spi-atmel broken in next
Subsystems (3)
Hi,
spi-atmel is now broken for all spi transfers in linux-next. This is caused by commit "spi: make sure all transfer has proper speed set". The reason is the following code in atmel_spi_transfer(...)
849 /* FIXME implement these protocol options!! */ 850 if (xfer->speed_hz) {
851 dev_dbg(&spi->dev, "no protocol options yet\n");
852 return -ENOPROTOOPT;
853 }
Now that xfer->speed_hz is always set all spi transfers will return with ENOPROTOOPT.
Not quite sure how to best fix it. Just removing the whole test makes it work again but I guess we should do something better with speed_hz here.
One little question; Is a per transfer speed setting really sensible? As far as I can see from grepping the kernel source it's not used by device drivers.
regards
Joachim Eastwood
------------------------------------------------------------------------------
The Go Parallel Website, sponsored by Intel - in partnership with Geeknet, is your hub for all things parallel software development, from weekly thought leadership blogs to news, videos, case studies, tutorials, tech docs, whitepapers, evaluation guides, and opinion stories. Check out the most recent posts - join the conversation now. http://goparallel.sourceforge.net/
A message that is not in the archive (sent off-list or not yet received): CAGhQ9VwoehjY4niDqcnb88va9vNLw=jFmxOnq6k0QP6GZJ4gqQ-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org
Quoted 12 lines · Joachim Eastwood
Hi,
spi-atmel is now broken for all spi transfers in linux-next. This is caused by commit "spi: make sure all transfer has proper speed set". The reason is the following code in
atmel_spi_transfer(...)849 /* FIXME implement these protocol options!! */ 850 if (
xfer->speed_hz) {
851dev_dbg(&spi->dev, "no protocol options yet\n");
852 return-ENOPROTOOPT;
853 }
Quoted 6 lines
Now that
xfer->speed_hzis always set all spi transfers will return with ENOPROTOOPT.Not quite sure how to best fix it. Just removing the whole test makes it work again but I guess we should do something better with
speed_hzhere.
Probably the best fix is to check instead if xfer->speed_hz is less than the bus speed. If it is then you can fail. If it is not, then continue.
g.
------------------------------------------------------------------------------
The Go Parallel Website, sponsored by Intel - in partnership with Geeknet, is your hub for all things parallel software development, from weekly thought leadership blogs to news, videos, case studies, tutorials, tech docs, whitepapers, evaluation guides, and opinion stories. Check out the most recent posts - join the conversation now. http://goparallel.sourceforge.net/
A message that is not in the archive (sent off-list or not yet received): CACxGe6u0o0V3Jm9-iyizKH5KX=dPe-CL0bH-Kc7KM_k6gAk69A-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org
PATCH 1/1 spi/atmel: fix speed_hz check in atmel_spi_transfer()
atmel_spi_transfer() would check speed_hz and fail if the speed was changed in the transfer. After commit "spi: make sure all transfer has proper speed set" this would happen on all transfers.
Change speed_hz check to only fail if a lower speed than max is requested.
---
Hi Grant,
Fix for spi-atmel as per your suggestion; check if speed_hz is less than bus max speed.
regards
Joachim Eastwood
drivers/spi/spi-atmel.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/spi/spi-atmel.c +2 -2
@@ -847,8 +847,8 @@ static int atmel_spi_transfer(struct spi_device *spi, struct spi_message *msg) }/* FIXME implement these protocol options!! */ -if (xfer->speed_hz) { -dev_dbg (&spi->dev,"no protocol options yet\n" ); +if (xfer->speed_hz < spi->max_speed_hz) { +dev_dbg (&spi->dev,"can't change speed in transfer\n" );return -ENOPROTOOPT ; }
Signature · 8 lines
1.8.0 ------------------------------------------------------------------------------ Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_feb
A message that is not in the archive (sent off-list or not yet received): 1361310297-28002-1-git-send-email-manabian-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org
Nicolas Ferre
Acked-byRe: [PATCH] spi/atmel: fix speed_hz check in atmel_spi_transfer()
On 02/19/2013 10:44 PM, Joachim Eastwood :
Quoted 9 lines
atmel_spi_transfer()would checkspeed_hzand fail if the speed was changed in the transfer. After commit "spi: make sure all transfer has proper speed set" this would happen on all transfers.Change
speed_hzcheck to only fail if a lower speed than max is requested.Signed-off-by Joachim Eastwood <[email protected]>
Thanks for spotting this Joachim.
Bye,
Quoted 28 lines
---
Hi Grant,Fix for spi-atmel as per your suggestion; check if
speed_hzis less than bus max speed.regards
Joachim Eastwooddrivers/spi/spi-atmel.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)drivers/spi/spi-atmel.c +2
-2@@ -847,8 +847,8 @@ static int atmel_spi_transfer(struct spi_device *spi, struct spi_message *msg) }/* FIXME implement these protocol options!! */ -if (xfer->speed_hz) { -dev_dbg (&spi->dev,"no protocol options yet\n" ); +if (xfer->speed_hz < spi->max_speed_hz) { +dev_dbg (&spi->dev,"can't change speed in transfer\n" );return -ENOPROTOOPT ; }
Signature · 7 lines
Nicolas Ferre ------------------------------------------------------------------------------ Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_feb
A message that is not in the archive (sent off-list or not yet received): [email protected]
Grant Likely
AppliedRe: [PATCH] spi/atmel: fix speed_hz check in atmel_spi_transfer()
Quoted 6 lines · Nicolas Ferre
On 02/19/2013 10:44 PM, Joachim Eastwood :
Quoted 5 lines
atmel_spi_transfer()would checkspeed_hzand fail if the speed was changed in the transfer. After commit "spi: make sure all transfer has proper speed set" this would happen on all transfers.
Quoted 8 lines
Quoted 4 lines
Change
speed_hzcheck to only fail if a lower speed than max is requested.Signed-off-by Joachim Eastwood <[email protected]>Acked-by Nicolas Ferre <[email protected]>Thanks for spotting this Joachim.
Applied, thanks.
g.
------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_feb