On 6/1/22 04:07, Arnd Bergmann wrote:
On Tue, May 31, 2022 at 6:01 PM Huacai Chen <chenhuacai@xxxxxxxxxx> wrote:
On Tue, May 31, 2022 at 7:15 PM Arnd Bergmann <arnd@xxxxxxxxxx> wrote:
On Tue, May 31, 2022 at 10:17 AM Huacai Chen <chenhuacai@xxxxxxxxxx> wrote:
https://git.kernel.org/pub/scm/linux/kernel/git/chenhuacai/linux-loongson.git/log/?h=loongarch-next
has been updated. Now this branch droped irqchip drivers and pci
drivers. But the existing irqchip drivers need some small adjustment
to avoid build errors [1], and I hope Marc can give an Acked-by.
Ok, glad you got that working.
What about the ACPI changes? I see that these are needed to get a clean build,
and as I understood it, they are supposed to get merged through the
acpica tree.
I think the acpica bits could be dropped with some effort too; the main
dependency on the various ACPI 6.5 tables are the SMP bits, which relies
on the new MADT CPUINTC tables. While the others also provide
information, they're not as fundamental as this, and even this CPUINTC
piece can be taken out given we can't run this branch on any real
LoongArch hardware after all (due to the irqchip changes being backed
out), I think we can just leave the remaining bits dummy-initialized
with some simple comment. We can review once the new branch with only
arch/loongarch changes is out.
This branch can be built with defconfig and allmodconfig (except
drivers/platform/surface/aggregator/controller.c, because it requires
8bit/16bit cmpxchg, which I was told to remove their support).
Right, that is ok to keep in there, we should fix that by adding a Kconfig
dependency for the driver. It looks like it has a CONFIG_ACPI dependency,
so it is currently limited to x86/arm64/ia64, which all have the short
cmpxchg(),
but in reality this driver can only work on x86 and arm64.
In case this isn't obvious to any non-native English speaker: the driver
is written for the Microsoft Surface, which only has x86 and arm64
variants to this date and the list is probably not going to expand in
the foreseeable future, so the word "work" here takes a quite literal
sense. ;-)
I agree a tiny fix for that driver could be added later that limits the
driver to X86 || ARM64. As a popular product line, adding support for
yet another architecture would be a news visible enough for the crowd
that they'll come and tweak the Kconfig themselves.
