BeagleBone kernel is required to be built in Thumb2 mode for fast boot.
Gold linker is not yet perfect for this, giving an internal error:
| arm-linux-gnueabi-ld: internal error in thumb_branch_common, at .../binutils-cross-2.22-r6/binutils-2.22/gold/arm.cc:4149
While this fix can be put in generic linux.inc for all the kernels, it's
better to test the Gold linker out as much as possible and only disable it
for known issues, like this one (until it gets fixed, of course).
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
Beaglebone kernel needs to be built in Thumb2 mode to achieve fast boot
times. Unfortunately, that build is broken under newer binutils-2.22 with
the below Assembler error. The fix is to pass the specific -march=armv7-a
flag to the Assembler to make it work.
| arch/arm/boot/compressed/head.S: Assembler messages:
| arch/arm/boot/compressed/head.S:127: Error: selected processor does not support requested special purpose register -- `mrs r2,cpsr'
| arch/arm/boot/compressed/head.S:134: Error: selected processor does not support requested special purpose register -- `mrs r2,cpsr'
| arch/arm/boot/compressed/head.S:136: Error: selected processor does not support requested special purpose register -- `msr cpsr_c,r2'
| make[2]: *** [arch/arm/boot/compressed/head.o] Error 1
| make[1]: *** [arch/arm/boot/compressed/vmlinux] Error 2
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
Acked-by: Koen Kooi <koen@circuitco.com>
The psp tree added a patch to limit the beaglebone to 500MHz when powered by USB, which triggered conflicts in the patches. The 3.2.16 series has been rediffed as well, no functional changes.
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
Also change defconfig to prepare for audio support and fix lcd35 pinmux
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
3.2.16 is the final release in the 3.2-stable series by Greg KH.
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
This makes the backlight on teh 7" LCD cape work again. Also add a patch to
print the hardware revision of the cape.
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
This fixes most PM operations: cpufreq, cpuidle and suspend. Resume does not
work on beaglebone, but 3 out of 4 ain't bad :)
The M3 firmware needed for suspend/resume is used as a binary, building it
from source and including it from sysroots is planned for future commits.
Beaglebone patches have been forward ported and runtime tested.
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
The function get_kernelversion uses generated header files to extract the
version number. But those files are not available until do_compile, and
do_compileconfigs is injected before that, hence the version becomes None
and breaks in several places.
Introduce a new task do_preparekernel that calls "make prepare" in the
kernel tree to generate all the necessary files. It's also safer to call
do_configure early on.
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
And move it to a separate .inc for other kernels to use.
Signed-off-by: Denys Dmytriyenko <denis@denix.org>
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
Backport Nishants OPP/hwmod fix to get more than 600MHz out of 37xx, but don't
activate it yet. 3630ES1.2 derived silicon gets misdetected:
xM revision A:
[ 0.000000] OMAP3630 ES1.0 (l2cache iva sgx neon isp 192mhz_clk )
new xM revision C:
[ 0.000000] OMAP3630 ES1.2 (l2cache neon isp 192mhz_clk )
Since the iva hwmod is missing the OPP tables will fail and clock it back to
600MHz. But with the patch the kernel hangs in early init. This problem isn't
specific to this kernel update, it just coincided with ES1.2 silicon arriving
on my desk.
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
Also add script used to generate patches and SRC_URI
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
This fixes below error found on some distros (Gentoo and Fedora):
*** No rule to make target `lib/firmware/./', needed by `lib/firmware/ti_3410.fw'. Stop.
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
It has been completely replaced with the 'pandaboard' machine config.
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
BitBake's fetch2 unpack() seems to be picky about multi-kernel's
file://configs/* and defaults to only looking into machine-specific
location. This fixes the issue, as tested by building these platforms:
beaglebone, am335x-evm and am180x-evm.
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
* Added a defconfig file for the am335x-evm and moved the common
configs directory from the beaglebone directory to the top
of the kernel files tree.
Signed-off-by: Chase Maupin <Chase.Maupin@ti.com>
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
Denys Dmytriyenko
Koen Kooi
Denys Dmytriyenko
Chase Maupin