[Qemu-devel] [PATCH] QXL: always show default cursor, even for unsupported chunks

[Qemu-devel] [PATCH] QXL: always show default cursor, even for unsupported chunks

[René Rebe]

Hi,

I was testing a Windows 10 guest with HiDPI scaling which

actually made the cursor disappear and only multiple chunks

warnings being printed.

When the cursor is using currently unsupported multiple chunks, the
builtin-left cursor is better than none.

PS: if TB damages the white-space just let me know, only recently 
started using it, ...

   - René Rebe <rene@exactcode.de>

diff -u -ur qemu-2.10.0-rc3/hw/display/qxl-render.c 
qemu-2.10.0-rc3-fixed/hw/display/qxl-render.c
--- qemu-2.10.0-rc3/hw/display/qxl-render.c    2017-08-15 
20:39:41.000000000 +0200
+++ qemu-2.10.0-rc3-fixed/hw/display/qxl-render.c    2017-08-17 
13:42:37.108953910 +0200
@@ -270,9 +270,10 @@
          }
          if (cursor->chunk.data_size != cursor->data_size) {
              fprintf(stderr, "%s: multiple chunks\n", __FUNCTION__);
-            return 1;
+            c = NULL; /* fall back to built-in left cursor */
+        } else {
+            c = qxl_cursor(qxl, cursor);
          }
-        c = qxl_cursor(qxl, cursor);
          if (c == NULL) {
              c = cursor_builtin_left_ptr();
          }

Re: [Qemu-devel] [PATCH] QXL: always show default cursor, even for unsupported chunks

[Dr. David Alan Gilbert]

cc'ing in Gerd

* René Rebe (rene@exactcode.com) wrote:
> Hi,
> 
> I was testing a Windows 10 guest with HiDPI scaling which
> 
> actually made the cursor disappear and only multiple chunks
> 
> warnings being printed.
> 
> When the cursor is using currently unsupported multiple chunks, the
> builtin-left cursor is better than none.
> 
> PS: if TB damages the white-space just let me know, only recently started
> using it, ...
> 
>   - René Rebe <rene@exactcode.de>
> 
> diff -u -ur qemu-2.10.0-rc3/hw/display/qxl-render.c
> qemu-2.10.0-rc3-fixed/hw/display/qxl-render.c
> --- qemu-2.10.0-rc3/hw/display/qxl-render.c    2017-08-15 20:39:41.000000000
> +0200
> +++ qemu-2.10.0-rc3-fixed/hw/display/qxl-render.c    2017-08-17
> 13:42:37.108953910 +0200
> @@ -270,9 +270,10 @@
>          }
>          if (cursor->chunk.data_size != cursor->data_size) {
>              fprintf(stderr, "%s: multiple chunks\n", __FUNCTION__);
> -            return 1;
> +            c = NULL; /* fall back to built-in left cursor */
> +        } else {
> +            c = qxl_cursor(qxl, cursor);
>          }
> -        c = qxl_cursor(qxl, cursor);
>          if (c == NULL) {
>              c = cursor_builtin_left_ptr();
>          }
> 
> 
--
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK

Re: [Qemu-devel] [PATCH] QXL: always show default cursor, even for unsupported chunks

[Eric Blake]

[-- Attachment #1: Type: text/plain, Size: 1588 bytes --]

On 08/18/2017 09:37 AM, Dr. David Alan Gilbert wrote:
> cc'ing in Gerd
> 
> * René Rebe (rene@exactcode.com) wrote:
>> Hi,
>>
>> I was testing a Windows 10 guest with HiDPI scaling which
>>
>> actually made the cursor disappear and only multiple chunks
>>
>> warnings being printed.
>>
>> When the cursor is using currently unsupported multiple chunks, the
>> builtin-left cursor is better than none.

This part is okay, but might be better written in an imperative sense
(the best commit messages are ones where you add an implicit "Apply this
patch to ..." prior to the first word after the blank line separating
the subject).

>>
>> PS: if TB damages the white-space just let me know, only recently started
>> using it, ...

This part belongs after a --- separator, as it is useful to reviewers
but adds nothing to git.

By the way, while your patch did not appear to be whitespace damaged on
a cursory glance, the fact that you used thunderbird to send it is
risky; thunderbird HAS been known to mess up patches that are pasted
inline, and sending patches as attachments is harder to read than direct
patches.  Using 'git send-email' excels at making sure the patch will be
in a format easy to apply.

>>
>>   - René Rebe <rene@exactcode.de>

And this is not a proper Signed-off-by: tag; we'll need one of those
before your patch can be considered.  More hints at:
http://wiki.qemu.org/Contribute/SubmitAPatch

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3266
Virtualization:  qemu.org | libvirt.org


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 619 bytes --]

Re: [Qemu-devel] [PATCH] QXL: always show default cursor, even for unsupported chunks

[Gerd Hoffmann]

On Fri, 2017-08-18 at 15:37 +0100, Dr. David Alan Gilbert wrote:
> cc'ing in Gerd
> 
> * René Rebe (rene@exactcode.com) wrote:
> > Hi,
> > 
> > I was testing a Windows 10 guest with HiDPI scaling which
> > 
> > actually made the cursor disappear and only multiple chunks
> > 
> > warnings being printed.
> > 
> > When the cursor is using currently unsupported multiple chunks, the
> > builtin-left cursor is better than none.

Better: add support for chunks.  Can you try

https://www.kraxel.org/cgit/qemu/log/?h=work/qxl-chunked-cursor

?

cheers,
  Gerd

Re: [Qemu-devel] [PATCH] QXL: always show default cursor, even for unsupported chunks

[René Rebe]

Hi,

On 08/23/2017 04:36 PM, Gerd Hoffmann wrote:
> On Fri, 2017-08-18 at 15:37 +0100, Dr. David Alan Gilbert wrote:
>> cc'ing in Gerd
>>
>> * René Rebe (rene@exactcode.com) wrote:
>>> Hi,
>>>
>>> I was testing a Windows 10 guest with HiDPI scaling which
>>>
>>> actually made the cursor disappear and only multiple chunks
>>>
>>> warnings being printed.
>>>
>>> When the cursor is using currently unsupported multiple chunks, the
>>> builtin-left cursor is better than none.
> 
> Better: add support for chunks.  Can you try
> 
> https://www.kraxel.org/cgit/qemu/log/?h=work/qxl-chunked-cursor
> 
> ?
> 
> cheers,
>    Gerd
> 

I applied that change on top of the current -rc3 tarball, and it works 
beautifully so far.

Thanks!

	René
-- 
   René Rebe, ExactCODE GmbH, Lietzenburger Str. 42, DE-10117 Berlin
   http://exactcode.com | http://t2-project.org | http://rene.rebe.de