Re: [PATCH] SPI: bcm2835: move to the transfer_one driver model

[Martin Sperl <kernel@xxxxxxxxxxxxxxxx>]

Hi Stephen!

> all obviously an option but time consuming and an effort...
> Let us see if I find some time to test...

I started with the foundation kernels (as I know I was working
with those at the time):
3.8.13  - d996a1b91b2bf3dc06f4f4f822a56f4496457aa1 - WORKS
3.9.11  - d5572370289f698b101f3d0198b1c99f17f0d278 - WORKS
3.10.38 - 1b49b450222df26e4abf7abb6d9302f72b2ed386 - FAIL
3.12.36 - ee9b8c7d46f2b1787b1e64604acafc70f70191cf - FAIL

I will now try the same with the mainline kernel comparing 3.9
and 3.10 and maybe I can come up with something...
(unfortunately this is still quite time-consuming - 
even cross-compiling)

Note that some comments in commit messages of rpi-linux talk 
about some backport of code from 3.13, so this may be the real
reason...

Here the commit inside the foundation kernel that I refer to:

commit cff74f34502d46a6160dba6572db5f936cfa80ae
Author: ghollingworth <gordon@xxxxxxxxxxxxxxx>
Date:   Fri Mar 21 12:14:41 2014 +0000

    Support eMMC 5.1 cards

    Already upstream and in 3.13


>From the diff of drivers/mmc it also looks as if there is more 
error-handling code in the generic mmc code which may trigger
now...

One other issue with the upstream kernel is that for 3.9 there is
no spi-bcm2835 driver, so testing against does not work - it only
got included with 3.10, so testing is futile...

I will try to see if the commit 0385850e67359838f7d63 which is
right before cff74f34502d46a6160db is working or not and then
check if that commit introduced the issue.

Martin



--
To unsubscribe from this list: send the line "unsubscribe linux-spi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html