qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Steven Sistare <steven.sistare@oracle.com>
To: Zheng Chuan <zhengchuan@huawei.com>,
	Alex Williamson <alex.williamson@redhat.com>
Cc: "Daniel P. Berrange" <berrange@redhat.com>,
	"Michael S. Tsirkin" <mst@redhat.com>,
	"Jason Zeng" <jason.zeng@linux.intel.com>,
	"Alex Bennée" <alex.bennee@linaro.org>,
	"Juan Quintela" <quintela@redhat.com>,
	qemu-devel@nongnu.org, "Eric Blake" <eblake@redhat.com>,
	"Dr. David Alan Gilbert" <dgilbert@redhat.com>,
	Xiexiangyou <xiexiangyou@huawei.com>,
	"Stefan Hajnoczi" <stefanha@redhat.com>,
	"Marc-André Lureau" <marcandre.lureau@redhat.com>,
	"Paolo Bonzini" <pbonzini@redhat.com>,
	"Philippe Mathieu-Daudé" <philmd@redhat.com>,
	"Markus Armbruster" <armbru@redhat.com>
Subject: Re: [PATCH V6 19/27] vfio-pci: cpr part 1 (fd and dma)
Date: Tue, 30 Nov 2021 11:11:55 -0500	[thread overview]
Message-ID: <59e9f793-667d-5fe1-bccf-9afef791a00e@oracle.com> (raw)
In-Reply-To: <803a8bee-ec82-a41b-38ad-57d7ef572019@huawei.com>

On 11/10/2021 2:48 AM, Zheng Chuan wrote:
> 
> Hi, steve
> 
> On 2021/8/11 1:06, Alex Williamson wrote:
>> On Fri,  6 Aug 2021 14:43:53 -0700
>> Steve Sistare <steven.sistare@oracle.com> wrote:
>> [...]
>>> +static int
>>> +vfio_region_remap(MemoryRegionSection *section, void *handle, Error **errp)
>>> +{
>>> +    MemoryRegion *mr = section->mr;
>>> +    VFIOContainer *container = handle;
>>> +    const char *name = memory_region_name(mr);
>>> +    ram_addr_t size = int128_get64(section->size);
>>> +    hwaddr offset, iova, roundup;
>>> +    void *vaddr;
>>> +
>>> +    if (vfio_listener_skipped_section(section) || memory_region_is_iommu(mr)) {
>>> +        return 0;
>>> +    }
>>> +
>>> +    offset = section->offset_within_address_space;
>>> +    iova = REAL_HOST_PAGE_ALIGN(offset);
> We should not do remap if it shares on host page with other structures.
> I think a judgement like int128_ge((int128_make64(iova), llend)) in vfio_listener_region_add() should be also added here to check it,
> otherwise it will remap no-exit dma which causes the live update failure.
> diff --git a/hw/vfio/cpr.c b/hw/vfio/cpr.c
> index 0981d31..d231841 100644
> --- a/hw/vfio/cpr.c
> +++ b/hw/vfio/cpr.c
> @@ -58,13 +58,21 @@ vfio_region_remap(MemoryRegionSection *section, void *handle, Error **errp)
>      ram_addr_t size = int128_get64(section->size);
>      hwaddr offset, iova, roundup;
>      void *vaddr;
> -
> +    Int128 llend;
> +
>      if (vfio_listener_skipped_section(section) || memory_region_is_iommu(mr)) {
>          return 0;
>      }
> 
>      offset = section->offset_within_address_space;
>      iova = REAL_HOST_PAGE_ALIGN(offset);
> +    llend = int128_make64(section->offset_within_address_space);
> +    llend = int128_add(llend, section->size);
> +    llend = int128_and(llend, int128_exts64(qemu_real_host_page_mask));
> +    if (int128_ge(int128_make64(iova), llend)) {
> +        return 0;
> +    }
> +
>      roundup = iova - offset;
>      size -= roundup;
>      size = REAL_HOST_PAGE_ALIGN(size);
> 
>>> +    roundup = iova - offset;
>>> +    size -= roundup;
>>> +    size = REAL_HOST_PAGE_ALIGN(size);
>>> +    vaddr = memory_region_get_ram_ptr(mr) +
>>> +            section->offset_within_region + roundup;
>>> +
>>> +    trace_vfio_region_remap(name, container->fd, iova, iova + size - 1, vaddr);
>>> +    return vfio_dma_map_vaddr(container, iova, size, vaddr, errp);
>>> +}

Thank you Zheng.  I intended to implement the logic you suggest, using 64-bit arithmetic,
but I botched it.  This should do the trick:

diff --git a/hw/vfio/cpr.c b/hw/vfio/cpr.c
index df334d9..bbdeaea 100644
--- a/hw/vfio/cpr.c
+++ b/hw/vfio/cpr.c
@@ -66,8 +66,8 @@ vfio_region_remap(MemoryRegionSection *section, void *handle,
     offset = section->offset_within_address_space;
     iova = REAL_HOST_PAGE_ALIGN(offset);
     roundup = iova - offset;
-    size -= roundup;
-    size = REAL_HOST_PAGE_ALIGN(size);
+    size -= roundup;                    /* adjust for starting alignment */
+    size &= qemu_real_host_page_mask;   /* adjust for ending alignment */
     end = iova + size;
     if (iova >= end) {
         return 0;

- Steve


  reply	other threads:[~2021-11-30 16:13 UTC|newest]

Thread overview: 44+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-08-06 21:43 [PATCH V6 00/27] Live Update Steve Sistare
2021-08-06 21:43 ` [PATCH V6 01/27] memory: qemu_check_ram_volatile Steve Sistare
2021-08-06 21:43 ` [PATCH V6 02/27] migration: fix populate_vfio_info Steve Sistare
2021-08-06 21:43 ` [PATCH V6 03/27] migration: qemu file wrappers Steve Sistare
2021-08-06 21:43 ` [PATCH V6 04/27] migration: simplify savevm Steve Sistare
2021-08-06 21:43 ` [PATCH V6 05/27] vl: start on wakeup request Steve Sistare
2021-08-06 21:43 ` [PATCH V6 06/27] cpr: reboot mode Steve Sistare
2021-08-06 21:43 ` [PATCH V6 07/27] cpr: reboot HMP interfaces Steve Sistare
2021-08-06 21:43 ` [PATCH V6 08/27] memory: flat section iterator Steve Sistare
2021-08-06 21:43 ` [PATCH V6 09/27] oslib: qemu_clear_cloexec Steve Sistare
2021-08-06 21:43 ` [PATCH V6 10/27] machine: memfd-alloc option Steve Sistare
2021-08-06 21:43 ` [PATCH V6 11/27] qapi: list utility functions Steve Sistare
2021-08-06 21:43 ` [PATCH V6 12/27] vl: helper to request re-exec Steve Sistare
2021-08-06 21:43 ` [PATCH V6 13/27] cpr: preserve extra state Steve Sistare
2021-08-06 21:43 ` [PATCH V6 14/27] cpr: restart mode Steve Sistare
2021-08-06 21:43 ` [PATCH V6 15/27] cpr: restart HMP interfaces Steve Sistare
2021-08-06 21:43 ` [PATCH V6 16/27] hostmem-memfd: cpr for memory-backend-memfd Steve Sistare
2021-08-06 21:43 ` [PATCH V6 17/27] pci: export functions for cpr Steve Sistare
2021-08-06 21:43 ` [PATCH V6 18/27] vfio-pci: refactor " Steve Sistare
2021-08-10 16:53   ` Alex Williamson
2021-08-23 16:52     ` Steven Sistare
2021-08-06 21:43 ` [PATCH V6 19/27] vfio-pci: cpr part 1 (fd and dma) Steve Sistare
2021-08-10 17:06   ` Alex Williamson
2021-08-23 19:43     ` Steven Sistare
2021-11-10  7:48     ` Zheng Chuan
2021-11-30 16:11       ` Steven Sistare [this message]
2021-08-06 21:43 ` [PATCH V6 20/27] vfio-pci: cpr part 2 (msi) Steve Sistare
2021-08-06 21:43 ` [PATCH V6 21/27] vfio-pci: cpr part 3 (intx) Steve Sistare
2022-03-29 11:03   ` Fam Zheng
2022-04-11 16:23     ` Steven Sistare
2022-04-12 11:01       ` Fam Zheng
2021-08-06 21:43 ` [PATCH V6 22/27] vhost: reset vhost devices for cpr Steve Sistare
2021-08-06 21:43 ` [PATCH V6 23/27] chardev: cpr framework Steve Sistare
2021-08-06 21:43 ` [PATCH V6 24/27] chardev: cpr for simple devices Steve Sistare
2021-08-06 21:43 ` [PATCH V6 25/27] chardev: cpr for pty Steve Sistare
2021-08-06 21:44 ` [PATCH V6 26/27] chardev: cpr for sockets Steve Sistare
2021-08-06 21:44 ` [PATCH V6 27/27] cpr: only-cpr-capable option Steve Sistare
2021-08-09 16:02 ` [PATCH V6 00/27] Live Update Steven Sistare
2021-08-21  8:54 ` Zheng Chuan
2021-08-23 21:36   ` Steven Sistare
2021-08-24  9:36     ` Zheng Chuan
2021-08-31 21:15       ` Steven Sistare
2021-10-27  6:16         ` Zheng Chuan
2021-10-27 12:25           ` Steven Sistare

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=59e9f793-667d-5fe1-bccf-9afef791a00e@oracle.com \
    --to=steven.sistare@oracle.com \
    --cc=alex.bennee@linaro.org \
    --cc=alex.williamson@redhat.com \
    --cc=armbru@redhat.com \
    --cc=berrange@redhat.com \
    --cc=dgilbert@redhat.com \
    --cc=eblake@redhat.com \
    --cc=jason.zeng@linux.intel.com \
    --cc=marcandre.lureau@redhat.com \
    --cc=mst@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=philmd@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=quintela@redhat.com \
    --cc=stefanha@redhat.com \
    --cc=xiexiangyou@huawei.com \
    --cc=zhengchuan@huawei.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).