kernel/drivers/gpu/host1x/job.c, branch linux-4.17.y

gpu: host1x: Check whether size of unpin isn't 0

2018-08-24T11:07:03Z

[ Upstream commit ec58923215dbbeef59ee82923ee94d745f73db58 ] Only gather pins are mapped by the Host1x driver, regular BO relocations are not. Check whether size of unpin isn't 0, otherwise IOVA allocation at 0x0 could be erroneously released. Signed-off-by: Dmitry Osipenko Signed-off-by: Thierry Reding Signed-off-by: Sasha Levin Signed-off-by: Greg Kroah-Hartman

gpu: host1x: Don't fail on NULL bo physical address

2017-08-17T15:57:06Z

Pinning a Host1x BO currently cannot fail and zero is a valid address for a BO when IOMMU is enabled. To avoid false errors remove checks for NULL BO physical addresses. Fixes: 404bfb78daf3 ("gpu: host1x: Add IOMMU support") Signed-off-by: Mikko Perttunen Reviewed-by: Dmitry Osipenko Signed-off-by: Thierry Reding

gpu: host1x: At first try a non-blocking allocation for the gather copy

2017-06-15T12:25:56Z

The blocking gather copy allocation is a major performance downside of the Host1x firewall, it may take hundreds milliseconds which is unacceptable for the real-time graphics operations. Let's try a non-blocking allocation first as a least invasive solution, it makes opentegra (Xorg driver) performance indistinguishable with/without the firewall. Signed-off-by: Dmitry Osipenko Reviewed-by: Erik Faye-Lund Signed-off-by: Thierry Reding

gpu: host1x: Check waits in the firewall

2017-06-15T12:24:41Z

Check waits in the firewall in a way it is done for relocations. Signed-off-by: Dmitry Osipenko Reviewed-by: Mikko Perttunen Reviewed-by: Erik Faye-Lund Signed-off-by: Thierry Reding

gpu: host1x: Forbid unrelated SETCLASS opcode in the firewall

2017-06-15T12:23:50Z

Several channels could be made to write the same unit concurrently via the SETCLASS opcode, trusting userspace is a bad idea. It should be possible to drop the per-client channel reservation and add a per-unit locking by inserting MLOCK's to the command stream to re-allow the SETCLASS opcode, but it will be much more work. Let's forbid the unit-unrelated class changes for now. Signed-off-by: Dmitry Osipenko Reviewed-by: Erik Faye-Lund Reviewed-by: Mikko Perttunen Signed-off-by: Thierry Reding

gpu: host1x: Forbid RESTART opcode in the firewall

2017-06-15T12:23:18Z

The RESTART opcode terminates the gather and restarts the CDMA fetching from a specified word << 2 relative to the CDMA start address. That shouldn't be allowed to be done by userspace. Signed-off-by: Dmitry Osipenko Reviewed-by: Erik Faye-Lund Reviewed-by: Mikko Perttunen Signed-off-by: Thierry Reding

gpu: host1x: Forbid relocation address shifting in the firewall

2017-06-15T12:22:32Z

Incorrectly shifted relocation address will cause a lower memory corruption and likely a hang on a write or a read of an arbitrary data in case of IOMMU absence. As of now, there is no known use for the address shifting and adding a proper shifts / sizes validation is a much more work. Let's forbid shifts in the firewall till a proper validation is implemented. Signed-off-by: Dmitry Osipenko Reviewed-by: Erik Faye-Lund Reviewed-by: Mikko Perttunen Signed-off-by: Thierry Reding

gpu: host1x: Do not leak BO's phys address to userspace

2017-06-15T12:22:03Z

Perform gathers coping before patching them, so that original gathers are left untouched. That's not as bad as leaking kernel addresses, but still doesn't feel right. Signed-off-by: Dmitry Osipenko Reviewed-by: Mikko Perttunen Signed-off-by: Thierry Reding

gpu: host1x: Correct host1x_job_pin() error handling

2017-06-15T12:21:46Z

In case of relocations / waitchecks patching failure the jobs pins stay referenced till DRM file get closed, wasting memory. Add the missed unpinning. Signed-off-by: Dmitry Osipenko Reviewed-by: Erik Faye-Lund Reviewed-by: Mikko Perttunen Signed-off-by: Thierry Reding

gpu: host1x: Initialize firewall class to the job's one

2017-06-15T12:21:23Z

The commands stream is prepended by the jobs class on the CDMA submission, so that explicitly setting a module class in the commands stream isn't necessary. The firewall initializes its class to 0 and the command stream that doesn't explicitly specify the class effectively bypasses the firewall. Signed-off-by: Dmitry Osipenko Reviewed-by: Mikko Perttunen Reviewed-by: Erik Faye-Lund Signed-off-by: Thierry Reding