Hi Christian,
On Fr, 2019-10-25 at 12:39 +0200, Christian Gmeiner wrote:
> etnaviv_iommuv2_dump_size(..) returns the number of PTE * SZ_4K but etnaviv_iommuv2_dump(..)
> increments buf pointer even if there is no PTE. This results in a bad buf pointer which gets
> used for memcpy(..).
>
> [ 264.408474] 8<--- cut here ---
> [ 264.412048] Unable to handle kernel paging request at virtual address f1a2c268
> [ 264.419321] pgd = e5846004
> [ 264.422069] [f1a2c268] *pgd=00000000
> [ 264.425702] Internal error: Oops: 805 [#1] SMP ARM
> [ 264.430520] Modules linked in:
> [ 264.433616] CPU: 2 PID: 130 Comm: kworker/2:2 Tainted: G W 5.4.0-rc4 #10
> [ 264.441643] Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
> [ 264.448227] Workqueue: events drm_sched_job_timedout
> [ 264.453237] PC is at memcpy+0x50/0x330
> [ 264.457012] LR is at 0x2
> [ 264.459572] pc : [<c0c04650>] lr : [<00000002>] psr: 200f0013
> [ 264.465863] sp : ec96fe64 ip : 00000002 fp : 00000140
> [ 264.471112] r10: 00003000 r9 : ec688040 r8 : 00000002
> [ 264.476364] r7 : 00000002 r6 : 00000002 r5 : 00000002 r4 : 00000002
> [ 264.482917] r3 : 00000002 r2 : 00000f60 r1 : f162a020 r0 : f1a2c268
> [ 264.489472] Flags: nzCv IRQs on FIQs on Mode SVC_32 ISA ARM Segment none
> [ 264.496635] Control: 10c5387d Table: 3d26804a DAC: 00000051
> [ 264.502407] Process kworker/2:2 (pid: 130, stack limit = 0xe8f69f3d)
> [ 264.508786] Stack: (0xec96fe64 to 0xec970000)
> [ 264.513180] fe60: f1622000 f162218c f162c000 414e5445 f1a2c268 00000ffc c0655a8c
> [ 264.521394] fe80: 00000000 0000012a f162c268 c064fd78 c0657350 c0187f64 00000001 00000000
> [ 264.529606] fea0: ed0f9c00 00000001 00000002 435d587d ec688140 ec688100 ed0f9c00 ec688040
> [ 264.537818] fec0: ed0f9c00 c1308b28 ec96ff1c c13e55b0 c13e41c8 c0657358 ec688260 ed0f9c18
> [ 264.546029] fee0: ec688100 c0641278 ec688260 ec2f6180 ee1ba700 ee1bda00 c1308b28 c0149b98
> [ 264.554240] ff00: 00000001 00000000 c0149ae4 c0c21fb0 00000000 00000000 c014a194 c1a4be34
> [ 264.562452] ff20: c1870740 00000000 c1015384 435d587d ffffe000 ec2f6180 ec2f6194 ee1ba700
> [ 264.570663] ff40: 00000008 ee1ba734 c1305900 ee1ba700 ffffe000 c014a0e4 ec9537a4 c0c28e64
> [ 264.578874] ff60: ec96e000 00000000 ec2be780 ec2f99c0 ec96e000 ec2f6180 c014a0b8 ec13fe90
> [ 264.587086] ff80: ec2be7b8 c0152890 ec96e000 ec2f99c0 c0152750 00000000 00000000 00000000
> [ 264.595296] ffa0: 00000000 00000000 00000000 c01010b4 00000000 00000000 00000000 00000000
> [ 264.603506] ffc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
> [ 264.611716] ffe0: 00000000 00000000 00000000 00000000 00000013 00000000 00000000 00000000
> [ 264.619944] [<c0c04650>] (memcpy) from [<c0655a8c>] (etnaviv_iommuv2_dump+0x58/0x60)
> [ 264.627738] [<c0655a8c>] (etnaviv_iommuv2_dump) from [<c064fd78>] (etnaviv_core_dump+0x140/0x45c)
> [ 264.636658] [<c064fd78>] (etnaviv_core_dump) from [<c0657358>] (etnaviv_sched_timedout_job+0x8c/0xb8)
> [ 264.645923] [<c0657358>] (etnaviv_sched_timedout_job) from [<c0641278>] (drm_sched_job_timedout+0x38/0x88)
> [ 264.655631] [<c0641278>] (drm_sched_job_timedout) from [<c0149b98>] (process_one_work+0x2c4/0x7e4)
> [ 264.664633] [<c0149b98>] (process_one_work) from [<c014a0e4>] (worker_thread+0x2c/0x59c)
> [ 264.672765] [<c014a0e4>] (worker_thread) from [<c0152890>] (kthread+0x140/0x158)
> [ 264.680200] [<c0152890>] (kthread) from [<c01010b4>] (ret_from_fork+0x14/0x20)
> [ 264.687448] Exception stack(0xec96ffb0 to 0xec96fff8)
> [ 264.692530] ffa0: 00000000 00000000 00000000 00000000
> [ 264.700741] ffc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
> [ 264.708949] ffe0: 00000000 00000000 00000000 00000000 00000013 00000000
> [ 264.715599] Code: f5d1f05c f5d1f07c e8b151f8 e2522020 (e8a051f8)
> [ 264.721727] ---[ end trace 8afcd79e9e2725b3 ]---
Code change is fine, but the commit message has a lot of lines over 80
characters and generally the kernel trace is mostly noise. I trimmed it
out and applied it to my fixes branch. If you don't agree with my
changes, please shout now as I'm planning on sending the fixes pull
tomorrow.
Regards,
Lucas
> Fixes: afb7b3b1deb4 ("drm/etnaviv: implement IOMMUv2 translation")
> Cc: stable@xxxxxxxxxxxxxxx
> Signed-off-by: Christian Gmeiner <christian.gmeiner@xxxxxxxxx>
> ---
> drivers/gpu/drm/etnaviv/etnaviv_iommu_v2.c | 6 ++++--
> 1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/etnaviv/etnaviv_iommu_v2.c b/drivers/gpu/drm/etnaviv/etnaviv_iommu_v2.c
> index 043111a1d60c..f8bf488e9d71 100644
> --- a/drivers/gpu/drm/etnaviv/etnaviv_iommu_v2.c
> +++ b/drivers/gpu/drm/etnaviv/etnaviv_iommu_v2.c
> @@ -155,9 +155,11 @@ static void etnaviv_iommuv2_dump(struct etnaviv_iommu_context *context, void *bu
>
> memcpy(buf, v2_context->mtlb_cpu, SZ_4K);
> buf += SZ_4K;
> - for (i = 0; i < MMUv2_MAX_STLB_ENTRIES; i++, buf += SZ_4K)
> - if (v2_context->mtlb_cpu[i] & MMUv2_PTE_PRESENT)
> + for (i = 0; i < MMUv2_MAX_STLB_ENTRIES; i++)
> + if (v2_context->mtlb_cpu[i] & MMUv2_PTE_PRESENT) {
> memcpy(buf, v2_context->stlb_cpu[i], SZ_4K);
> + buf += SZ_4K;
> + }
> }
>
> static void etnaviv_iommuv2_restore_nonsec(struct etnaviv_gpu *gpu,
