[U-Boot] sunxi: SATA link timeout fix
After updating u-boot from v2016.01 to 2017.09 i got a "SATA link 0 timeout." on my Cubietruck board. I tracked this down to this patch.
http://git.denx.de/?p=u-boot.git;a=commitdiff_plain;h=9fbb0c3aa49f4604b0342c....
After adding the "mdelay(500);" line that was lost in the path the error is gone.
Thank you.
Werner
diff --git a/board/sunxi/board.c b/board/sunxi/board.c index 70e0143..b38941f 100644 --- a/board/sunxi/board.c +++ b/board/sunxi/board.c @@ -216,6 +216,8 @@ int board_init(void) satapwr_pin = sunxi_name_to_gpio(CONFIG_SATAPWR); gpio_request(satapwr_pin, "satapwr"); gpio_direction_output(satapwr_pin, 1); + /* Give attached sata device time to power-up to avoid link timeouts */ + mdelay(500); #endif #ifdef CONFIG_MACPWR macpwr_pin = sunxi_name_to_gpio(CONFIG_MACPWR);
On Sun, Nov 5, 2017 at 7:11 PM, Werner Böllmann Werner.Boellmann@fh-dortmund.de wrote:
After updating u-boot from v2016.01 to 2017.09 i got a "SATA link 0 timeout." on my Cubietruck board. I tracked this down to this patch.
http://git.denx.de/?p=u-boot.git;a=commitdiff_plain;h=9fbb0c3aa49f4604b0342c....
Look like mdelay missing while moving to board file.
thanks!
On Tue, Nov 7, 2017 at 5:43 PM, Jagan Teki jagannadh.teki@gmail.com wrote:
On Sun, Nov 5, 2017 at 7:11 PM, Werner Böllmann Werner.Boellmann@fh-dortmund.de wrote:
After updating u-boot from v2016.01 to 2017.09 i got a "SATA link 0 timeout." on my Cubietruck board. I tracked this down to this patch.
http://git.denx.de/?p=u-boot.git;a=commitdiff_plain;h=9fbb0c3aa49f4604b0342c....
Look like mdelay missing while moving to board file.
Can someone test this?
thanks!
On Wed, Nov 8, 2017 at 11:32 AM, Jagan Teki jagannadh.teki@gmail.com wrote:
On Tue, Nov 7, 2017 at 5:43 PM, Jagan Teki jagannadh.teki@gmail.com wrote:
On Sun, Nov 5, 2017 at 7:11 PM, Werner Böllmann Werner.Boellmann@fh-dortmund.de wrote:
After updating u-boot from v2016.01 to 2017.09 i got a "SATA link 0 timeout." on my Cubietruck board. I tracked this down to this patch.
http://git.denx.de/?p=u-boot.git;a=commitdiff_plain;h=9fbb0c3aa49f4604b0342c....
Look like mdelay missing while moving to board file.
Can someone test this?
I boot my Cubietruck off a SATA drive without issues with any of those intervening releases but it's a SSD so there would be no spin up time compared to spinning rust, I don't have a spare spinning disk to test it with though.
On Wed, Nov 8, 2017 at 5:35 PM, Peter Robinson pbrobinson@gmail.com wrote:
On Wed, Nov 8, 2017 at 11:32 AM, Jagan Teki jagannadh.teki@gmail.com wrote:
On Tue, Nov 7, 2017 at 5:43 PM, Jagan Teki jagannadh.teki@gmail.com wrote:
On Sun, Nov 5, 2017 at 7:11 PM, Werner Böllmann Werner.Boellmann@fh-dortmund.de wrote:
After updating u-boot from v2016.01 to 2017.09 i got a "SATA link 0 timeout." on my Cubietruck board. I tracked this down to this patch.
http://git.denx.de/?p=u-boot.git;a=commitdiff_plain;h=9fbb0c3aa49f4604b0342c....
Look like mdelay missing while moving to board file.
Can someone test this?
I boot my Cubietruck off a SATA drive without issues with any of those intervening releases but it's a SSD so there would be no spin up time compared to spinning rust, I don't have a spare spinning disk to test it with though.
So w/o the mdelay work for you?
On Thu, Nov 09, 2017 at 02:48:18PM +0530, Jagan Teki wrote:
On Wed, Nov 8, 2017 at 5:35 PM, Peter Robinson pbrobinson@gmail.com wrote:
On Wed, Nov 8, 2017 at 11:32 AM, Jagan Teki jagannadh.teki@gmail.com wrote:
On Tue, Nov 7, 2017 at 5:43 PM, Jagan Teki jagannadh.teki@gmail.com wrote:
On Sun, Nov 5, 2017 at 7:11 PM, Werner Böllmann Werner.Boellmann@fh-dortmund.de wrote:
After updating u-boot from v2016.01 to 2017.09 i got a "SATA link 0 timeout." on my Cubietruck board. I tracked this down to this patch.
http://git.denx.de/?p=u-boot.git;a=commitdiff_plain;h=9fbb0c3aa49f4604b0342c....
Look like mdelay missing while moving to board file.
Can someone test this?
I boot my Cubietruck off a SATA drive without issues with any of those intervening releases but it's a SSD so there would be no spin up time compared to spinning rust, I don't have a spare spinning disk to test it with though.
So w/o the mdelay work for you?
I guess the point of this patch was that you need to have the delay, but the time between the two function calls might cover it, and apparently in some case doesn't.
So whether it works in one particular case is entirely moot.
Maxime
On Thu, Nov 9, 2017 at 3:06 PM, Maxime Ripard maxime.ripard@free-electrons.com wrote:
On Thu, Nov 09, 2017 at 02:48:18PM +0530, Jagan Teki wrote:
On Wed, Nov 8, 2017 at 5:35 PM, Peter Robinson pbrobinson@gmail.com wrote:
On Wed, Nov 8, 2017 at 11:32 AM, Jagan Teki jagannadh.teki@gmail.com wrote:
On Tue, Nov 7, 2017 at 5:43 PM, Jagan Teki jagannadh.teki@gmail.com wrote:
On Sun, Nov 5, 2017 at 7:11 PM, Werner Böllmann Werner.Boellmann@fh-dortmund.de wrote:
After updating u-boot from v2016.01 to 2017.09 i got a "SATA link 0 timeout." on my Cubietruck board. I tracked this down to this patch.
http://git.denx.de/?p=u-boot.git;a=commitdiff_plain;h=9fbb0c3aa49f4604b0342c....
Look like mdelay missing while moving to board file.
Can someone test this?
I boot my Cubietruck off a SATA drive without issues with any of those intervening releases but it's a SSD so there would be no spin up time compared to spinning rust, I don't have a spare spinning disk to test it with though.
So w/o the mdelay work for you?
I guess the point of this patch was that you need to have the delay, but the time between the two function calls might cover it, and apparently in some case doesn't.
So whether it works in one particular case is entirely moot.
Then, better some more debugging needed in this case.
thanks!
On Thu, Nov 09, 2017 at 05:12:08PM +0530, Jagan Teki wrote:
On Thu, Nov 9, 2017 at 3:06 PM, Maxime Ripard maxime.ripard@free-electrons.com wrote:
On Thu, Nov 09, 2017 at 02:48:18PM +0530, Jagan Teki wrote:
On Wed, Nov 8, 2017 at 5:35 PM, Peter Robinson pbrobinson@gmail.com wrote:
On Wed, Nov 8, 2017 at 11:32 AM, Jagan Teki jagannadh.teki@gmail.com wrote:
On Tue, Nov 7, 2017 at 5:43 PM, Jagan Teki jagannadh.teki@gmail.com wrote:
On Sun, Nov 5, 2017 at 7:11 PM, Werner Böllmann Werner.Boellmann@fh-dortmund.de wrote: > After updating u-boot from v2016.01 to 2017.09 i got a "SATA link 0 timeout." on my Cubietruck board. > I tracked this down to this patch. > > http://git.denx.de/?p=u-boot.git;a=commitdiff_plain;h=9fbb0c3aa49f4604b0342c....
Look like mdelay missing while moving to board file.
Can someone test this?
I boot my Cubietruck off a SATA drive without issues with any of those intervening releases but it's a SSD so there would be no spin up time compared to spinning rust, I don't have a spare spinning disk to test it with though.
So w/o the mdelay work for you?
I guess the point of this patch was that you need to have the delay, but the time between the two function calls might cover it, and apparently in some case doesn't.
So whether it works in one particular case is entirely moot.
Then, better some more debugging needed in this case.
I'm not sure why that is needed, if we need the delay, we should add it explicitly, just like this patch does.
Maxime
On Thu, Nov 9, 2017 at 5:36 PM, Maxime Ripard maxime.ripard@free-electrons.com wrote:
On Thu, Nov 09, 2017 at 05:12:08PM +0530, Jagan Teki wrote:
On Thu, Nov 9, 2017 at 3:06 PM, Maxime Ripard maxime.ripard@free-electrons.com wrote:
On Thu, Nov 09, 2017 at 02:48:18PM +0530, Jagan Teki wrote:
On Wed, Nov 8, 2017 at 5:35 PM, Peter Robinson pbrobinson@gmail.com wrote:
On Wed, Nov 8, 2017 at 11:32 AM, Jagan Teki jagannadh.teki@gmail.com wrote:
On Tue, Nov 7, 2017 at 5:43 PM, Jagan Teki jagannadh.teki@gmail.com wrote: > On Sun, Nov 5, 2017 at 7:11 PM, Werner Böllmann > Werner.Boellmann@fh-dortmund.de wrote: >> After updating u-boot from v2016.01 to 2017.09 i got a "SATA link 0 timeout." on my Cubietruck board. >> I tracked this down to this patch. >> >> http://git.denx.de/?p=u-boot.git;a=commitdiff_plain;h=9fbb0c3aa49f4604b0342c.... > > Look like mdelay missing while moving to board file.
Can someone test this?
I boot my Cubietruck off a SATA drive without issues with any of those intervening releases but it's a SSD so there would be no spin up time compared to spinning rust, I don't have a spare spinning disk to test it with though.
So w/o the mdelay work for you?
I guess the point of this patch was that you need to have the delay, but the time between the two function calls might cover it, and apparently in some case doesn't.
So whether it works in one particular case is entirely moot.
Then, better some more debugging needed in this case.
I'm not sure why that is needed, if we need the delay, we should add it explicitly, just like this patch does.
Since it works on particular case, can't we verify non-working area? adding delay is hard for me but look like this patch needed.
thanks!
On Thu, Nov 9, 2017 at 5:56 PM, Jagan Teki jagannadh.teki@gmail.com wrote:
On Thu, Nov 9, 2017 at 5:36 PM, Maxime Ripard maxime.ripard@free-electrons.com wrote:
On Thu, Nov 09, 2017 at 05:12:08PM +0530, Jagan Teki wrote:
On Thu, Nov 9, 2017 at 3:06 PM, Maxime Ripard maxime.ripard@free-electrons.com wrote:
On Thu, Nov 09, 2017 at 02:48:18PM +0530, Jagan Teki wrote:
On Wed, Nov 8, 2017 at 5:35 PM, Peter Robinson pbrobinson@gmail.com wrote:
On Wed, Nov 8, 2017 at 11:32 AM, Jagan Teki jagannadh.teki@gmail.com wrote: > On Tue, Nov 7, 2017 at 5:43 PM, Jagan Teki jagannadh.teki@gmail.com wrote: >> On Sun, Nov 5, 2017 at 7:11 PM, Werner Böllmann >> Werner.Boellmann@fh-dortmund.de wrote: >>> After updating u-boot from v2016.01 to 2017.09 i got a "SATA link 0 timeout." on my Cubietruck board. >>> I tracked this down to this patch. >>> >>> http://git.denx.de/?p=u-boot.git;a=commitdiff_plain;h=9fbb0c3aa49f4604b0342c.... >> >> Look like mdelay missing while moving to board file. > > Can someone test this?
I boot my Cubietruck off a SATA drive without issues with any of those intervening releases but it's a SSD so there would be no spin up time compared to spinning rust, I don't have a spare spinning disk to test it with though.
So w/o the mdelay work for you?
I guess the point of this patch was that you need to have the delay, but the time between the two function calls might cover it, and apparently in some case doesn't.
So whether it works in one particular case is entirely moot.
Then, better some more debugging needed in this case.
I'm not sure why that is needed, if we need the delay, we should add it explicitly, just like this patch does.
Since it works on particular case, can't we verify non-working area? adding delay is hard for me but look like this patch needed.
Updated commit message and Applied on u-boot-sunxi/master
participants (4)
-
Jagan Teki -
Maxime Ripard -
Peter Robinson -
Werner Böllmann