Re: [Nouveau] v4.8-rc2 crashes while probing nvidia graphics card on arm64

2016-08-15 Thread Punit Agrawal
I made a hash of LKML email in the previous mail. Re-sending with the address corrected. Apologies for the duplicate. Punit Agrawal writes: > Hi, > > While trying out v4.8-rc2 on Juno r2 (arm64), I ran into the following > crash when probing the nvidia graphics card attached to the

[Nouveau] v4.8-rc2 crashes while probing nvidia graphics card on arm64

2016-08-15 Thread Punit Agrawal
Hi, While trying out v4.8-rc2 on Juno r2 (arm64), I ran into the following crash when probing the nvidia graphics card attached to the PCIe slot. So I tried rc1 and got the same crash. The card probes without any errors on v4.7. Anybody familiear with the recent changes knows what might have led

Re: [Nouveau] [RFC PATCH] drm/nouveau/fb/nv50: set DMA mask before mapping scratch page

2016-06-23 Thread Punit Agrawal
Ard Biesheuvel writes: > On 23 June 2016 at 11:55, Punit Agrawal wrote: >> Ard Biesheuvel writes: >> >>> The 100c08 scratch page is mapped using dma_map_page() before the TTM >>> layer has had a chance to set the DMA mask. This means we are still >>>

Re: [Nouveau] [RFC PATCH] drm/nouveau/fb/nv50: set DMA mask before mapping scratch page

2016-06-23 Thread Punit Agrawal
Ard Biesheuvel writes: > The 100c08 scratch page is mapped using dma_map_page() before the TTM > layer has had a chance to set the DMA mask. This means we are still > running with the default of 32 when this code executes, and this causes > problems for platforms with no memory below 4 GB (such a