On 2022-04-19 15:40, Will Deacon wrote:
On Thu, Apr 14, 2022 at 01:42:33PM +0100, Robin Murphy wrote:
Stop calling bus_set_iommu() since it's now unnecessary. With device
probes now replayed for every IOMMU instance registration, the whole
sorry ordering workaround for legacy DT bindings goes too, hooray!
Ha, I hope you tested this!
Oh alright then, since it's you... :)
I've hacked up a Juno DT with the old bindings, and (after needing a
while to remember that they're fundamentally incompatible with
disable_bypass), can confirm that with my whole dev branch including
this series applied, it boots and creates IOMMU groups as expected. I
then made the mistake of trying without the branch to check whether the
squawks from iommu_setup_dma_ops() were new or not, and... well... plain
rc3 doesn't even boot on the same setup - it's somehow blowing up in the
failure cleanup path of iommu_bus_init(), apparently calling
iommu_release_device() on something where dev->iommu->iommu_dev is NULL,
for reasons that are far from clear and I'm not sure I can really be
bothered to debug further... :/
Cheers,
Robin.
Signed-off-by: Robin Murphy <robin.mur...@arm.com>
---
drivers/iommu/arm/arm-smmu/arm-smmu.c | 84 +--------------------------
1 file changed, 2 insertions(+), 82 deletions(-)
Assuming it works,
Acked-by: Will Deacon <w...@kernel.org>
Will
_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu