[PATCH] iommu/rockchip: disable fetch dte time limit

[Sven Püschel]

Hi,

pinging this Patch, as I didn't receive any more feedback/reviews. The 
feedback only affects the comment and doesn't affect the commit, 
therefore I currently don't see a reason for a v2.

We've noticed that this patches fixes hangups when using the RGA3 
peripheral on the rk3588 (where I'm in the process of upstreaming a 
driver [1], which depends on this changeset). Based on the vendor commit 
messages, it should also fix VOP and screen black issues.

Sincerely
     Sven

[1] 
https://lore.kernel.org/linux-media/20260428-spu-rga3-v5-0-eb7f5d019d86@pengutronix.de/

On 11/28/25 8:26 AM, Sven Püschel wrote:
>
> On 11/27/25 5:32 AM, Shawn Lin wrote:
>> 在 2025/11/26 星期三 19:45, Sven Püschel 写道:
>>> From: Simon Xue <xxm at rock-chips.com>
>>>
>>> Disable the Bit 31 of the AUTO_GATING iommu register, as it causes
>>> hangups with the RGA3 (Raster Graphics Acceleration 3) peripheral.
>>> The RGA3 register description of the TRM already states that the bit
>>> must be set to 1. The vendor kernel sets the bit unconditionally to
>>> 1 to fix VOP (Video Output Processor) screen black issues. This patch
>>> squashes the 2 vendor kernel commits with the following commit 
>>> messages:
>>>
>>> Master fetch data and cpu update page table may work in parallel, may
>>> have the following procedure:
>>>
>>>     master                  cpu
>>>     fetch dte               update page tabl
>>>             |                       |
>>>     (make dte invalid)  <-  zap iotlb entry
>>>             |                       |
>>>     fetch dte again
>>>     (make dte invalid)  <-  zap iotlb entry
>>>             |                       |
>>>     fetch dte again
>>>     (make dte invalid)  <-  zap iotlb entry
>>>             |                       |
>>>     fetch dte again
>>>     (make iommu block)  <-  zap iotlb entry
>>>
>>> New iommu version has the above bug, if fetch dte consecutively four
>>> times, then it will be blocked. Fortunately, we can set bit 31 of
>>> register MMU_AUTO_GATING to 1 to make it work as old version which does
>>> not have this issue.
>>>
>>> This issue only appears on RV1126 so far, so make a workaround 
>>> dedicated
>>> to "rockchip,rv1126" machine type.
>>>
>>> iommu/rockchip: fix vop blocked and screen black on RK356X and RK3588
>>>
>>> RK3568 and RK3588 has the same issue as RV1126/RV1109 that caused by
>>> dte fetch time limit, So we can set BIT(31) of register 0x24 default
>>> to 1 as a workaround.
>>>
>>> Signed-off-by: Simon Xue <xxm at rock-chips.com>
>>> Signed-off-by: Sven Püschel <s.pueschel at pengutronix.de>
>>> ---
>>> During testing of a newly developed driver for the RGA3 peripheral [1]
>>> (Raster Graphic Acceleration 3) of the RK3588 some sporadic hangs
>>> have been observed. The upstream rockchip-iommu driver is used to 
>>> handle
>>> the RGA3 IOMMU register space.
>>>
>>> After a closer look at the TRM for the RK3588, the RGA3 iommu register
>>> description of the RGA3_MMU_AUTO_GATING register (offset 0x24) mentions
>>
>> It's 0xF24 per RGA3 chapter.
>
> yeah, sorry. I was already thinking relative to the rga3 iommu address 
> space in my head, but didn't really mention that the chapter says 
> 0xF24 with the iommu related registers starting at 0xF00.
>
> Sincerely
>     Sven
>
>>
>>> a mmu_bug_fixed_disable bit, which must be set to 1 but defaults to 0.
>>>
>>> Looking at the commits in the vendor kernel, the bit is unconditionally
>>> set to 1 and mentions that it fixes a blocked VOP (Video Output
>>> Processor) [3]. Therefore squash the relevant vendor commits
>>> [2] and [3] into a single patch, combine the commit messages and keep
>>> the Signed-off-by line from the original author.
>>>
>>> [1] 
>>> https://lore.kernel.org/all/20251007-spu-rga3-v1-0-36ad85570402@pengutronix.de/
>>> [2] 
>>> https://github.com/rockchip-linux/kernel/commit/7f8158fb41b5cc8e738aaeebc3637c50ebd74cae
>>> [3] 
>>> https://github.com/rockchip-linux/kernel/commit/6a355e5f9a2069a2309e240791bc3aad63b7324e
>>> ---
>>>   drivers/iommu/rockchip-iommu.c | 8 ++++++++
>>>   1 file changed, 8 insertions(+)
>>>
>>> diff --git a/drivers/iommu/rockchip-iommu.c 
>>> b/drivers/iommu/rockchip-iommu.c
>>> index 0861dd469bd86..2d0dabb0d101a 100644
>>> --- a/drivers/iommu/rockchip-iommu.c
>>> +++ b/drivers/iommu/rockchip-iommu.c
>>> @@ -76,6 +76,8 @@
>>>   #define SPAGE_ORDER 12
>>>   #define SPAGE_SIZE (1 << SPAGE_ORDER)
>>>   +#define DISABLE_FETCH_DTE_TIME_LIMIT BIT(31)
>>> +
>>>    /*
>>>     * Support mapping any size that fits in one page table:
>>>     *   4 KiB to 4 MiB
>>> @@ -930,6 +932,7 @@ static int rk_iommu_enable(struct rk_iommu *iommu)
>>>       struct iommu_domain *domain = iommu->domain;
>>>       struct rk_iommu_domain *rk_domain = to_rk_domain(domain);
>>>       int ret, i;
>>> +    u32 auto_gate;
>>>         ret = clk_bulk_enable(iommu->num_clocks, iommu->clocks);
>>>       if (ret)
>>> @@ -948,6 +951,11 @@ static int rk_iommu_enable(struct rk_iommu *iommu)
>>> rk_ops->mk_dtentries(rk_domain->dt_dma));
>>>           rk_iommu_base_command(iommu->bases[i], RK_MMU_CMD_ZAP_CACHE);
>>>           rk_iommu_write(iommu->bases[i], RK_MMU_INT_MASK, 
>>> RK_MMU_IRQ_MASK);
>>> +
>>> +        /* Workaround for iommu blocked, BIT(31) default to 1 */
>>> +        auto_gate = rk_iommu_read(iommu->bases[i], 
>>> RK_MMU_AUTO_GATING);
>>> +        auto_gate |= DISABLE_FETCH_DTE_TIME_LIMIT;
>>> +        rk_iommu_write(iommu->bases[i], RK_MMU_AUTO_GATING, 
>>> auto_gate);
>>>       }
>>>         ret = rk_iommu_enable_paging(iommu);
>>>
>>> ---
>>> base-commit: 30f09200cc4aefbd8385b01e41bde2e4565a6f0e
>>> change-id: 20251126-spu-iommudtefix-cd0c5244c74a
>>>
>>> Best regards,
>>
>>