qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Avi Kivity <avi@redhat.com>
To: liu ping fan <qemulist@gmail.com>
Cc: Jan Kiszka <jan.kiszka@siemens.com>,
	Marcelo Tosatti <mtosatti@redhat.com>,
	"qemu-devel@nongnu.org" <qemu-devel@nongnu.org>,
	Anthony Liguori <anthony@codemonkey.ws>,
	Paolo Bonzini <pbonzini@redhat.com>
Subject: Re: [Qemu-devel] [big lock] Discussion about the convention of device's DMA each other after breaking down biglock
Date: Mon, 24 Sep 2012 09:44:59 +0200	[thread overview]
Message-ID: <50600F7B.5080106@redhat.com> (raw)
In-Reply-To: <CAJnKYQmsiXPw=BknEZZb2+eEyKoKW5dakwv5TUVGR3mHLc-+VA@mail.gmail.com>

On 09/24/2012 08:33 AM, liu ping fan wrote:
> On Wed, Sep 19, 2012 at 5:50 PM, Avi Kivity <avi@redhat.com> wrote:
> > On 09/19/2012 12:34 PM, Jan Kiszka wrote:
> >>
> >> What about the following:
> >>
> >> What we really need to support in practice is MMIO access triggers RAM
> >> access of device model. Scenarios where a device access triggers another
> >> MMIO access could likely just be rejected without causing troubles.
> >>
> >> So, when we dispatch a request to a device, we mark that the current
> >> thread is in a MMIO dispatch and reject any follow-up c_p_m_rw that does
> >> _not_ target RAM, ie. is another, nested MMIO request - independent of
> >> its destination. How much of the known issues would this solve? And what
> >> would remain open?
> >
> > Various iommu-like devices re-dispatch I/O, like changing endianness or
> > bitband.  I don't know whether it targets I/O rather than RAM.
> >
> Have not found the exact code. But I think the call chain may look
> like this: dev mmio-handler --> c_p_m_rw() --> iommu mmio-handler -->
> c_p_m_rw()
> And I think you worry about the case for "c_p_m_rw() --> iommu
> mmio-handler". Right? How about introduce an member can_nest for
> MemoryRegionOps of iommu's mr?
>

I would rather push the iommu logic into the memory API:

  memory_region_init_iommu(MemoryRegion *mr, const char *name,
                           MemoryRegion *target, MemoryRegionIOMMUOps *ops,
                           unsigned size)

  struct MemoryRegionIOMMUOps {
      target_physical_addr_t (*translate)(target_physical_addr_t addr,
bool write);
      void (*fault)(target_physical_addr_t addr);
  };

I'll look at a proposal for this.  It's a generalized case of
memory_region_init_alias().

-- 
I have a truly marvellous patch that fixes the bug which this
signature is too narrow to contain.

  reply	other threads:[~2012-09-24  7:45 UTC|newest]

Thread overview: 45+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-09-19  3:02 [Qemu-devel] [big lock] Discussion about the convention of device's DMA each other after breaking down biglock liu ping fan
2012-09-19  8:06 ` Avi Kivity
2012-09-19  9:00   ` liu ping fan
2012-09-19  9:07     ` Avi Kivity
2012-09-19  9:11       ` liu ping fan
2012-09-19  9:14         ` Paolo Bonzini
2012-09-19  9:19           ` liu ping fan
2012-09-19  9:23             ` Avi Kivity
2012-09-19  9:27               ` Jan Kiszka
2012-09-19  9:28                 ` Jan Kiszka
2012-09-20  7:51               ` liu ping fan
2012-09-20  7:54                 ` Paolo Bonzini
2012-09-20  8:09                   ` liu ping fan
2012-09-20  8:27                     ` Paolo Bonzini
2012-09-20  9:07                 ` Avi Kivity
2012-09-21  7:27                   ` liu ping fan
2012-09-21  8:21                     ` Paolo Bonzini
2012-09-19  9:21           ` Avi Kivity
2012-09-19  9:51             ` Paolo Bonzini
2012-09-19 10:06               ` Avi Kivity
2012-09-19 10:19                 ` Paolo Bonzini
2012-09-19 10:27                   ` Avi Kivity
2012-09-19  9:34     ` Jan Kiszka
2012-09-19  9:50       ` Avi Kivity
2012-09-19 10:18         ` Jan Kiszka
2012-09-24  6:33         ` liu ping fan
2012-09-24  7:44           ` Avi Kivity [this message]
2012-09-24  8:32             ` liu ping fan
2012-09-24  9:42               ` Avi Kivity
2012-09-27  3:13                 ` liu ping fan
2012-09-27  9:16                   ` Avi Kivity
2012-09-27  9:29                     ` Paolo Bonzini
2012-09-27  9:34                       ` Avi Kivity
2012-09-27  9:36                         ` Paolo Bonzini
2012-09-27 10:08                           ` Avi Kivity
2012-09-27 10:22                             ` Paolo Bonzini
2012-09-27 10:48                               ` Avi Kivity
2012-09-29  9:20                     ` liu ping fan
2012-09-30  8:13                       ` Avi Kivity
2012-09-30  8:48                         ` liu ping fan
2012-09-30 11:18                           ` Avi Kivity
2012-09-30 11:04                         ` Blue Swirl
2012-09-30 11:17                           ` Avi Kivity
2012-09-30 11:48                             ` Blue Swirl
2012-09-20  8:11       ` liu ping fan

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=50600F7B.5080106@redhat.com \
    --to=avi@redhat.com \
    --cc=anthony@codemonkey.ws \
    --cc=jan.kiszka@siemens.com \
    --cc=mtosatti@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=qemulist@gmail.com \
    /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).