Re: My progress on armhf
On Mon, Aug 22, 2011 at 09:00:06PM +0000, Phil Endecott wrote:
> Right, I've also been looking at that. They do have a git tree as well. I'm
> not sure if the patches are extracted from the git tree, or if the git tree is
> built from the patches. One gotcha with that kernel is that it doesn't have the
> select4 syscall which breaks some Debian things; you might want to backport that.
The patches do look a lot like they were generated from a git tree.
It would make sense to do that. Of course since they have to turn it
into something useable by the awful LTIB, it seems a list of sequential
patches to apply had to be made, so they did. The names even appear
like they were generated from the git commit title.
> That's interesting. Is this being discussed somewhere, or did you work it out
> by poking around in Linus' tree? (I'm curious as to whether it has gone from
> FSL to Linux via Linaro, or not.)
I saw lots of commits from freescale email address in Linus's git tree.
The fact that MX53_LOCO is mentioned (which seems to be the code name
of the MX53QSB) to me was a good sign.
For example:
commit 3b18c349285a534755e8f84c273aeafac41fdf3b
Author: Yong Shen <yong.shen@freescale.com>
Date: Wed Jan 12 17:14:47 2011 +0800
ARM i.MX53 enable LOCO board bootup
1. Add Kconfig and Makefile entries
2. Add board definition
3. enable uart and fec for LOCO board
Signed-off-by: Yong Shen <yong.shen@freescale.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
It happens to be pretty close to when the BSP files were done for the
QSB as well. Probably not a coincidence.
Based on what I see on my system, I have to assume this is related to
the QSB:
root@lucid-desktop:~# cat /proc/cpuinfo
Processor : ARMv7 Processor rev 5 (v7l)
BogoMIPS : 999.42
Features : swp half thumb fastmult vfp edsp neon vfpv3
CPU implementer : 0x41
CPU architecture: 7
CPU variant : 0x2
CPU part : 0xc08
CPU revision : 5
Hardware : Freescale MX53 LOCO Board
Revision : 53021
Serial : 0000000000000000
My board is an i.MX53 rev 2.1, which is apparently what revision 53021
means.
The reason I am waiting for 3.1 to be released before trying anything is:
commit d9c927833a42b4eaae4addd031f780f4530f7a2d
Author: Arnaud Patard (Rtp) <arnaud.patard@rtp-net.org>
Date: Fri Aug 5 09:32:41 2011 +0200
iMX: Fix build for iMX53
Commit fad107086d5a869c1c07e5bb35b7b57a10ecf578 fixed the wrong test for MX51
as the MX51 addresses are wrong for MX50 and MX53 but now it's MX51 only,
UART_PADDR is not defined anymore when building for MX50/MX53.
Signed-off-by: Arnaud Patard <arnaud.patard@rtp-net.org>
Tested-by: Steev Klimaszewski <steev@genesi-usa.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Looks important, and is after 3.0 was released by Linus.
--
Len Sorensen
Reply to: