svn commit: r564088 - in head/x11: linux-nvidia-libs nvidia-driver nvidia-driver/files nvidia-settings nvidia-xconfig

[Craig Leres]

On 2/7/21 10:16 PM, Craig Leres wrote:
> 
> On 2/7/21 5:38 PM, Alexey Dokuchaev wrote:
>> On Sun, Feb 07, 2021 at 01:04:15PM -0800, Craig Leres wrote:
>>> On 2/6/21 12:57 AM, Alexey Dokuchaev wrote:
>>>> Thanks for the report, should be fixed in r564208 (or so I hope, I 
>>>> could
>>>> not reproduce it, but there is upstream issue which suggested similar,
>>>> albeit not completely identical fix*).
>>>>
>>>> *)https://github.com/NVIDIA/nvidia-settings/issues/59
>>> My poudriere build server failed to build this version this morning.
>>> I'll append the specific errors, let me know if you'd like to see the
>>> whole log.
>> No, it's fine.  Could you apply suggested patch with explicit $(MKDIR)
>> from that issue #59 and rerun your tests?  Now that I've studied their
>> makefiles closer, they actually do call it explicitly as part of the
>> recipe for other similar targets. :(
> 
> I've appended a patchset that does this. It builds in poudriere by 
> itself, tomorrow I'll report how it did when part of a ~1000 port build 
> (that when it's been failing).

The patched version of nvidia-settings built ok this morning (along with 
1091 other packages). So I think the patchset is good.

However, ever since booting with nvidia-driver-460.39 I've been seeing 
intermittent xterm glitches related to running firefox and kicad. It 
looks like some kind of hardware buffer swapping... Like something is 
switching between slightly different versions of a window, sometimes 
once every 5 seconds and sometimes very fast. And sometimes xterm 
scrolling results in the line doubling that doesn't really exist.

		Craig