[PATCH] drm: lcdif: change burst size to 256B

[Marco Felsch]

On 22-07-28, Marco Felsch wrote:
> On 22-07-28, Liu Ying wrote:
> > On Wed, 2022-07-27 at 05:56 +0200, Marco Felsch wrote:
> > > Hi Marek, Liu,
> > > 
> > > On 22-07-26, Liu Ying wrote:
> > > > On Tue, 2022-07-26 at 16:19 +0200, Marek Vasut wrote:
> > > > > On 7/26/22 11:43, Marco Felsch wrote:
> > > > > > FIFO underruns are seen if a AXI bus master with a higher
> > > > > > priority
> > > > > > do a
> > > > > > lot of memory access. Increase the burst size to 256B to avoid
> > > > > > such
> > > > > > underruns and to improve the memory access efficiency.
> > > > > 
> > > > > Sigh, this again ...
> > > 
> > > I know.. we also tried the PANIC mode but this somehow didn't worked
> > > as
> > > documented. So this was the only way to reduce the underruns without
> > > adapting the interconnect prio for the hdmi-lcdif.
> > > 
> > > > > > diff --git a/drivers/gpu/drm/mxsfb/lcdif_kms.c
> > > > > > b/drivers/gpu/drm/mxsfb/lcdif_kms.c
> > > > > > index 1bec1279c8b5..1f22ea5896d5 100644
> > > > > > --- a/drivers/gpu/drm/mxsfb/lcdif_kms.c
> > > > > > +++ b/drivers/gpu/drm/mxsfb/lcdif_kms.c
> > > > > > @@ -143,8 +143,20 @@ static void lcdif_set_mode(struct
> > > > > > lcdif_drm_private *lcdif, u32 bus_flags)
> > > > > >   	       CTRLDESCL0_1_WIDTH(m->crtc_hdisplay),
> > > > > >   	       lcdif->base + LCDC_V8_CTRLDESCL0_1);
> > > > > >   
> > > > > > -	writel(CTRLDESCL0_3_PITCH(lcdif->crtc.primary->state-
> > > > > > >fb-
> > > > > > > pitches[0]),
> > > > > > 
> > > > > > -	       lcdif->base + LCDC_V8_CTRLDESCL0_3);
> > > > > > +	/*
> > > > > > +	 * Undocumented P_SIZE and T_SIZE bit fields but
> > > > > > according the
> > > > > > +	 * downstream kernel they control the AXI burst size.
> > > > > > As of now
> > > > > > there
> > > > 
> > > > I'm not sure if it is AXI burst size or any other burst size,
> > > > though it
> > > > seems to be AXI burst size.
> > > > 
> > > > Cc'ing Jian who mentioned 'burst size' and changed it from 128B to
> > > > 256B
> > > > in the downstream kernel.
> > > 
> > > Thanks.
> > 
> > Jian told me that it's AXI burst size.
> 
> Thanks for asking him. Do you know anything about the PANIC mode? We
> tested it by:
>  - using the interconnect patchsets [1]
>  - added a patch for configuring the hdmi-lcdif interconnect via DT [not
>    send upstream yet]
>  - setting the PANIC threshold to thresh-low:1/2 and tresh-high:3/4 and
>    enabled the INT_ENABLE_D1_PLANE_PANIC_EN within
>    LCDC_V8_INT_ENABLE_D1 (like the downstream kernel) [no send upstream
>    yet]
>  - configured the 'LCDIF_NOC_HURRY' to 0x7 (highest prio) like you do
>    within your downstream TF-A.
> 
> But this didn't worked for us and for Marek if I got him correctly. My
> question is: Is the PANIC mode working as documented or are there some
> missing bits?
> 
> You can test if the PANIC mode is working on the downstream kernel by
> reducing the AXI burst size back to 64B, connect a display with at least
> 1080P and do some heavy memory access. If panic mode is working you
> shouldn't see any display artifacts. I tested this before increasing the
> AXI burst size by setting the HDMI-LCDIF prio staticlly to a high prio
> like 0x5 and it was working with a 64B AXI burst size.

[1] https://lore.kernel.org/linux-arm-kernel/20220703091132.1412063-1-peng.fan@oss.nxp.com/
    https://lore.kernel.org/all/20220708085632.1918323-1-peng.fan@oss.nxp.com/