Please test updated U-Boot Beagle revision detection

GoogleGroups
1

U-Boot (v1) patch [1] to detect Beagle revisions was reported to detect rev C3 wrongly (C1/2 was detected instead [2]). It seems that the state of GPIO 172 wasn't detected properly. I now changed the patch to _disable_ pull of GPIO 172 as stated by Gerald in [3].

As I don't have a C3 (and C4 prototype) and we change the pull configuration of GPIO 172, it would be nice if you could help testing. For this, updated patch and U-Boot binary in attachment. The question is if board revision matches the revision reported by U-Boot.

If we get positive testing on Ax/Bx, C1/2, C3 and C4 I'd like to send updated patch to U-Boot list.

Many thanks for testing and best regards

Dirk

Btw.: Gerald: Any news on the other necessary C4 changes?

[1] http://lists.denx.de/pipermail/u-boot/2009-December/065502.html

[2] http://lists.denx.de/pipermail/u-boot/2009-December/065575.html

[3] http://groups.google.com/group/beagleboard/msg/6f75fae713c7e3a2

u-boot.bin (175 KB)

uboot_beagle_revc4_detection_patch.txt (4.44 KB)

2

No. Khasim has not provided anything.

Gerald

3

No. Khasim has not provided anything.

I was travelling for Tech Days, I am back to work this week. I will be
sending all the patches in next two days.

Sorry to keep you all waiting.

Regards,
Khasim

4

No. Khasim has not provided anything.

I was travelling for Tech Days, I am back to work this week. I will be
sending all the patches in next two days.

Please note that we are not talking about patches. We are talking about detailed _information_ (register values?) what has to be changed for C4.

http://groups.google.com/group/beagleboard/msg/6048a53abb3782b5

Gerald: Have you been able to test the new revision detection on C4?

Best regards

Dirk

5

The revision detect works on Rev C4…

Gerald

6

Again, sorry for delayed testing, here are my observations :

7

Table Reduced version

8

Hi Khasim,

Do you like to try u-boot.bin from [1] and let us know what it reports on the (hopefully various) boards you have?

Many thanks

Dirk

[1] http://groups.google.com/group/beagleboard/msg/312ed6e276a3d1ce

9

Hi Khasim,

Do you like to try u-boot.bin from [1] and let us know what it reports on
the (hopefully various) boards you have?

On C4 - Reports as "C4"
On C3 - Reports as "C3"
On C2 - Reports as "C3"
On B5 - Reports as "C4"

Regards,
Khasim

10

Hi Khasim,

Do you like to try u-boot.bin from [1] and let us know what it reports on
the (hopefully various) boards you have?

On C4 - Reports as "C4"
On C3 - Reports as "C3"
On C2 - Reports as "C3"
On B5 - Reports as "C4"

:frowning:

Looks to me that we are not able to properly detect all revisions, then.

Gerald: Any further idea? Is there some incompatible HW configuration between the board revisions preventing us from properly detecting them?

The recent patch [1] implements what you described in

http://groups.google.com/group/beagleboard/msg/6f75fae713c7e3a2

This would result in different U-Boot (binary/build target) for C4 :frowning:

Best regards

Dirk

[1] http://groups.google.com/group/beagleboard/msg/312ed6e276a3d1ce

11

B5 has 3530ES2.x, while Cx has 3530ES3.x, so you could use that :slight_smile: It does leave B7....

regards,

Koen

12

As was already stated earlier by Steve statement, Khasim’s C4 board will NOT work. The resistor value is incorreect on his board. R112 needs to be a 510 ohm instead of the 10K that is there now.

Gerald

13

As was already stated earlier by Steve statement, Khasim's C4 board will NOT
work. The resistor value is incorreect on his board. R112 needs to be a 510
ohm instead of the 10K that is there now.

I will try correcting R112, But my B5 board doesn't give GPIO 172 as
"1", in that case do we have to take both silicon and GPIO's in
consideration ?

Regards,
Khasim

14

On B5, there is nothing connected to GPIO172. If you setup the internal pullups, it should show Hi. We didn’t start the board ID until Rev C. So, all Bx boards if setup correctly should read all ones.

Gerald

15

On B5, there is nothing connected to GPIO172. If you setup the internal
pullups, it should show Hi. We didn't start the board ID until Rev C. So,
all Bx boards if setup correctly should read all ones.

But we are putting 172 "pulldown", so 172 if not driven externally &
is always low on Bx. Does this mean C4 (111) and Bx(101) ??

Regards,
Khasim

16

On the Rev B boards, there are no resistors at all. They would all be a one if the pullups were enabled. Rev Bx would be “111”, rev C3 is a “011” and Rev C4 is “101”

Gerald

17

On the Rev B boards, there are no resistors at all. They would all be a one
if the pullups were enabled. Rev Bx would be "111", rev C3 is a "011" and
Rev C4 is "101"

After I get the R112 Corrected, I see the below:

(173,172,171)
(1,1,1) - B5
(1,1,0) - C2
(1,1,0) - C3
(1,0,1) - C4

I think this is as expected, but I had to put GPIO 172 as pull up in
the MUX. We might have to report C2/C3 together.

Dirk, you need to change the patch to put GPIO 172 back to pull up,

MUX_VAL(CP(MCSPI1_CLK), (IEN | PTU | EN | M4)) /*GPIO_171*/\
MUX_VAL(CP(MCSPI1_SIMO), (IEN | PTU | EN | M4)) /*GPIO_172*/\
MUX_VAL(CP(MCSPI1_SOMI), (IEN | PTU | EN | M4)) /*GPIO_173*/\

Regards,
Khasim

18

What you read is correct.

Gerald

19

On the Rev B boards, there are no resistors at all. They would all be a one
if the pullups were enabled. Rev Bx would be "111", rev C3 is a "011" and
Rev C4 is "101"

After I get the R112 Corrected, I see the below:

(173,172,171)
(1,1,1) - B5
(1,1,0) - C2
(1,1,0) - C3
(1,0,1) - C4

I think this is as expected, but I had to put GPIO 172 as pull up in
the MUX. We might have to report C2/C3 together.

Dirk, you need to change the patch to put GPIO 172 back to pull up,

MUX_VAL(CP(MCSPI1_CLK), (IEN | PTU | EN | M4)) /*GPIO_171*/\
MUX_VAL(CP(MCSPI1_SIMO), (IEN | PTU | EN | M4)) /*GPIO_172*/\
MUX_VAL(CP(MCSPI1_SOMI), (IEN | PTU | EN | M4)) /*GPIO_173*/\

Ok, thanks!

Updated patch and u-boot.bin for testing in attachment.

Thanks

Dirk

uboot_beagle_revc4_detection_patch.txt (4.25 KB)

u-boot.bin (175 KB)