On 21/02/2021 04.17, Marc Zyngier wrote:
On 2021-02-20 19:13, Krzysztof Kozlowski wrote:
On Thu, Feb 18, 2021 at 11:01:21PM +0900, Hector Martin wrote:
On 16/02/2021 03.51, Krzysztof Kozlowski wrote:
Also fix a bug checking the return value, which should use IS_ERR().
No, no, no. We never, never combine fixing bugs with some rework.
However devm_ioremap() returns NULL so where is the error?
Sorry, this was a commit message mistake. The code is correct and so
is the
patch: just the NULL check is correct for the previous variant and
IS_ERR is
correct for devm_ioremap_resource. I confused myself while writing the
commit message after the fact.
Did you test your patches on existing platforms? If not, please mark all
of them as RFT on next submission, so Greg does not pick them too fast.
I unfortunately don't have any Exynos devices where I could test the
code (I
have a couple but no serial connections, and I have no idea if mailine
would
run on them). I'll mark v3 as RFT.
If you have one of Odroid boards with Exynos, then you can nicely test
Exynos. Others - depends, on board.
Anyway I can test them for you. I just want to be sure that Greg waits
for this testing.
Worse case, QEMU has some Exynos4210 emulation that is usable.
That's a good point; better than nothing, certainly.
Does anyone have a known good example of booting an exynos kernel under
qemu? I tried building a plain 5.11 arm exynos_defconfig and booting it,
but without much luck:
$ qemu-system-arm -kernel arch/arm/boot/zImage -append
"console=ttySAC0,115200n8 debug" -dtb
arch/arm/boot/dts/exynos4210-universal_c210.dtb -nographic -serial
mon:stdio -M smdkc210 -smp 2
(I also tried without the -dtb option, in case qemu provides something
usable)
Of course I'll still mark v3 as RFT, I just thought I might as well try
qemu.
--
Hector Martin (marcan@xxxxxxxxx)
Public Key: https://mrcn.st/pub
