On Tue, 2021-01-05 at 00:23 +0100, Borislav Petkov wrote: > On Mon, Jan 04, 2021 at 02:22:50PM +0100, Borislav Petkov wrote: > > Hi folks, > > > > syscore_resume() doesn't like when the AMD iommu driver enables > > interrupts in its ->resume hook when I resume the box from suspend to > > RAM. > > > > All kinds of warnings get triggered too: > > > > [ 17.386830] WARNING: CPU: 0 PID: 2 at kernel/time/timekeeping.c:824 > > ktime_get+0x8d/0xa0 > > [ 17.386830] WARNING: CPU: 0 PID: 2 at kernel/time/timekeeping.c:867 > > ktime_get_with_offset+0xda/0xf0 > > [ 17.386830] WARNING: CPU: 0 PID: 2 at kernel/time/timekeeping.c:824 > > ktime_get+0x8d/0xa0 > > [ 17.386830] WARNING: CPU: 0 PID: 2 at kernel/time/timekeeping.c:867 > > ktime_get_with_offset+0xda/0xf0 > > [ 17.386830] WARNING: CPU: 0 PID: 1576 at drivers/base/syscore.c:103 > > syscore_resume+0x12d/0x160 > > > > but when I comment out the > > > > amd_iommu_enable_interrupts() > > > > call in the resume hook, all is fine and quiet and box resumes. > > > > I'll try to bisect later to try to pinpoint it better because I don't > > see what recent change would've caused this. But someone might have a > > better idea so CC the usual suspects. > > Ok, bisected to: > > # first bad commit: [d1adcfbb520c43c10fc22fcdccdd4204e014fb53] iommu/amd: Fix > IOMMU interrupt generation in X2APIC mode > > That patch reverts cleanly ontop of -rc2 and with it reverted, the > machine resumes fine.
diff --git a/drivers/iommu/amd/init.c b/drivers/iommu/amd/init.c
index f54cd79b43e4..6a1f7048dacc 100644
--- a/drivers/iommu/amd/init.c
+++ b/drivers/iommu/amd/init.c
@@ -1973,8 +1973,6 @@ static int iommu_setup_msi(struct amd_iommu *iommu)
return r;
}
- iommu->int_enabled = true;
-
return 0;
}
@@ -2169,6 +2167,7 @@ static int iommu_init_irq(struct amd_iommu *iommu)
if (ret)
return ret;
+ iommu->int_enabled = true;
enable_faults:
iommu_feature_enable(iommu, CONTROL_EVT_INT_EN);
?
smime.p7s
Description: S/MIME cryptographic signature
