On 9/4/2019 4:43 PM, Thomas Zimmermann wrote:
Am 04.09.19 um 10:35 schrieb Feng Tang:
> Hi Daniel,
> On Wed, Sep 04, 2019 at 10:11:11AM +0200, Daniel Vetter wrote:
>> On Wed, Sep 4, 2019 at 8:53 AM Thomas Zimmermann <tzimmermann(a)suse.de>
>>> Am 04.09.19 um 08:27 schrieb Feng Tang:
>>>>> Thank you for testing. But don't get too excited, because the
>>>>> simulates a bug that was present in the original mgag200 code. A
>>>>> significant number of frames are simply skipped. That is apparently
>>>>> reason why it's faster.
>>>> Thanks for the detailed info, so the original code skips time-consuming
>>>> work inside atomic context on purpose. Is there any space to optmise it?
>>>> If 2 scheduled update worker are handled at almost same time, can one be
>>> To my knowledge, there's only one instance of the worker. Re-scheduling
>>> the worker before a previous instance started, will not create a second
>>> instance. The worker's instance will complete all pending updates. So in
>>> some way, skipping workers already happens.
>> So I think that the most often fbcon update from atomic context is the
>> blinking cursor. If you disable that one you should be back to the old
>> performance level I think, since just writing to dmesg is from process
>> context, so shouldn't change.
> Hmm, then for the old driver, it should also do the most update in
> non-atomic context?
> One other thing is, I profiled that updating a 3MB shadow buffer needs
> 20 ms, which transfer to 150 MB/s bandwidth. Could it be related with
> the cache setting of DRM shadow buffer? say the orginal code use a
> cachable buffer?
>> Bunch of tricks, but tbh I haven't tested them.
> Thomas has suggested to disable curson by
> echo 0 > /sys/devices/virtual/graphics/fbcon/cursor_blink
> We tried that way, and no change for the performance data.
There are several ways of disabling the cursor. On my test system, I entered
before the test and got better performance. Did you try this as well?
There's no obvious change on our system.
>> In any case, I still strongly advice you don't print anything to dmesg
>> or fbcon while benchmarking, because dmesg/printf are anything but
>> fast, especially if a gpu driver is involved. There's some efforts to
>> make the dmesg/printk side less painful (untangling the console_lock
>> from printk), but fundamentally printing to the gpu from the kernel
>> through dmesg/fbcon won't be cheap. It's just not something we
>> optimize beyond "make sure it works for emergencies".
>>> Best regards
>>>>> Best regards
>>>> dri-devel mailing list
>>> Thomas Zimmermann
>>> Graphics Driver Developer
>>> SUSE Linux GmbH, Maxfeldstrasse 5, 90409 Nuernberg, Germany
>>> GF: Felix Imendörffer, Mary Higgins, Sri Rasiah
>>> HRB 21284 (AG Nürnberg)
>>> dri-devel mailing list
>> Daniel Vetter
>> Software Engineer, Intel Corporation
>> +41 (0) 79 365 57 48 - http://blog.ffwll.ch
LKP mailing list