imx: ipu: fix occasional hang following disabling of channel
PATCH 1/1 imx: ipu: fix occasional hang following disabling of channel
It has been observed that some IMX6SDL SoC will hang appx 0.1% of boots between disabling the ipu channel and disabling DMA preceeding an OS boot unless an extra delay is added here.
The specific cause is still under investigation.
--- drivers/video/ipu_common.c | 1 + 1 file changed, 1 insertion(+)
drivers/video/ipu_common.c +1 -0
@@ -1109,6 +1109,7 @@ int32_t ipu_disable_channel(ipu_channel_t channel)if ((channel ==MEM_BG_SYNC ) || (channel ==MEM_FG_SYNC ) || (channel ==MEM_DC_SYNC )) {ipu_dp_dc_disable (channel,0 ); +mdelay (50 ); }/* Disable DMA channel(s) */
Signature · 1 line
1.9.1
PATCH 1/1 imx: ipu: fix occasional hang following disabling of channel
Hi Tim,
Does your code base contain this patch[1]?
It probably addressed the same issue you meet.
[1] http://lists.denx.de/pipermail/u-boot/2014-February/174533.html
Regards,
Liu Ying
Quoted 26 lines · Tim Harvey
It has been observed that some IMX6SDL SoC will hang appx 0.1% of boots between disabling the ipu channel and disabling DMA preceeding an OS boot unless an extra delay is added here.
The specific cause is still under investigation.
Signed-off-by Tim Harvey <[email protected]>--- drivers/video/ipu_common.c | 1 + 1 file changed, 1 insertion(+)drivers/video/ipu_common.c +1
-0@@ -1109,6 +1109,7 @@ int32_t ipu_disable_channel(ipu_channel_t channel)if ((channel ==MEM_BG_SYNC ) || (channel ==MEM_FG_SYNC ) || (channel ==MEM_DC_SYNC )) {ipu_dp_dc_disable (channel,0 ); +mdelay (50 ); }/* Disable DMA channel(s) */ Signature · 1 line
1.9.1
PATCH 1/1 imx: ipu: fix occasional hang following disabling of channel
Quoted 9 lines · Liu Ying
Hi Tim,
Does your code base contain this patch[1]?
It probably addressed the same issue you meet.[1] http://lists.denx.de/pipermail/u-boot/2014-February/174533.html
Regards,
Liu Ying
Liu,
It certainly does not - this issue was encountered back on 2014-04. I was hoping you might have some knowledge of this one!
I will re-test for this specific issue now that I am re-based on master.
Thanks,
Tim