From: "Tao Wu(吴涛@Eng)" <lepton@google.com>
To: "Marc-André Lureau" <marcandre.lureau@redhat.com>
Cc: qemu-devel@nongnu.org, kraxel@redhat.com, airlied@redhat.com,
liqiang6-s@360.cn
Subject: Re: [Qemu-devel] [PATCH] virtio-gpu: fix bug in host memory calculation.
Date: Thu, 9 Nov 2017 10:20:05 -0800 [thread overview]
Message-ID: <CAOvqHj5_aPibqECTZ+i0FTcCjg0r8N4fEMauZU8XmpBdvkSQzw@mail.gmail.com> (raw)
In-Reply-To: <152319465.38481678.1510223677165.JavaMail.zimbra@redhat.com>
Thanks. Sent out a new version add comments to say that we rely on
pixman create_bits to fail.
On Thu, Nov 9, 2017 at 2:34 AM, Marc-André Lureau
<marcandre.lureau@redhat.com> wrote:
> Hi
>
> ----- Original Message -----
>> The old code treats bits as bytes when calculating host memory usage.
>> Change it to be consistent with allocation logic in pixman library.
>>
>
> Good catch
>
>> Signed-off-by: Tao Wu <lepton@google.com>
>> ---
>> hw/display/virtio-gpu.c | 13 +++++++++++--
>> 1 file changed, 11 insertions(+), 2 deletions(-)
>>
>> diff --git a/hw/display/virtio-gpu.c b/hw/display/virtio-gpu.c
>> index 43bbe09ea0..428786f291 100644
>> --- a/hw/display/virtio-gpu.c
>> +++ b/hw/display/virtio-gpu.c
>> @@ -322,6 +322,15 @@ static pixman_format_code_t get_pixman_format(uint32_t
>> virtio_gpu_format)
>> }
>> }
>>
>> +static uint32_t calc_image_hostmem(pixman_format_code_t pformat,
>> + uint32_t width, uint32_t height)
>> +{
>> + /* copied from pixman/pixman-bits-image.c, skip integer overflow check.
>> */
>
> So we rely on pixman create_bits() to fail if overflow happened? perhaps it's worth a comment.
>
>> + int bpp = PIXMAN_FORMAT_BPP(pformat);
>> + int stride = ((width * bpp + 0x1f) >> 5) * sizeof(uint32_t);
>> + return height * stride;
>> +}
>> +
>> static void virtio_gpu_resource_create_2d(VirtIOGPU *g,
>> struct virtio_gpu_ctrl_command
>> *cmd)
>> {
>> @@ -366,7 +375,7 @@ static void virtio_gpu_resource_create_2d(VirtIOGPU *g,
>> return;
>> }
>>
>> - res->hostmem = PIXMAN_FORMAT_BPP(pformat) * c2d.width * c2d.height;
>> + res->hostmem = calc_image_hostmem(pformat, c2d.width, c2d.height);
>> if (res->hostmem + g->hostmem < g->conf.max_hostmem) {
>> res->image = pixman_image_create_bits(pformat,
>> c2d.width,
>> @@ -1087,7 +1096,7 @@ static int virtio_gpu_load(QEMUFile *f, void *opaque,
>> size_t size,
>> return -EINVAL;
>> }
>>
>> - res->hostmem = PIXMAN_FORMAT_BPP(pformat) * res->width *
>> res->height;
>> + res->hostmem = calc_image_hostmem(pformat, res->width, res->height);
>>
>> res->addrs = g_new(uint64_t, res->iov_cnt);
>> res->iov = g_new(struct iovec, res->iov_cnt);
>> --
>> 2.15.0.403.gc27cc4dac6-goog
>>
>
> looks good otherwise,
>
> Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
>
>>
prev parent reply other threads:[~2017-11-09 18:20 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-11-08 18:05 [Qemu-devel] [PATCH] virtio-gpu: fix bug in host memory calculation Tao Wu
2017-11-09 10:34 ` Marc-André Lureau
2017-11-09 18:17 ` Tao Wu
2017-11-09 18:19 ` Marc-André Lureau
2017-11-09 18:20 ` Tao Wu(吴涛@Eng) [this message]
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=CAOvqHj5_aPibqECTZ+i0FTcCjg0r8N4fEMauZU8XmpBdvkSQzw@mail.gmail.com \
--to=lepton@google.com \
--cc=airlied@redhat.com \
--cc=kraxel@redhat.com \
--cc=liqiang6-s@360.cn \
--cc=marcandre.lureau@redhat.com \
--cc=qemu-devel@nongnu.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).