On 08/26/2016 04:38 PM, Lucas Stach wrote:
> Am Freitag, den 26.08.2016, 16:30 +0200 schrieb Marek Vasut:
>> Allocate the framebuffer memory as coherent, otherwise the framebuffer
>> will suffer from artifacts when displaying scrolling text or video.
>> This can be replicated on i.MX6SX (armv7), which has more complex memory
>> architecture compared to the i.MX23/28 (armv5).
>>
>> Signed-off-by: Marek Vasut <marex@xxxxxxx>
>> Cc: Tomi Valkeinen <tomi.valkeinen@xxxxxx>
>> Cc: Fabio Estevam <fabio.estevam@xxxxxxx>
>> Cc: Lucas Stach <l.stach@xxxxxxxxxxxxxx>
>> Cc: Shawn Guo <shawnguo@xxxxxxxxxx>
>> ---
>> drivers/video/fbdev/mxsfb.c | 10 ++++++----
>> 1 file changed, 6 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/video/fbdev/mxsfb.c b/drivers/video/fbdev/mxsfb.c
>> index 4e6608c..bd1a310 100644
>> --- a/drivers/video/fbdev/mxsfb.c
>> +++ b/drivers/video/fbdev/mxsfb.c
>> @@ -800,6 +800,7 @@ static int mxsfb_init_fbinfo(struct mxsfb_info *host,
>> struct fb_videomode *vmode)
>> {
>> int ret;
>> + struct device *dev = &host->pdev->dev;
>> struct fb_info *fb_info = &host->fb_info;
>> struct fb_var_screeninfo *var = &fb_info->var;
>> dma_addr_t fb_phys;
>> @@ -825,12 +826,11 @@ static int mxsfb_init_fbinfo(struct mxsfb_info *host,
>>
>> /* Memory allocation for framebuffer */
>> fb_size = SZ_2M;
>> - fb_virt = alloc_pages_exact(fb_size, GFP_DMA);
>> + fb_virt = dma_alloc_coherent(dev, PAGE_ALIGN(fb_size), &fb_phys,
>> + GFP_KERNEL);
>
> You probably want writecombined memory here instead of coherent. Using
> coherent memory for framebuffers sucks performance wise.
Ah, true. That'd be dma_{alloc,free}_wc() , right ?
--
Best regards,
Marek Vasut
--
To unsubscribe from this list: send the line "unsubscribe linux-fbdev" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
