takeshi has quit [Quit: Connection closed for inactivity]
_whitelogger has joined #linux-rockchip
lkcl has quit [Ping timeout: 260 seconds]
lkcl has joined #linux-rockchip
matthias_bgg has joined #linux-rockchip
lurchi_ is now known as lurchi__
matthias_bgg has quit [Quit: Leaving]
lurchi__ is now known as lurchi_
lurchi_ has quit [Ping timeout: 240 seconds]
BenG83 has joined #linux-rockchip
Aussie_matt has joined #linux-rockchip
aalm has quit [Ping timeout: 260 seconds]
aalm has joined #linux-rockchip
descend-irc has joined #linux-rockchip
phinxy has joined #linux-rockchip
phinxy has left #linux-rockchip [#linux-rockchip]
lurchi_ has joined #linux-rockchip
gnufan has joined #linux-rockchip
cnxsoft has quit [Quit: cnxsoft]
ted_ has joined #linux-rockchip
ted_ has quit [Read error: Connection reset by peer]
lurchi_ has quit [Ping timeout: 240 seconds]
lkcl has quit [Ping timeout: 248 seconds]
Aussie_matt has quit [Read error: Connection reset by peer]
phinxy has joined #linux-rockchip
phinxy has quit [Read error: Connection reset by peer]
<LongChair>
ayaka : how is one supposed to detect if a codec is supported with mpp ? seems that trying to use like vp9 on 3288 leads to pretty bad crash. is there a proper way to detect that ?
<LongChair>
the only thing i have found is a function that relies on build flags, which is not depeding,on what hw supports
descend-irc has quit [Quit: Connection closed for inactivity]
lurchi_ has joined #linux-rockchip
lurchi_ is now known as lurchi__
lurchi__ is now known as lurchi_
lurchi_ is now known as lurchi__
BenG83 has quit [Quit: Leaving]
lurchi__ is now known as lurchi_
lurchi_ is now known as lurchi__
lurchi__ is now known as lurchi_
lurchi_ is now known as lurchi__
cyteen has quit [Ping timeout: 240 seconds]
cyteen has joined #linux-rockchip
lurchi__ is now known as lurchi_
lurchi_ is now known as lurchi__
lurchi__ is now known as lurchi_
Myy has joined #linux-rockchip
<Myy>
Miaou
<Myy>
I've been able to repair the rockchip_forwardports VPU driver for the 4.13 kernel, when using the IOMMU_DMA API
<Myy>
Turns out that trying the API one more time, getting rid of the #ifdef to be sure that the API was used clearly showed that the issue with IOVA calls was that the iova_cache was not set up, causing NPE in the IOVA code.
<Myy>
The function in the IOMMU DMA API that should take care of that is called "iommu_dma_init"
<Myy>
But this function is not exported
<Myy>
So, I patched the kernel to export this function, added the "iommu_dma_init()" call before the iommu_dma_init_domain(...)" call
<Myy>
And it works !
<Myy>
On a 4.13 kernel
<Myy>
I'll try to pull/push today, using wzyy2 version of rockchip_forwardports
<Myy>
This will still require a kernel patch though, to export iommu_dma_init().
<Myy>
The testing was done with MPV using MPP (rkdec) output on some anime mp4 I had on my HDD.
<Myy>
I'll still try with a real H264 in order to be sure that everything works fine with the decoder though...