On 18 Feb 2015 17:08, "Alexandre Courbot" <<a href="mailto:gnurou@gmail.com">gnurou@gmail.com</a>> wrote: > > On Wed, Feb 18, 2015 at 8:01 AM, Ben Skeggs <<a href="mailto:skeggsb@gmail.com">skeggsb@gmail.com</a>> wrote: > > On Tue, Feb 17, 2015 at 5:47 PM, Alexandre Courbot <<a href="mailto:acourbot@nvidia.com">acourbot@nvidia.com</a>> wrote: > >> Having a RAM device does not make sense for chips like GK20A which have > >> no dedicated video memory. The dummy RAM device that we used so far > >> works as a temporary band-aid, but in the long-term it is desirable for > >> the driver to be able to work without any kind of VRAM. > >> > >> This patch adds a few conditionals in places where a RAM device was > >> assumed to be present and allows some more objects to be allocated from > >> the TT domain, allowing Nouveau to handle GPUs for which > >> pfb->ram == NULL. > >> > >> Signed-off-by: Alexandre Courbot <<a href="mailto:acourbot@nvidia.com">acourbot@nvidia.com</a>> > >> --- > >> drm/nouveau/nouveau_display.c | 8 +++++++- > >> drm/nouveau/nouveau_ttm.c | 3 +++ > >> drm/nouveau/nv84_fence.c | 14 +++++++++++--- > >> drm/nouveau/nvkm/engine/device/base.c | 9 ++++++--- > >> drm/nouveau/nvkm/subdev/clk/base.c | 2 +- > >> drm/nouveau/nvkm/subdev/fb/base.c | 26 ++++++++++++++++++-------- > >> drm/nouveau/nvkm/subdev/ltc/gf100.c | 10 +++++++++- > >> 7 files changed, 55 insertions(+), 17 deletions(-) > >> > >> diff --git a/drm/nouveau/nouveau_display.c b/drm/nouveau/nouveau_display.c > >> index 860b0e2d4181..68ee0af22eea 100644 > >> --- a/drm/nouveau/nouveau_display.c > >> +++ b/drm/nouveau/nouveau_display.c > >> @@ -869,13 +869,19 @@ nouveau_display_dumb_create(struct drm_file *file_priv, struct drm_device *dev, > >> struct drm_mode_create_dumb *args) > >> { > >> struct nouveau_bo *bo; > >> + uint32_t domain; > >> int ret; > >> > >> args->pitch = roundup(args->width * (args->bpp / 8), 256); > >> args->size = args->pitch * args->height; > >> args->size = roundup(args->size, PAGE_SIZE); > >> > >> - ret = nouveau_gem_new(dev, args->size, 0, NOUVEAU_GEM_DOMAIN_VRAM, 0, 0, &bo); > >> + if (nvxx_fb(&nouveau_drm(dev)->device)->ram) > > For these checks in the drm, it's probably better to use > > nouveau_drm(dev)->device.info.ram_size. > > I wonder - in other places (e.g. clock, ltc) we don't have access to > nouveau_drm, so IIUC we need to rely on pfb->ram there. Correct. >Wouldn't it be > more confusing to use two different ways to check the presence of VRAM > when we could stick to a single one? It's best to think of nvkm/ as a separate entity, and it will be at some point (drm load on its own, inside a vm), and drm might not be able to access it's internal structures. That's not the case now, so the code is fine as-is for the moment. But it's worth keeping in mind. Thanks, Ben.