drm/nouveau: use dma.max rather than pushbuf size for checking GET validity

Some upcoming G80 DMA changes will depend on this, but it's split out for
bisectibility just in case it causes some unexpected issues.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
This commit is contained in:
Ben Skeggs 2010-01-08 10:53:40 +10:00
parent 1dee7a930b
commit dff3632149

View file

@ -130,7 +130,7 @@ READ_GET(struct nouveau_channel *chan, uint32_t *get)
val = nvchan_rd32(chan, chan->user_get); val = nvchan_rd32(chan, chan->user_get);
if (val < chan->pushbuf_base || if (val < chan->pushbuf_base ||
val >= chan->pushbuf_base + chan->pushbuf_bo->bo.mem.size) { val > chan->pushbuf_base + (chan->dma.max << 2)) {
/* meaningless to dma_wait() except to know whether the /* meaningless to dma_wait() except to know whether the
* GPU has stalled or not * GPU has stalled or not
*/ */