From: Scott Wood <scottwood@freescale.com> To: Alexander Graf <agraf@suse.de> Cc: <kvm-ppc@vger.kernel.org>, KVM list <kvm@vger.kernel.org> Subject: Re: [PATCH 17/38] KVM: PPC: BookE: Add support for vcpu->mode Date: Mon, 20 Aug 2012 20:41:32 -0500 [thread overview] Message-ID: <5032E74C.7010100@freescale.com> (raw) In-Reply-To: <6F911453-9056-4070-8279-7ADA8C28FD47@suse.de> On 08/15/2012 04:29 AM, Alexander Graf wrote: > > On 15.08.2012, at 03:17, Scott Wood wrote: > >> On 08/14/2012 07:26 PM, Alexander Graf wrote: >>> >>> On 15.08.2012, at 02:17, Scott Wood wrote: >>> >>>> On 08/14/2012 06:04 PM, Alexander Graf wrote: >>>>> Generic KVM code might want to know whether we are inside guest context >>>>> or outside. It also wants to be able to push us out of guest context. >>>>> >>>>> Add support to the BookE code for the generic vcpu->mode field that describes >>>>> the above states. >>>>> >>>>> Signed-off-by: Alexander Graf <agraf@suse.de> >>>>> --- >>>>> arch/powerpc/kvm/booke.c | 11 +++++++++++ >>>>> 1 files changed, 11 insertions(+), 0 deletions(-) >>>>> >>>>> diff --git a/arch/powerpc/kvm/booke.c b/arch/powerpc/kvm/booke.c >>>>> index bcf87fe..70a86c0 100644 >>>>> --- a/arch/powerpc/kvm/booke.c >>>>> +++ b/arch/powerpc/kvm/booke.c >>>>> @@ -501,6 +501,15 @@ static int kvmppc_prepare_to_enter(struct kvm_vcpu *vcpu) >>>>> continue; >>>>> } >>>>> >>>>> + if (vcpu->mode == EXITING_GUEST_MODE) { >>>>> + r = 1; >>>>> + break; >>>>> + } >>>>> + >>>>> + /* Going into guest context! Yay! */ >>>>> + vcpu->mode = IN_GUEST_MODE; >>>>> + smp_wmb(); >>>>> + >>>>> break; >>>>> } >>>> >>>> Normally on entry to this function mode should be OUTSIDE_GUEST_MODE, >>>> right? How could it possibly be EXITING_GUEST_MODE then, since that >>>> only replaces IN_GUEST_MODE? >>>> >>>> This doesn't match what x86 does with mode on entry. Mode is supposed >>>> to be set to IN_GUEST_MODE before requests are checked. >>>> >>>> I'm not sure what the point of EXITING_GUEST_MODE is at all, compared to >>>> just waiting until after interrupts are disabled before setting >>>> IN_GUEST_MODE (which we do on ppc, but not on x86 even though it seems >>>> like a trivial change), plus the existing ordering between mode and >>>> requests. >>> >>> Well, the only real use case I could find for the mode was the remote >>> vcpu kick. If we're not outside of guest mode, we get an IPI to >>> notify us that requests are outstanding. >> >> I'm curious why this is done so differently for broadcast requests than >> for single-cpu requests. >> >>> So I only get us into OUTSIDE_GUEST_MODE when we really exit >>> __vcpu_run, thus are in user space. That doesn't reflect what x86 >>> does, right, but so doesn't our whole loop concept. >> >> OK. We still need to do ordering like x86 does, because otherwise >> there's a race where we could check requests before the request bit is >> set, and still have make_all_cpus_request see OUTSIDE_GUEST_MODE and not >> send an IPI. > > Could you please send a patch showing what workflow you envision? I'll try to do this tomorrow. > The code as is should work, just be inefficient at times, right? No, we could fail to send the IPI -- couldn't that result in the guest using a stale TLB entry? -Scott
WARNING: multiple messages have this Message-ID (diff)
From: Scott Wood <scottwood@freescale.com> To: Alexander Graf <agraf@suse.de> Cc: kvm-ppc@vger.kernel.org, KVM list <kvm@vger.kernel.org> Subject: Re: [PATCH 17/38] KVM: PPC: BookE: Add support for vcpu->mode Date: Tue, 21 Aug 2012 01:41:32 +0000 [thread overview] Message-ID: <5032E74C.7010100@freescale.com> (raw) In-Reply-To: <6F911453-9056-4070-8279-7ADA8C28FD47@suse.de> On 08/15/2012 04:29 AM, Alexander Graf wrote: > > On 15.08.2012, at 03:17, Scott Wood wrote: > >> On 08/14/2012 07:26 PM, Alexander Graf wrote: >>> >>> On 15.08.2012, at 02:17, Scott Wood wrote: >>> >>>> On 08/14/2012 06:04 PM, Alexander Graf wrote: >>>>> Generic KVM code might want to know whether we are inside guest context >>>>> or outside. It also wants to be able to push us out of guest context. >>>>> >>>>> Add support to the BookE code for the generic vcpu->mode field that describes >>>>> the above states. >>>>> >>>>> Signed-off-by: Alexander Graf <agraf@suse.de> >>>>> --- >>>>> arch/powerpc/kvm/booke.c | 11 +++++++++++ >>>>> 1 files changed, 11 insertions(+), 0 deletions(-) >>>>> >>>>> diff --git a/arch/powerpc/kvm/booke.c b/arch/powerpc/kvm/booke.c >>>>> index bcf87fe..70a86c0 100644 >>>>> --- a/arch/powerpc/kvm/booke.c >>>>> +++ b/arch/powerpc/kvm/booke.c >>>>> @@ -501,6 +501,15 @@ static int kvmppc_prepare_to_enter(struct kvm_vcpu *vcpu) >>>>> continue; >>>>> } >>>>> >>>>> + if (vcpu->mode = EXITING_GUEST_MODE) { >>>>> + r = 1; >>>>> + break; >>>>> + } >>>>> + >>>>> + /* Going into guest context! Yay! */ >>>>> + vcpu->mode = IN_GUEST_MODE; >>>>> + smp_wmb(); >>>>> + >>>>> break; >>>>> } >>>> >>>> Normally on entry to this function mode should be OUTSIDE_GUEST_MODE, >>>> right? How could it possibly be EXITING_GUEST_MODE then, since that >>>> only replaces IN_GUEST_MODE? >>>> >>>> This doesn't match what x86 does with mode on entry. Mode is supposed >>>> to be set to IN_GUEST_MODE before requests are checked. >>>> >>>> I'm not sure what the point of EXITING_GUEST_MODE is at all, compared to >>>> just waiting until after interrupts are disabled before setting >>>> IN_GUEST_MODE (which we do on ppc, but not on x86 even though it seems >>>> like a trivial change), plus the existing ordering between mode and >>>> requests. >>> >>> Well, the only real use case I could find for the mode was the remote >>> vcpu kick. If we're not outside of guest mode, we get an IPI to >>> notify us that requests are outstanding. >> >> I'm curious why this is done so differently for broadcast requests than >> for single-cpu requests. >> >>> So I only get us into OUTSIDE_GUEST_MODE when we really exit >>> __vcpu_run, thus are in user space. That doesn't reflect what x86 >>> does, right, but so doesn't our whole loop concept. >> >> OK. We still need to do ordering like x86 does, because otherwise >> there's a race where we could check requests before the request bit is >> set, and still have make_all_cpus_request see OUTSIDE_GUEST_MODE and not >> send an IPI. > > Could you please send a patch showing what workflow you envision? I'll try to do this tomorrow. > The code as is should work, just be inefficient at times, right? No, we could fail to send the IPI -- couldn't that result in the guest using a stale TLB entry? -Scott
next prev parent reply other threads:[~2012-08-21 1:41 UTC|newest] Thread overview: 150+ messages / expand[flat|nested] mbox.gz Atom feed top 2012-08-14 23:04 [PULL 00/38] ppc patch queue 2012-08-15 Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 01/38] PPC: epapr: create define for return code value of success Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 02/38] KVM: PPC: use definitions in epapr header for hcalls Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 03/38] KVM: PPC: add pvinfo for hcall opcodes on e500mc/e5500 Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 04/38] KVM: PPC: Add support for ePAPR idle hcall in host kernel Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 05/38] KVM: PPC: ev_idle hcall support for e500 guests Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 06/38] PPC: select EPAPR_PARAVIRT for all users of epapr hcalls Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 07/38] powerpc/fsl-soc: use CONFIG_EPAPR_PARAVIRT for hcalls Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 08/38] PPC: Don't use hardcoded opcode for ePAPR hcall invocation Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 09/38] KVM: PPC: PR: Use generic tracepoint for guest exit Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 10/38] KVM: PPC: Expose SYNC cap based on mmu notifiers Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 11/38] KVM: PPC: BookE: Expose remote TLB flushes in debugfs Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 12/38] KVM: PPC: E500: Fix clear_tlb_refs Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 13/38] KVM: PPC: Book3S HV: Fix incorrect branch in H_CEDE code Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 14/38] KVM: PPC: Quieten message about allocating linear regions Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 15/38] powerpc/epapr: export epapr_hypercall_start Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 16/38] KVM: PPC: BookE: Add check_requests helper function Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-15 0:10 ` Scott Wood 2012-08-15 0:10 ` Scott Wood 2012-08-15 0:13 ` Alexander Graf 2012-08-15 0:13 ` Alexander Graf 2012-08-15 0:20 ` Scott Wood 2012-08-15 0:20 ` Scott Wood 2012-08-15 18:28 ` Marcelo Tosatti 2012-08-15 18:28 ` Marcelo Tosatti 2012-08-14 23:04 ` [PATCH 17/38] KVM: PPC: BookE: Add support for vcpu->mode Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-15 0:17 ` Scott Wood 2012-08-15 0:17 ` Scott Wood 2012-08-15 0:26 ` Alexander Graf 2012-08-15 0:26 ` Alexander Graf 2012-08-15 1:17 ` Scott Wood 2012-08-15 1:17 ` Scott Wood 2012-08-15 9:29 ` Alexander Graf 2012-08-15 9:29 ` Alexander Graf 2012-08-21 1:41 ` Scott Wood [this message] 2012-08-21 1:41 ` Scott Wood 2012-08-15 1:25 ` Scott Wood 2012-08-15 1:25 ` Scott Wood 2012-08-14 23:04 ` [PATCH 18/38] KVM: PPC: E500: Implement MMU notifiers Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-15 1:20 ` Scott Wood 2012-08-15 1:20 ` Scott Wood 2012-08-15 9:38 ` Alexander Graf 2012-08-15 9:38 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 19/38] KVM: PPC: Add cache flush on page map Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-15 1:23 ` Scott Wood 2012-08-15 1:23 ` Scott Wood 2012-08-15 9:52 ` Alexander Graf 2012-08-15 9:52 ` Alexander Graf 2012-08-15 17:26 ` Scott Wood 2012-08-15 17:26 ` Scott Wood 2012-08-15 17:27 ` Alexander Graf 2012-08-15 17:27 ` Alexander Graf 2012-08-15 17:47 ` Scott Wood 2012-08-15 17:47 ` Scott Wood 2012-08-15 18:01 ` Alexander Graf 2012-08-15 18:01 ` Alexander Graf 2012-08-15 18:16 ` Scott Wood 2012-08-15 18:16 ` Scott Wood 2012-08-15 18:27 ` Alexander Graf 2012-08-15 18:27 ` Alexander Graf 2012-08-15 18:29 ` Alexander Graf 2012-08-15 18:29 ` Alexander Graf 2012-08-15 18:33 ` Scott Wood 2012-08-15 18:33 ` Scott Wood 2012-08-15 18:51 ` Alexander Graf 2012-08-15 18:51 ` Alexander Graf 2012-08-15 18:56 ` Scott Wood 2012-08-15 18:56 ` Scott Wood 2012-08-15 18:58 ` Alexander Graf 2012-08-15 18:58 ` Alexander Graf 2012-08-15 19:05 ` Scott Wood 2012-08-15 19:05 ` Scott Wood 2012-08-15 19:29 ` Alexander Graf 2012-08-15 19:29 ` Alexander Graf 2012-08-15 19:53 ` Scott Wood 2012-08-15 19:53 ` Scott Wood 2012-08-14 23:04 ` [PATCH 20/38] KVM: PPC: BookE: Add some more trace points Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 21/38] KVM: PPC: BookE: No duplicate request != 0 check Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 22/38] KVM: PPC: Use same kvmppc_prepare_to_enter code for booke and book3s_pr Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 23/38] KVM: PPC: Book3s: PR: Add (dumb) MMU Notifier support Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 24/38] KVM: PPC: BookE: Drop redundant vcpu->mode set Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 25/38] KVM: PPC: Book3S: PR: Only do resched check once per exit Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 26/38] KVM: PPC: Exit guest context while handling exit Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 27/38] KVM: PPC: Book3S: PR: Indicate we're out of guest mode Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 28/38] KVM: PPC: Consistentify vcpu exit path Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 29/38] KVM: PPC: Book3S: PR: Rework irq disabling Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-17 21:47 ` Benjamin Herrenschmidt 2012-08-17 21:47 ` Benjamin Herrenschmidt 2012-09-28 0:52 ` Alexander Graf 2012-09-28 0:52 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 30/38] KVM: PPC: Move kvm_guest_enter call into generic code Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 31/38] KVM: PPC: Ignore EXITING_GUEST_MODE mode Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 32/38] KVM: PPC: Add return value in prepare_to_enter Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 33/38] KVM: PPC: Add return value to core_check_requests Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 34/38] KVM: PPC: booke: Add watchdog emulation Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 35/38] booke: Added ONE_REG interface for IAC/DAC debug registers Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:44 ` Scott Wood 2012-08-14 23:44 ` Scott Wood 2012-08-14 23:47 ` Alexander Graf 2012-08-14 23:47 ` Alexander Graf 2012-08-15 0:06 ` Scott Wood 2012-08-15 0:06 ` Scott Wood 2012-08-14 23:04 ` [PATCH 36/38] KVM: PPC: 44x: Initialize PVR Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 37/38] KVM: PPC: BookE: Add MCSR SPR support Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-14 23:04 ` [PATCH 38/38] ppc: e500_tlb memset clears nothing Alexander Graf 2012-08-14 23:04 ` Alexander Graf 2012-08-15 10:07 ` Avi Kivity 2012-08-15 10:07 ` Avi Kivity 2012-08-15 10:09 ` Alexander Graf 2012-08-15 10:09 ` Alexander Graf 2012-08-15 10:10 ` Avi Kivity 2012-08-15 10:10 ` Avi Kivity
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=5032E74C.7010100@freescale.com \ --to=scottwood@freescale.com \ --cc=agraf@suse.de \ --cc=kvm-ppc@vger.kernel.org \ --cc=kvm@vger.kernel.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.