* Thomas Zimmermann [240220 10:42]:
> The changes below look good. You can test by instrumenting
> drm_fb_helper_deferred_io() with printk(). For testing, you can reduce the
> write-back frequency by setting helper->fbdefio.delay to a higher value. If
> you set it to HZ, it should only do a write-
Hi
Am 20.02.24 um 11:16 schrieb Tony Lindgren:
* Thomas Zimmermann [240220 09:32]:
Am 20.02.24 um 09:56 schrieb Tony Lindgren:
Oh right, yes omapdrm is operating on memory.
With the latest kernels, you should see a warning if helpers operate on the
wrong type of memory. [1][2]
OK
[1]
htt
* Thomas Zimmermann [240220 09:32]:
> Am 20.02.24 um 09:56 schrieb Tony Lindgren:
> > Oh right, yes omapdrm is operating on memory.
>
> With the latest kernels, you should see a warning if helpers operate on the
> wrong type of memory. [1][2]
OK
> [1]
> https://elixir.bootlin.com/linux/v6.8-rc
Hi
Am 20.02.24 um 09:56 schrieb Tony Lindgren:
* Thomas Zimmermann [240219 16:43]:
Am 19.02.24 um 15:19 schrieb Tony Lindgren:
--- a/drivers/gpu/drm/omapdrm/Kconfig
+++ b/drivers/gpu/drm/omapdrm/Kconfig
@@ -5,6 +5,7 @@ config DRM_OMAP
depends on ARCH_OMAP2PLUS
select DRM_KMS_H
* Thomas Zimmermann [240219 16:43]:
> Am 19.02.24 um 15:19 schrieb Tony Lindgren:
> > --- a/drivers/gpu/drm/omapdrm/Kconfig
> > +++ b/drivers/gpu/drm/omapdrm/Kconfig
> > @@ -5,6 +5,7 @@ config DRM_OMAP
> > depends on ARCH_OMAP2PLUS
> > select DRM_KMS_HELPER
> > select FB_DMAMEM_HELPERS
Hi
Am 19.02.24 um 15:19 schrieb Tony Lindgren:
Commit 95da53d63dcf ("drm/omapdrm: Use regular fbdev I/O helpers")
broke console because there is no damage handling in fb_sys_write()
unlike we have in drm_fb_helper_sys_write().
Let's fix the issue by using deferred ops with fb helpers for damage
Commit 95da53d63dcf ("drm/omapdrm: Use regular fbdev I/O helpers")
broke console because there is no damage handling in fb_sys_write()
unlike we have in drm_fb_helper_sys_write().
Let's fix the issue by using deferred ops with fb helpers for damage.
Fixes: 95da53d63dcf ("drm/omapdrm: Use regular
