Skip to contentSkip to conversation

Revert "am335x_evm.h: If mmcdev and bootpart switch to mmcdev 1, so should mmcroot."

PATCH New Tom Rini 3 participants 9 messages Oct 8, 2013

lore
Share & tools
lore

Tom Rini

author
#0 lore raw

PATCH 1/1 Revert "am335x_evm.h: If mmcdev and bootpart switch to mmcdev 1, so should mmcroot."

Upon further inspection and review and chatting with kernel folks, what happens here is that what mmcblk# a device gets is based on probe order. So a system with an SD card inserted with place eMMC on mmcblk1, but without an SD card, it will be on mmcblk0. So U-boot can only provide a best guess. In this case, if no SD card is present, we would want to pass mmcblk0p2 still. If an SD card is present, it woudl be able to provide a uEnv.txt that would be loaded (even if the kernel is NOT there) which can still update mmcroot variable.

This reverts commit 827512fb1154c05c6eb1e2259e936df55c98a535.

Cc Robert P. J. Day <[email protected]>
---
 include/configs/am335x_evm.h |    1 -
 1 file changed, 1 deletion(-)
include/configs/am335x_evm.h +0 -1
index e8e5275..f746e48 100644
@@ -171,7 +171,6 @@
 	"run mmcboot;" \
 	"setenv mmcdev 1; " \
 	"setenv bootpart 1:2; " \
-	"setenv mmcroot /dev/mmcblk1p2 ro; " \
 	"run mmcboot;" \
 	"run nandboot;"
 
Signature · 1 line
1.7.9.5
#1 lore raw

PATCH 1/1 Revert "am335x_evm.h: If mmcdev and bootpart switch to mmcdev 1, so should mmcroot."

Quoted 30 lines · Tom Rini

Upon further inspection and review and chatting with kernel folks, what happens here is that what mmcblk# a device gets is based on probe order. So a system with an SD card inserted with place eMMC on mmcblk1, but without an SD card, it will be on mmcblk0. So U-boot can only provide a best guess. In this case, if no SD card is present, we would want to pass mmcblk0p2 still. If an SD card is present, it woudl be able to provide a uEnv.txt that would be loaded (even if the kernel is NOT there) which can still update mmcroot variable.

This reverts commit 827512fb1154c05c6eb1e2259e936df55c98a535.

Cc Robert P. J. Day <[email protected]>
---
 include/configs/am335x_evm.h |    1 -
 1 file changed, 1 deletion(-)
include/configs/am335x_evm.h +0 -2
index e8e5275..f746e48 100644
@@ -171,7 +171,6 @@
 	"run mmcboot;" \
 	"setenv mmcdev 1; " \
 	"setenv bootpart 1:2; " \
-	"setenv mmcroot /dev/mmcblk1p2 ro; " \
 	"run mmcboot;" \
 	"run nandboot;"
 
--
Quoted 1 line

1.7.9.5

are you sure about this? note that, in the above, you've *already*
tried to mmcboot off of device mmcblk0 and, for one reason or another, that failed. so u-boot then *explicitly* switches to mmc dev 1 with:

setenv mmcdev 1
setenv bootpart 1:2

why would it not make sense to also switch mmcroot there as well?

rday

Signature · 7 lines
========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================
author
#2 lore raw

PATCH 1/1 Revert "am335x_evm.h: If mmcdev and bootpart switch to mmcdev 1, so should mmcroot."

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 38 lines · Robert P. J. Day
Quoted 30 lines · Tom Rini

Upon further inspection and review and chatting with kernel folks, what happens here is that what mmcblk# a device gets is based on probe order. So a system with an SD card inserted with place eMMC on mmcblk1, but without an SD card, it will be on mmcblk0. So U-boot can only provide a best guess. In this case, if no SD card is present, we would want to pass mmcblk0p2 still. If an SD card is present, it woudl be able to provide a uEnv.txt that would be loaded (even if the kernel is NOT there) which can still update mmcroot variable.

This reverts commit 827512fb1154c05c6eb1e2259e936df55c98a535.

Cc Robert P. J. Day <[email protected]>
---
 include/configs/am335x_evm.h |    1 -
 1 file changed, 1 deletion(-)
include/configs/am335x_evm.h +0 -2
index e8e5275..f746e48 100644
@@ -171,7 +171,6 @@
 	"run mmcboot;" \
 	"setenv mmcdev 1; " \
 	"setenv bootpart 1:2; " \
-	"setenv mmcroot /dev/mmcblk1p2 ro; " \
 	"run mmcboot;" \
 	"run nandboot;"
 
--
Quoted 1 line

1.7.9.5

are you sure about this? note that, in the above, you've *already*
tried to mmcboot off of device mmcblk0 and, for one reason or another, that failed. so u-boot then *explicitly* switches to mmc dev 1 with:

Quoted 4 lines

setenv mmcdev 1
setenv bootpart 1:2

why would it not make sense to also switch mmcroot there as well?

Yes, I'm very sure after talking with the kernel folks and playing with a tree with all of the EDMA stuff fixed up so we can use modern kernels again. But you can see this on the vendor kernel BBB ships with too. If you have an SD card and eMMC, SD card will be mmcblk0 and eMMC mmcblk1. If you however just root from eMMC, it will become mmcblk0. The only time you have mmcblk1 is when you have both SD card inserted and eMMC present.

Note that omap5 uevm has a different issue I shall be posting a patch for soon, need to get the board back into my setup.

- --
Tom
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQIcBAEBAgAGBQJSVCqWAAoJENk4IS6UOR1WUtkP+wXraIrib77S9sWFY75230Dd +Q4+x9HZ7AOsWX3M9SVSpKIzVhAMwrFoJWGEGwkd43YTMfDyTFT8FkJvLYu2jMRd yk2LhDbEuju5Jy/yabrU8iq5VHCWnFsyNJ8S4dpshDK0H1ZL7mDrsMiPc183C8op os+fYgN9IEOMf+vUVUBuWMtHGR0/UqyTwtrt+cwBsa2uzJQpOdfewssJ7+WY4B5I 4KZ4TkjxODrD2JGT5TbmU4oVPP4qq+nHOHdw7evOs//KdrPdhWyF/O/pyVKgDCB8 TL4xLTtityCgZF1RjI3p0pSiEVlJWTMq1RdnRWedhfh47Rdz3uZxHNkmFKMWs1sf lVORRzt+ZYveW3ERcWgiRAycQ7EwfFEV2J/VtI6nh0P7KkjAP7M/fcbe2MM3U9ua ckeF+w24rsprkXH8VAgU0M+UR+hqUC2me4YIY2oVDfTS0jPhdvw4W9P7SN4J+Lsg KirIjOLGlG1dOQWDzGu/4DaUG2l0t0N2VBAXUV10R5fJdvVzx+ex6KT/tnMdcpza PvobZbJAQoDTlf2KaU9fqIjMApQB1DD58WLuAW4XP6ni+ciPC8N2SZTIQnCeAqE+ fLgl6+1kaS0rOo8946GCAelQL8eADF68m4FOIzubzu27In4KGp29E7r1fhlLpBGv ULxGh0iJeonV6wbLr939
=/UqY
-----END PGP SIGNATURE-----

author
#3 lore raw

PATCH 1/1 Revert "am335x_evm.h: If mmcdev and bootpart switch to mmcdev 1, so should mmcroot."

In reply to Tom Rini
Quoted 9 lines · Tom Rini

Upon further inspection and review and chatting with kernel folks, what happens here is that what mmcblk# a device gets is based on probe order. So a system with an SD card inserted with place eMMC on mmcblk1, but without an SD card, it will be on mmcblk0. So U-boot can only provide a best guess. In this case, if no SD card is present, we would want to pass mmcblk0p2 still. If an SD card is present, it woudl be able to provide a uEnv.txt that would be loaded (even if the kernel is NOT there) which can still update mmcroot variable.

Quoted 4 lines

This reverts commit 827512fb1154c05c6eb1e2259e936df55c98a535.

Cc Robert P. J. Day <[email protected]>

Applied to u-boot/master.

Signature · 8 lines
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20131008/ad58f2a1/attachment.pgp>
#4 lore raw

PATCH 1/1 Revert "am335x_evm.h: If mmcdev and bootpart switch to mmcdev 1, so should mmcroot."

On Tue, 8 Oct 2013, Tom Rini wrote:

... snip ...

Quoted 1 line

Applied to u-boot/master.

dumb question but what does it mean to say "Applied to
u-boot/master" when it clearly has not been applied to master? i can see posts like that, but doing a "git pull" produces nothing. i am on the u-boot mainline, and the "master" branch, so what am i misunderstanding? thanks.

rday

Signature · 7 lines
========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================
#5 lore raw

PATCH 1/1 Revert "am335x_evm.h: If mmcdev and bootpart switch to mmcdev 1, so should mmcroot."

Quoted 11 lines · Robert P. J. Day

On Tue, 8 Oct 2013, Tom Rini wrote:

... snip ...

Quoted 1 line

Applied to u-boot/master.

dumb question but what does it mean to say "Applied to
u-boot/master" when it clearly has not been applied to master? i can see posts like that, but doing a "git pull" produces nothing. i am on the u-boot mainline, and the "master" branch, so what am i misunderstanding? thanks.

The public git/http server is only synced every 6hours...

Regards,

Signature · 2 lines
Robert Nelson
http://www.rcn-ee.com/
#6 lore raw

PATCH 1/1 Revert "am335x_evm.h: If mmcdev and bootpart switch to mmcdev 1, so should mmcroot."

Quoted 7 lines · Robert Nelson
Quoted 6 lines · Robert P. J. Day

On Tue, 8 Oct 2013, Tom Rini wrote:

... snip ...

Quoted 1 line

Applied to u-boot/master.

Quoted 7 lines
Quoted 5 lines

dumb question but what does it mean to say "Applied to
u-boot/master" when it clearly has not been applied to master? i can see posts like that, but doing a "git pull" produces nothing. i am on the u-boot mainline, and the "master" branch, so what am i misunderstanding? thanks.

The public git/http server is only synced every 6hours...

i'm not convinced. i'm looking at this posting from yesterday:

http://lists.denx.de/pipermail/u-boot/2013-October/164517.html

timestamped "Mon, Oct 07, 2013 at 10:02:08PM +0200" which contains a change i'm interested in and, as of this minute, it doesn't appear in master.

rday

author
#7 lore raw

PATCH 1/1 Revert "am335x_evm.h: If mmcdev and bootpart switch to mmcdev 1, so should mmcroot."

Quoted 17 lines · Robert P. J. Day
Quoted 7 lines · Robert Nelson
Quoted 6 lines · Robert P. J. Day

On Tue, 8 Oct 2013, Tom Rini wrote:

... snip ...

Quoted 1 line

Applied to u-boot/master.

Quoted 7 lines
Quoted 5 lines

dumb question but what does it mean to say "Applied to
u-boot/master" when it clearly has not been applied to master? i can see posts like that, but doing a "git pull" produces nothing. i am on the u-boot mainline, and the "master" branch, so what am i misunderstanding? thanks.

The public git/http server is only synced every 6hours...

Quoted 7 lines

i'm not convinced. i'm looking at this posting from yesterday:

http://lists.denx.de/pipermail/u-boot/2013-October/164517.html

timestamped "Mon, Oct 07, 2013 at 10:02:08PM +0200" which contains a change i'm interested in and, as of this minute, it doesn't appear in master.

That's when Albert sent his email. My reply is "Tue Oct 8 21:33:30 CEST 2013", so yes, we haven't hit the re-sync window just yet.

Signature · 8 lines
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20131008/1f7a5160/attachment.pgp>
#8 lore raw

PATCH 1/1 Revert "am335x_evm.h: If mmcdev and bootpart switch to mmcdev 1, so should mmcroot."

Quoted 20 lines · Tom Rini
Quoted 19 lines · Robert P. J. Day
Quoted 7 lines · Robert Nelson
Quoted 6 lines · Robert P. J. Day

On Tue, 8 Oct 2013, Tom Rini wrote:

... snip ...

Quoted 1 line

Applied to u-boot/master.

Quoted 7 lines
Quoted 5 lines

dumb question but what does it mean to say "Applied to
u-boot/master" when it clearly has not been applied to master? i can see posts like that, but doing a "git pull" produces nothing. i am on the u-boot mainline, and the "master" branch, so what am i misunderstanding? thanks.

The public git/http server is only synced every 6hours...

i'm not convinced. i'm looking at this posting from yesterday:

Quoted 8 lines
Quoted 5 lines

http://lists.denx.de/pipermail/u-boot/2013-October/164517.html

timestamped "Mon, Oct 07, 2013 at 10:02:08PM +0200" which contains a change i'm interested in and, as of this minute, it doesn't appear in master.

That's when Albert sent his email. My reply is "Tue Oct 8 21:33:30 CEST 2013", so yes, we haven't hit the re-sync window just yet.

ah, got it.

rday

Signature · 7 lines
========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================