On 09/25/17 17:54 -0700, Dan Williams wrote: > On Mon, Sep 25, 2017 at 5:45 PM, Haozhong Zhang > <haozhong.zhang@xxxxxxxxx> wrote: > > On 09/25/17 17:33 -0700, Dan Williams wrote: > >> On Sun, Sep 17, 2017 at 6:56 PM, Haozhong Zhang > >> <haozhong.zhang@xxxxxxxxx> wrote: > >> > In kvm_vcpu_trigger_posted_interrupt() and pi_pre_block(), KVM > >> > assumes that PI notification events should not be suppressed when the > >> > target vCPU is not blocked. > >> > > >> > vmx_update_pi_irte() sets the SN field before changing an interrupt > >> > from posting to remapping, but it does not check the vCPU mode. > >> > Therefore, the change of SN field may break above the assumption. > >> > Besides, I don't see reasons to suppress notification events here, so > >> > remove the changes of SN field to avoid race condition. > >> > > >> > Signed-off-by: Haozhong Zhang <haozhong.zhang@xxxxxxxxx> > >> > >> Missing a "Fixes:" line and a "Cc: <stable@xxxxxxxxxxxxxxx>"? > > > > "Fixes" was added when these two patches were committed. I cc'ed to > > stable mailing list when sent these two patches. > > Yes, but that doesn't notify the stable team to do the backport. You > actually need to put "Cc: <stable@xxxxxxxxxxxxxxx>" in the patch > directly so the automatic scripts pick it up when it hits mainline. > > Otherwise, you need to send the patch to the stable with the upstream > commit id noted at the top of the patch. See the options in > Documentation/process/stable-kernel-rules.rst. Got it. I'll send them to the stable mailing list with upstream commit ids. Thanks, Haozhong
