Message ID | 1568123461-I95235a4f9e6929241afd5150f1ceceb45f5e28af@changeid |
---|---|
State | Superseded |
Headers | show |
Series | um: virtio support | expand |
On Tue, 2019-09-10 at 15:51 +0200, Johannes Berg wrote: > From: Johannes Berg <johannes.berg@intel.com> > > UML has its own platform-specific barrier.h under arch/x86/um/, > which should get used. Fix the build system to use it, and then > fix the barrier.h to actually compile. > > Signed-off-by: Johannes Berg <johannes.berg@intel.com> > --- > I'm debugging some odd stuff with virtio, and the next patch is to > remove dma_rmb/dma_wmb from this, or at least make them conditional > on virtio support, since if we have virtio support we really need > memory barriers there, not just compiler barriers. > . > At least in theory. My problem isn't actually fixed by this. Oh, forgot to edit this - my problem turned out to be the "unbacked memory" part that I fixed in the virtio patch. Still, I think we need this. johannes
diff --git a/arch/um/include/asm/Kbuild b/arch/um/include/asm/Kbuild index b352ed09260e..398006d27e40 100644 --- a/arch/um/include/asm/Kbuild +++ b/arch/um/include/asm/Kbuild @@ -1,5 +1,4 @@ # SPDX-License-Identifier: GPL-2.0 -generic-y += barrier.h generic-y += bpf_perf_event.h generic-y += bug.h generic-y += compat.h diff --git a/arch/x86/um/asm/barrier.h b/arch/x86/um/asm/barrier.h index f31e5d903161..eb0654f39fd2 100644 --- a/arch/x86/um/asm/barrier.h +++ b/arch/x86/um/asm/barrier.h @@ -2,14 +2,7 @@ #ifndef _ASM_UM_BARRIER_H_ #define _ASM_UM_BARRIER_H_ -#include <asm/asm.h> -#include <asm/segment.h> -#include <asm/cpufeatures.h> -#include <asm/cmpxchg.h> -#include <asm/nops.h> - -#include <linux/kernel.h> -#include <linux/irqflags.h> +#include <asm/alternative.h> /* * Force strict CPU ordering.