Holger Freyther escreveu:
On Sunday 03 August 2008 17:59:02 Cesar Eduardo Barros wrote:
I knew they made a difference on 2.6.24, so I put them on the 2.6.26
defconfig before even compiling it. In fact, I copied these three
changes from a file called fix-defconfig-gta01.patch from some OE
recipe. I really have no idea if they actually are needed for 2.6.26
(but they wouldn't hurt on a GTA01, which is what I'm testing on, so...).
Well, I'm interested in fixing root causes. We do not have lowlevel debugging
enabled on the config. So setting these UARTs to 0,1,2,3,..., 255 shouldn't
make any difference. If it does we have something to fix and should put
#ifdef WHATEVER_LOW_LEVEL_DE... around these users?
so what difference does it make on 2.6.24 and 2.6.26? I would be very happy if
defconfig-2.6.2x would work for gta01 out of the box but I need someone's
help to make that happen.
I updated to the latest stable-2.6.26 and took the opportunity to try
with a clean defconfig-2.6.26 (since CONFIG_MTD_NAND_S3C2410_DEBUG has
now been disabled on it). It hangs on the very first line ("Starting
kernel..." IIRC). However, the hang didn't happen on qemu; only on the
real hardware.
Changing both variables back to =0 made it work again.
I haven't tested on 2.6.24.
--
Cesar Eduardo Barros
[EMAIL PROTECTED]
[EMAIL PROTECTED]
