if (stop) {
+ err = pasemi_smb_waitready(smbus);
+ if (err)
+ goto reset_out;
+ }
}
return 0;
--
2.33.0
^ permalink raw reply related [flat|nested] 9+ messages in thread* [PATCH] i2c: pasemi: Wait for write xfers to finish
@ 2022-03-29 18:38 ` Martin Povišer
0 siblings, 0 replies; 9+ messages in thread
From: Martin Povišer @ 2022-03-29 18:38 UTC (permalink / raw)
To: Hector Martin, Sven Peter, Michael Ellerman
Cc: Alyssa Rosenzweig, Benjamin Herrenschmidt, Paul Mackerras,
Jean Delvare, Olof Johansson, linux-arm-kernel, linuxppc-dev,
linux-i2c, linux-kernel, Janne Grunau, Martin Povišer
Wait for completion of write transfers before returning from the driver.
At first sight it may seem advantageous to leave write transfers queued
for the controller to carry out on its own time, but there's a couple of
issues with it:
* Driver doesn't check for FIFO space.
* The queued writes can complete while the driver is in its I2C read
transfer path which means it will get confused by the raising of
XEN (the 'transaction ended' signal). This can cause a spurious
ENODATA error due to premature reading of the MRXFIFO register.
Adding the wait fixes some unreliability issues with the driver. There's
some efficiency cost to it (especially with pasemi_smb_waitready doing
its polling), but that will be alleviated once the driver receives
interrupt support.
Fixes: beb58aa39e6e ("i2c: PA Semi SMBus driver")
Signed-off-by: Martin Povišer <povik+lin@cutebit.org>
---
Tested on Apple's t8103 chip. To my knowledge the PA Semi controller
in its pre-Apple occurences behaves the same as far as this patch is
concerned.
drivers/i2c/busses/i2c-pasemi-core.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/drivers/i2c/busses/i2c-pasemi-core.c b/drivers/i2c/busses/i2c-pasemi-core.c
index 7728c8460dc0..9028ffb58cc0 100644
--- a/drivers/i2c/busses/i2c-pasemi-core.c
+++ b/drivers/i2c/busses/i2c-pasemi-core.c
@@ -137,6 +137,12 @@ static int pasemi_i2c_xfer_msg(struct i2c_adapter *adapter,
TXFIFO_WR(smbus, msg->buf[msg->len-1] |
(stop ? MTXFIFO_STOP : 0));
+
+ if (stop) {
+ err = pasemi_smb_waitready(smbus);
+ if (err)
+ goto reset_out;
+ }
}
return 0;
--
2.33.0
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply related [flat|nested] 9+ messages in thread* Re: [PATCH] i2c: pasemi: Wait for write xfers to finish
2022-03-29 18:38 ` Martin Povišer
(?)
@ 2022-03-29 19:58 ` Sven Peter
-1 siblings, 0 replies; 9+ messages in thread
From: Sven Peter @ 2022-03-29 19:58 UTC (permalink / raw)
To: Martin Povišer, Hector Martin, Michael Ellerman
Cc: Alyssa Rosenzweig, Benjamin Herrenschmidt, Paul Mackerras,
Jean Delvare, Olof Johansson, linux-arm-kernel, linuxppc-dev,
linux-i2c, linux-kernel, Janne Grunau
On Tue, Mar 29, 2022, at 20:38, Martin Povišer wrote:
> Wait for completion of write transfers before returning from the driver.
> At first sight it may seem advantageous to leave write transfers queued
> for the controller to carry out on its own time, but there's a couple of
> issues with it:
>
> * Driver doesn't check for FIFO space.
Maybe we should also check that in a follow-up patch :-)
>
> * The queued writes can complete while the driver is in its I2C read
> transfer path which means it will get confused by the raising of
> XEN (the 'transaction ended' signal). This can cause a spurious
> ENODATA error due to premature reading of the MRXFIFO register.
>
> Adding the wait fixes some unreliability issues with the driver. There's
> some efficiency cost to it (especially with pasemi_smb_waitready doing
> its polling), but that will be alleviated once the driver receives
> interrupt support.
>
> Fixes: beb58aa39e6e ("i2c: PA Semi SMBus driver")
> Signed-off-by: Martin Povišer <povik+lin@cutebit.org>
> ---
Reviewed-by: Sven Peter <sven@svenpeter.dev>
>
> Tested on Apple's t8103 chip. To my knowledge the PA Semi controller
> in its pre-Apple occurences behaves the same as far as this patch is
> concerned.
>
> drivers/i2c/busses/i2c-pasemi-core.c | 6 ++++++
> 1 file changed, 6 insertions(+)
>
> diff --git a/drivers/i2c/busses/i2c-pasemi-core.c
> b/drivers/i2c/busses/i2c-pasemi-core.c
> index 7728c8460dc0..9028ffb58cc0 100644
> --- a/drivers/i2c/busses/i2c-pasemi-core.c
> +++ b/drivers/i2c/busses/i2c-pasemi-core.c
> @@ -137,6 +137,12 @@ static int pasemi_i2c_xfer_msg(struct i2c_adapter
> *adapter,
>
> TXFIFO_WR(smbus, msg->buf[msg->len-1] |
> (stop ? MTXFIFO_STOP : 0));
> +
> + if (stop) {
> + err = pasemi_smb_waitready(smbus);
> + if (err)
> + goto reset_out;
> + }
Looks like pasemi_smb_xfer doesn't suffer from the same issue.
I wonder if every device connected to the bus on the original PA Semi boards
only used that path and that's why no one noticed it.
Sven
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [PATCH] i2c: pasemi: Wait for write xfers to finish
@ 2022-03-29 19:58 ` Sven Peter
0 siblings, 0 replies; 9+ messages in thread
From: Sven Peter @ 2022-03-29 19:58 UTC (permalink / raw)
To: Martin Povišer, Hector Martin, Michael Ellerman
Cc: linux-kernel, Olof Johansson, Paul Mackerras, Alyssa Rosenzweig,
Jean Delvare, Janne Grunau, linuxppc-dev, linux-arm-kernel,
linux-i2c
On Tue, Mar 29, 2022, at 20:38, Martin Povišer wrote:
> Wait for completion of write transfers before returning from the driver.
> At first sight it may seem advantageous to leave write transfers queued
> for the controller to carry out on its own time, but there's a couple of
> issues with it:
>
> * Driver doesn't check for FIFO space.
Maybe we should also check that in a follow-up patch :-)
>
> * The queued writes can complete while the driver is in its I2C read
> transfer path which means it will get confused by the raising of
> XEN (the 'transaction ended' signal). This can cause a spurious
> ENODATA error due to premature reading of the MRXFIFO register.
>
> Adding the wait fixes some unreliability issues with the driver. There's
> some efficiency cost to it (especially with pasemi_smb_waitready doing
> its polling), but that will be alleviated once the driver receives
> interrupt support.
>
> Fixes: beb58aa39e6e ("i2c: PA Semi SMBus driver")
> Signed-off-by: Martin Povišer <povik+lin@cutebit.org>
> ---
Reviewed-by: Sven Peter <sven@svenpeter.dev>
>
> Tested on Apple's t8103 chip. To my knowledge the PA Semi controller
> in its pre-Apple occurences behaves the same as far as this patch is
> concerned.
>
> drivers/i2c/busses/i2c-pasemi-core.c | 6 ++++++
> 1 file changed, 6 insertions(+)
>
> diff --git a/drivers/i2c/busses/i2c-pasemi-core.c
> b/drivers/i2c/busses/i2c-pasemi-core.c
> index 7728c8460dc0..9028ffb58cc0 100644
> --- a/drivers/i2c/busses/i2c-pasemi-core.c
> +++ b/drivers/i2c/busses/i2c-pasemi-core.c
> @@ -137,6 +137,12 @@ static int pasemi_i2c_xfer_msg(struct i2c_adapter
> *adapter,
>
> TXFIFO_WR(smbus, msg->buf[msg->len-1] |
> (stop ? MTXFIFO_STOP : 0));
> +
> + if (stop) {
> + err = pasemi_smb_waitready(smbus);
> + if (err)
> + goto reset_out;
> + }
Looks like pasemi_smb_xfer doesn't suffer from the same issue.
I wonder if every device connected to the bus on the original PA Semi boards
only used that path and that's why no one noticed it.
Sven
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [PATCH] i2c: pasemi: Wait for write xfers to finish
@ 2022-03-29 19:58 ` Sven Peter
0 siblings, 0 replies; 9+ messages in thread
From: Sven Peter @ 2022-03-29 19:58 UTC (permalink / raw)
To: Martin Povišer, Hector Martin, Michael Ellerman
Cc: Alyssa Rosenzweig, Benjamin Herrenschmidt, Paul Mackerras,
Jean Delvare, Olof Johansson, linux-arm-kernel, linuxppc-dev,
linux-i2c, linux-kernel, Janne Grunau
On Tue, Mar 29, 2022, at 20:38, Martin Povišer wrote:
> Wait for completion of write transfers before returning from the driver.
> At first sight it may seem advantageous to leave write transfers queued
> for the controller to carry out on its own time, but there's a couple of
> issues with it:
>
> * Driver doesn't check for FIFO space.
Maybe we should also check that in a follow-up patch :-)
>
> * The queued writes can complete while the driver is in its I2C read
> transfer path which means it will get confused by the raising of
> XEN (the 'transaction ended' signal). This can cause a spurious
> ENODATA error due to premature reading of the MRXFIFO register.
>
> Adding the wait fixes some unreliability issues with the driver. There's
> some efficiency cost to it (especially with pasemi_smb_waitready doing
> its polling), but that will be alleviated once the driver receives
> interrupt support.
>
> Fixes: beb58aa39e6e ("i2c: PA Semi SMBus driver")
> Signed-off-by: Martin Povišer <povik+lin@cutebit.org>
> ---
Reviewed-by: Sven Peter <sven@svenpeter.dev>
>
> Tested on Apple's t8103 chip. To my knowledge the PA Semi controller
> in its pre-Apple occurences behaves the same as far as this patch is
> concerned.
>
> drivers/i2c/busses/i2c-pasemi-core.c | 6 ++++++
> 1 file changed, 6 insertions(+)
>
> diff --git a/drivers/i2c/busses/i2c-pasemi-core.c
> b/drivers/i2c/busses/i2c-pasemi-core.c
> index 7728c8460dc0..9028ffb58cc0 100644
> --- a/drivers/i2c/busses/i2c-pasemi-core.c
> +++ b/drivers/i2c/busses/i2c-pasemi-core.c
> @@ -137,6 +137,12 @@ static int pasemi_i2c_xfer_msg(struct i2c_adapter
> *adapter,
>
> TXFIFO_WR(smbus, msg->buf[msg->len-1] |
> (stop ? MTXFIFO_STOP : 0));
> +
> + if (stop) {
> + err = pasemi_smb_waitready(smbus);
> + if (err)
> + goto reset_out;
> + }
Looks like pasemi_smb_xfer doesn't suffer from the same issue.
I wonder if every device connected to the bus on the original PA Semi boards
only used that path and that's why no one noticed it.
Sven
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH] i2c: pasemi: Wait for write xfers to finish
2022-03-29 18:38 ` Martin Povišer
(?)
@ 2022-04-15 21:12 ` Wolfram Sang
-1 siblings, 0 replies; 9+ messages in thread
From: Wolfram Sang @ 2022-04-15 21:12 UTC (permalink / raw)
To: Martin Povišer
Cc: Hector Martin, Sven Peter, Michael Ellerman, Alyssa Rosenzweig,
Benjamin Herrenschmidt, Paul Mackerras, Jean Delvare,
Olof Johansson, linux-arm-kernel, linuxppc-dev, linux-i2c,
linux-kernel, Janne Grunau
[-- Attachment #1: Type: text/plain, Size: 1045 bytes --]
On Tue, Mar 29, 2022 at 08:38:17PM +0200, Martin Povišer wrote:
> Wait for completion of write transfers before returning from the driver.
> At first sight it may seem advantageous to leave write transfers queued
> for the controller to carry out on its own time, but there's a couple of
> issues with it:
>
> * Driver doesn't check for FIFO space.
>
> * The queued writes can complete while the driver is in its I2C read
> transfer path which means it will get confused by the raising of
> XEN (the 'transaction ended' signal). This can cause a spurious
> ENODATA error due to premature reading of the MRXFIFO register.
>
> Adding the wait fixes some unreliability issues with the driver. There's
> some efficiency cost to it (especially with pasemi_smb_waitready doing
> its polling), but that will be alleviated once the driver receives
> interrupt support.
>
> Fixes: beb58aa39e6e ("i2c: PA Semi SMBus driver")
> Signed-off-by: Martin Povišer <povik+lin@cutebit.org>
Applied to for-current, thanks!
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [PATCH] i2c: pasemi: Wait for write xfers to finish
@ 2022-04-15 21:12 ` Wolfram Sang
0 siblings, 0 replies; 9+ messages in thread
From: Wolfram Sang @ 2022-04-15 21:12 UTC (permalink / raw)
To: Martin Povišer
Cc: Hector Martin, Sven Peter, Michael Ellerman, Alyssa Rosenzweig,
Benjamin Herrenschmidt, Paul Mackerras, Jean Delvare,
Olof Johansson, linux-arm-kernel, linuxppc-dev, linux-i2c,
linux-kernel, Janne Grunau
[-- Attachment #1.1: Type: text/plain, Size: 1045 bytes --]
On Tue, Mar 29, 2022 at 08:38:17PM +0200, Martin Povišer wrote:
> Wait for completion of write transfers before returning from the driver.
> At first sight it may seem advantageous to leave write transfers queued
> for the controller to carry out on its own time, but there's a couple of
> issues with it:
>
> * Driver doesn't check for FIFO space.
>
> * The queued writes can complete while the driver is in its I2C read
> transfer path which means it will get confused by the raising of
> XEN (the 'transaction ended' signal). This can cause a spurious
> ENODATA error due to premature reading of the MRXFIFO register.
>
> Adding the wait fixes some unreliability issues with the driver. There's
> some efficiency cost to it (especially with pasemi_smb_waitready doing
> its polling), but that will be alleviated once the driver receives
> interrupt support.
>
> Fixes: beb58aa39e6e ("i2c: PA Semi SMBus driver")
> Signed-off-by: Martin Povišer <povik+lin@cutebit.org>
Applied to for-current, thanks!
[-- Attachment #1.2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
[-- Attachment #2: Type: text/plain, Size: 176 bytes --]
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [PATCH] i2c: pasemi: Wait for write xfers to finish
@ 2022-04-15 21:12 ` Wolfram Sang
0 siblings, 0 replies; 9+ messages in thread
From: Wolfram Sang @ 2022-04-15 21:12 UTC (permalink / raw)
To: Martin Povišer
Cc: Sven Peter, Hector Martin, linux-kernel, Olof Johansson,
Paul Mackerras, linux-arm-kernel, Jean Delvare, Janne Grunau,
linuxppc-dev, Alyssa Rosenzweig, linux-i2c
[-- Attachment #1: Type: text/plain, Size: 1045 bytes --]
On Tue, Mar 29, 2022 at 08:38:17PM +0200, Martin Povišer wrote:
> Wait for completion of write transfers before returning from the driver.
> At first sight it may seem advantageous to leave write transfers queued
> for the controller to carry out on its own time, but there's a couple of
> issues with it:
>
> * Driver doesn't check for FIFO space.
>
> * The queued writes can complete while the driver is in its I2C read
> transfer path which means it will get confused by the raising of
> XEN (the 'transaction ended' signal). This can cause a spurious
> ENODATA error due to premature reading of the MRXFIFO register.
>
> Adding the wait fixes some unreliability issues with the driver. There's
> some efficiency cost to it (especially with pasemi_smb_waitready doing
> its polling), but that will be alleviated once the driver receives
> interrupt support.
>
> Fixes: beb58aa39e6e ("i2c: PA Semi SMBus driver")
> Signed-off-by: Martin Povišer <povik+lin@cutebit.org>
Applied to for-current, thanks!
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 9+ messages in thread