Re: [PATCH v8 09/20] acpi/ghes: don't hard-code the number of sources for HEST table

From: Mauro Carvalho Chehab
Date: Tue May 20 2025 - 02:48:40 EST


Em Mon, 12 May 2025 15:42:27 -0400
"Michael S. Tsirkin" <mst@xxxxxxxxxx> escreveu:

> On Fri, Mar 07, 2025 at 08:14:38PM +0100, Mauro Carvalho Chehab wrote:
> > The current code is actually dependent on having just one error
> > structure with a single source, as any change there would cause
> > migration issues.
> >
> > As the number of sources should be arch-dependent, as it will depend on
> > what kind of notifications will exist, and how many errors can be
> > reported at the same time, change the logic to be more flexible,
> > allowing the number of sources to be defined when building the
> > HEST table by the caller.
> >
> > Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@xxxxxxxxxx>
> > Reviewed-by: Jonathan Cameron <Jonathan.Cameron@xxxxxxxxxx>
> > Reviewed-by: Igor Mammedov <imammedo@xxxxxxxxxx>
>
> Breaks CI:
>
> https://gitlab.com/mstredhat/qemu/-/jobs/10007863974
>
> I dropped the patchset for now, pls rebase and repost.

Compilation was happening fine for me. I suspect that some rebase
dropped a change here:

diff --git a/target/arm/kvm.c b/target/arm/kvm.c
index a256f9d817cd..a1492bcaeb8d 100644
--- a/target/arm/kvm.c
+++ b/target/arm/kvm.c
@@ -2360,7 +2360,7 @@ void kvm_arch_on_sigbus_vcpu(CPUState *c, int code, void *addr)
*/
if (code == BUS_MCEERR_AR) {
kvm_cpu_synchronize_state(c);
- if (!acpi_ghes_memory_errors(ags, ACPI_HEST_SRC_ID_SEA,
+ if (!acpi_ghes_memory_errors(ags, ACPI_HEST_SRC_ID_SYNC,
paddr)) {
kvm_inject_arm_sea(c);
} else {

but somehow, ninja didn't detect and was keep using an already
compiled kvm.o object. Anyway, fixed at the new series:

https://lore.kernel.org/qemu-devel/9404b305e0fefdb14dce4839c3d96ccedf0533e0.1747722973.git.mchehab+huawei@xxxxxxxxxx/T/#u

And, just in case, I removed ./build directory and recompiled it
again.

Thanks,
Mauro