From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
To: "David Gibson" <david@gibson.dropbear.id.au>,
"Cédric Le Goater" <clg@kaod.org>
Cc: linuxppc-dev@lists.ozlabs.org, kvm@vger.kernel.org,
kvm-ppc@vger.kernel.org
Subject: Re: [PATCH 06/19] KVM: PPC: Book3S HV: add a GET_ESB_FD control to the XIVE native device
Date: Mon, 11 Feb 2019 17:42:52 +1100 [thread overview]
Message-ID: <9dc53c6bccd6fd303481217f768ec173992431e2.camel@kernel.crashing.org> (raw)
In-Reply-To: <20190211023842.GE7230@umbus.fritz.box>
On Mon, 2019-02-11 at 13:38 +1100, David Gibson wrote:
>
> 1) All in kernel
>
> The offset always maps directly to guest irq number and the kernel
> somehow binds it either to an IPI or a host irq as necessary.
> Cédric's original code attempts this, but the mechanism of keeping a
> pointer to the VMA can't work.
Why do you need a pointer to the VMA anyway ? unmap_mapping_range()
doesn't need a VMA for the unmap part, and faults/mmaps have the VMA.
> But.. remapping the irqs should be sufficiently infrequent that it
> might be ok to consider simply stepping through all the hosting
> process's VMAs to do this.
Which unmap_mapping_range() does for you as I explained previously. You
only need the address space. See how spufs does it (among others).
> 2) Remapped in qemu (using memory regions)
>
> I _think_ (in hindsight) was Cédric's been discussing as the
> alternative in more recent posts.
>
> Qemu maps the IPI pages at one place and the passthrough IRQ pages
> somewhere else. The IPIs are mapped into the guest as one memory
> region, then any passthrough IRQ pages are mapped over that using
> overlapping memory regions.
>
> I don't think this approach will work well, because it could require a
> bunch of separate KVM memory slots, which are fairly scarce.
>
> 3) Remapped in qemu (using mmap())
>
> This is the approach I (and I think Paul) have been suggested in
> contrast to (1).
>
> Qemu maps the IPI pages and maps those into the guest. When we need
> to set up a passthrough IRQ, qemu mmap()s its pages directly over the
> IPI pages, and it remains mapped into the guest with the same memory
> region / memslot as the IPIs are already using. If the passthrough
> device is removed we have to remap the IPI pages back into place.
>
> 4) Dedicated irq numbers
>
> We never re-use regular guest irq numbers for passthrough irqs,
> instead we put them somewhere else and keep those mapped to the
> passthrough irq pages.
>
> I was favouring this approach, but it does mean there will be a guest
> visible difference between kernel_irqchip=on and off which isn't
> great.
>
>
> (1) is the most elegant _interface_, but as we've seen it's
> problematic to implement. Looking at the for_all_vmas() approach
> could be interesting, but otherwise option (3) might be the most
> practical.
>
> --
next prev parent reply other threads:[~2019-02-11 6:46 UTC|newest]
Thread overview: 135+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-01-07 18:43 [PATCH 00/19] KVM: PPC: Book3S HV: add XIVE native exploitation mode Cédric Le Goater
2019-01-07 18:43 ` [PATCH 01/19] powerpc/xive: export flags for the XIVE native exploitation mode hcalls Cédric Le Goater
2019-01-09 3:33 ` David Gibson
2019-01-09 13:08 ` Michael Ellerman
2019-01-09 13:38 ` Cédric Le Goater
2019-01-07 18:43 ` [PATCH 02/19] powerpc/xive: add OPAL extensions for the XIVE native exploitation support Cédric Le Goater
2019-01-09 4:26 ` David Gibson
2019-01-07 18:43 ` [PATCH 03/19] KVM: PPC: Book3S HV: check the IRQ controller type Cédric Le Goater
2019-01-09 4:27 ` David Gibson
2019-01-22 4:56 ` Paul Mackerras
2019-01-23 16:24 ` Cédric Le Goater
2019-02-04 0:50 ` David Gibson
2019-02-04 10:16 ` Cédric Le Goater
2019-01-07 18:43 ` [PATCH 04/19] KVM: PPC: Book3S HV: export services for the XIVE native exploitation device Cédric Le Goater
2019-01-11 4:09 ` David Gibson
2019-01-07 18:43 ` [PATCH 05/19] KVM: PPC: Book3S HV: add a new KVM device for the XIVE native exploitation mode Cédric Le Goater
2019-01-22 5:05 ` Paul Mackerras
2019-01-23 16:28 ` Cédric Le Goater
2019-01-28 17:35 ` Cédric Le Goater
2019-01-30 4:29 ` Paul Mackerras
2019-01-30 7:01 ` Cédric Le Goater
2019-01-31 3:01 ` Paul Mackerras
2019-02-01 17:03 ` Cédric Le Goater
2019-02-04 4:25 ` David Gibson
2019-02-04 11:19 ` Cédric Le Goater
2019-02-05 5:26 ` David Gibson
2019-01-07 18:43 ` [PATCH 06/19] KVM: PPC: Book3S HV: add a GET_ESB_FD control to the XIVE native device Cédric Le Goater
2019-01-22 5:09 ` Paul Mackerras
2019-01-23 16:48 ` Cédric Le Goater
2019-02-04 4:45 ` David Gibson
2019-02-04 11:30 ` Cédric Le Goater
2019-02-05 5:28 ` David Gibson
2019-02-05 12:55 ` Cédric Le Goater
2019-02-06 1:23 ` David Gibson
2019-02-06 7:21 ` Cédric Le Goater
2019-02-07 2:49 ` David Gibson
2019-02-07 9:03 ` Cédric Le Goater
2019-02-08 5:15 ` David Gibson
2019-02-08 7:58 ` Cédric Le Goater
2019-02-08 21:53 ` Paul Mackerras
2019-02-09 9:41 ` Cédric Le Goater
2019-02-11 2:38 ` David Gibson
2019-02-11 6:42 ` Benjamin Herrenschmidt [this message]
2019-02-12 22:07 ` Cédric Le Goater
2019-01-07 18:43 ` [PATCH 07/19] KVM: PPC: Book3S HV: add a GET_TIMA_FD control to " Cédric Le Goater
2019-01-07 18:43 ` [PATCH 08/19] KVM: PPC: Book3S HV: add a VC_BASE control to the " Cédric Le Goater
2019-01-22 5:14 ` Paul Mackerras
2019-01-23 16:56 ` Cédric Le Goater
2019-02-04 4:49 ` David Gibson
2019-02-04 15:36 ` Cédric Le Goater
2019-01-07 18:43 ` [PATCH 09/19] KVM: PPC: Book3S HV: add a SET_SOURCE " Cédric Le Goater
2019-02-04 4:57 ` David Gibson
2019-02-04 19:07 ` Cédric Le Goater
2019-02-05 5:35 ` David Gibson
2019-02-05 13:39 ` Cédric Le Goater
2019-01-07 18:43 ` [PATCH 10/19] KVM: PPC: Book3S HV: add a EISN attribute to kvmppc_xive_irq_state Cédric Le Goater
2019-01-07 18:43 ` [PATCH 11/19] KVM: PPC: Book3S HV: add support for the XIVE native exploitation mode hcalls Cédric Le Goater
2019-01-22 5:23 ` Paul Mackerras
2019-01-23 6:44 ` Benjamin Herrenschmidt
2019-01-23 8:48 ` Cédric Le Goater
2019-01-23 10:26 ` Paul Mackerras
2019-01-23 10:48 ` Cédric Le Goater
2019-01-23 21:23 ` Benjamin Herrenschmidt
2019-01-07 18:43 ` [PATCH 12/19] KVM: PPC: Book3S HV: record guest queue page address Cédric Le Goater
2019-02-04 5:15 ` David Gibson
2019-02-04 15:37 ` Cédric Le Goater
2019-01-07 18:43 ` [PATCH 13/19] KVM: PPC: Book3S HV: add a SYNC control for the XIVE native migration Cédric Le Goater
2019-02-04 5:17 ` David Gibson
2019-02-04 15:39 ` Cédric Le Goater
2019-01-07 18:43 ` [PATCH 14/19] KVM: PPC: Book3S HV: add a control to make the XIVE EQ pages dirty Cédric Le Goater
2019-02-04 5:18 ` David Gibson
2019-02-04 15:46 ` Cédric Le Goater
2019-02-05 5:30 ` David Gibson
2019-01-07 18:43 ` [PATCH 15/19] KVM: PPC: Book3S HV: add get/set accessors for the source configuration Cédric Le Goater
2019-02-04 5:21 ` David Gibson
2019-02-04 16:07 ` Cédric Le Goater
2019-02-05 5:32 ` David Gibson
2019-02-05 13:03 ` Cédric Le Goater
2019-02-06 1:23 ` David Gibson
2019-02-06 1:24 ` David Gibson
2019-02-06 7:07 ` Cédric Le Goater
2019-02-07 2:48 ` David Gibson
2019-02-07 9:13 ` Cédric Le Goater
2019-02-08 5:15 ` David Gibson
2019-02-14 16:50 ` Cédric Le Goater
2019-01-07 18:43 ` [PATCH 16/19] KVM: PPC: Book3S HV: add get/set accessors for the EQ configuration Cédric Le Goater
2019-02-04 5:24 ` David Gibson
2019-02-05 17:45 ` Cédric Le Goater
2019-01-07 19:10 ` [PATCH 17/19] KVM: PPC: Book3S HV: add get/set accessors for the VP XIVE state Cédric Le Goater
2019-01-07 19:10 ` [PATCH 18/19] KVM: PPC: Book3S HV: add passthrough support Cédric Le Goater
2019-01-22 5:26 ` Paul Mackerras
2019-01-23 6:45 ` Benjamin Herrenschmidt
2019-01-23 10:30 ` Paul Mackerras
2019-01-23 11:07 ` Cédric Le Goater
2019-01-28 6:13 ` Paul Mackerras
2019-01-28 18:26 ` Cédric Le Goater
2019-01-29 2:45 ` Paul Mackerras
2019-01-29 13:47 ` Cédric Le Goater
2019-01-30 6:20 ` Paul Mackerras
2019-01-30 15:54 ` Cédric Le Goater
2019-01-31 2:48 ` Paul Mackerras
2019-01-29 4:12 ` Paul Mackerras
2019-01-29 17:44 ` Cédric Le Goater
2019-01-30 5:55 ` Paul Mackerras
2019-01-30 7:06 ` Cédric Le Goater
2019-01-23 21:25 ` Benjamin Herrenschmidt
2019-01-24 8:41 ` Cédric Le Goater
2019-01-28 4:43 ` Paul Mackerras
2019-01-29 13:46 ` Cédric Le Goater
2019-01-07 19:10 ` [PATCH 19/19] KVM: introduce a KVM_DELETE_DEVICE ioctl Cédric Le Goater
2019-01-22 5:42 ` Paul Mackerras
2019-01-23 18:39 ` Cédric Le Goater
2019-01-23 21:32 ` Benjamin Herrenschmidt
2019-02-04 5:26 ` [PATCH 17/19] KVM: PPC: Book3S HV: add get/set accessors for the VP XIVE state David Gibson
2019-02-04 18:57 ` Cédric Le Goater
2019-02-05 5:33 ` David Gibson
2019-02-05 11:58 ` Cédric Le Goater
2019-02-06 1:19 ` David Gibson
2019-01-22 4:46 ` [PATCH 00/19] KVM: PPC: Book3S HV: add XIVE native exploitation mode Paul Mackerras
2019-01-23 19:07 ` Cédric Le Goater
2019-01-23 21:35 ` Benjamin Herrenschmidt
2019-01-26 8:25 ` Cédric Le Goater
2019-02-04 5:36 ` David Gibson
2019-02-05 11:31 ` Cédric Le Goater
2019-02-05 22:13 ` Paul Mackerras
2019-02-06 1:18 ` David Gibson
2019-02-06 7:35 ` Cédric Le Goater
2019-02-07 2:51 ` David Gibson
2019-02-07 8:31 ` Cédric Le Goater
2019-02-08 5:07 ` David Gibson
2019-02-08 7:38 ` Cédric Le Goater
2019-01-28 5:51 ` Paul Mackerras
2019-01-29 13:51 ` Cédric Le Goater
2019-01-30 5:40 ` Paul Mackerras
2019-01-30 15:36 ` Cédric Le Goater
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=9dc53c6bccd6fd303481217f768ec173992431e2.camel@kernel.crashing.org \
--to=benh@kernel.crashing.org \
--cc=clg@kaod.org \
--cc=david@gibson.dropbear.id.au \
--cc=kvm-ppc@vger.kernel.org \
--cc=kvm@vger.kernel.org \
--cc=linuxppc-dev@lists.ozlabs.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: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).