gpu: nvgpu: Move debug dump in MMU fault handler

Move this debug dump to after the regular MMU fault prints. This
makes it so that the MMU fault prints happen first. That means
that when you scan the log you will see the MMU fault first, then
all the channel, engine, pbdma, and (potentially) host1x data.

Change-Id: I9e371bb95f3c8d21df1c375ed45e1f0b78810a7c
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1940516
GVS: Gerrit_Virtual_Submit
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
This commit is contained in:
Alex Waterman
2018-10-31 13:20:43 -07:00
committed by mobile promotions
parent 032b37bee5
commit 1f4ab1b36e

View File

@@ -1649,7 +1649,6 @@ static bool gk20a_fifo_handle_mmu_fault_locked(
} else {
fault_id = gk20a_readl(g, fifo_intr_mmu_fault_id_r());
fake_fault = false;
gk20a_debug_dump(g);
}
@@ -1825,6 +1824,10 @@ static bool gk20a_fifo_handle_mmu_fault_locked(
}
}
if (!fake_fault) {
gk20a_debug_dump(g);
}
/* clear interrupt */
gk20a_writel(g, fifo_intr_mmu_fault_id_r(), (u32)fault_id);