linux-kernel - Re: [PATCH v2] drm/panthor: Simplify mmu_hw_do_operation

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <e8afaa94-1ba8-4a7b-8e3d-123d108c932a@arm.com>
Date: Fri, 15 Aug 2025 15:53:40 +0100
From: Karunika Choo <karunika.choo@....com>
To: Steven Price <steven.price@....com>,
 Boris Brezillon <boris.brezillon@...labora.com>,
 Liviu Dudau <liviu.dudau@....com>, Daniel Stone <daniel@...ishbar.org>
Cc: dri-devel@...ts.freedesktop.org, linux-kernel@...r.kernel.org,
 Chia-I Wu <olvaffe@...il.com>, nd@....com
Subject: Re: [PATCH v2] drm/panthor: Simplify mmu_hw_do_operation_locked

On 15/08/2025 15:10, Steven Price wrote:
> On 15/08/2025 15:01, Karunika Choo wrote:
>> On 15/08/2025 14:42, Steven Price wrote:
>>> The only callers to mmu_hw_do_operation_locked() pass an 'op' of either
>>> AS_COMAND_FLUSH_MEM or AS_COMMAND_FLUSH_PT. This means the code paths
>>> after that are dead. Removing those paths means the
>>> mmu_hw_do_flush_on_gpu_ctrl() function might has well be inlined.
>>>
>>> Simplify everything by having a switch statement for the type of 'op'
>>> (warning if we get an unexpected value) and removing the dead cases.
>>>
>>> Suggested-by: Daniel Stone <daniel@...ishbar.org>
>>> Signed-off-by: Steven Price <steven.price@....com>
>>> ---
>>> Changes from v1:
>>>  * As well as removing dead code, inline mmu_hw_do_flush_on_gpu_ctrl
>>>
>>>  drivers/gpu/drm/panthor/panthor_mmu.c | 57 ++++++++++++---------------
>>>  1 file changed, 26 insertions(+), 31 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/panthor/panthor_mmu.c b/drivers/gpu/drm/panthor/panthor_mmu.c
>>> index 367c89aca558..9d77e7c16ed2 100644
>>> --- a/drivers/gpu/drm/panthor/panthor_mmu.c
>>> +++ b/drivers/gpu/drm/panthor/panthor_mmu.c
>>> @@ -569,15 +569,37 @@ static void lock_region(struct panthor_device *ptdev, u32 as_nr,
>>>  	write_cmd(ptdev, as_nr, AS_COMMAND_LOCK);
>>>  }
>>>  
>>> -static int mmu_hw_do_flush_on_gpu_ctrl(struct panthor_device *ptdev, int as_nr,
>>> -				       u32 op)
>>> +static int mmu_hw_do_operation_locked(struct panthor_device *ptdev, int as_nr,
>>> +				      u64 iova, u64 size, u32 op)
>>>  {
>>>  	const u32 l2_flush_op = CACHE_CLEAN | CACHE_INV;
>>> -	u32 lsc_flush_op = 0;
>>> +	u32 lsc_flush_op;
>>>  	int ret;
>>>  
>>> -	if (op == AS_COMMAND_FLUSH_MEM)
>>> +	lockdep_assert_held(&ptdev->mmu->as.slots_lock);
>>> +
>>> +	switch (op) {
>>> +	case AS_COMMAND_FLUSH_MEM:
>>>  		lsc_flush_op = CACHE_CLEAN | CACHE_INV;
>>> +		break;
>>> +	case AS_COMMAND_FLUSH_PT:
>>> +		lsc_flush_op = 0;
>>> +		break;
>>> +	default:
>>> +		drm_WARN(&ptdev->base, 1, "Unexpected AS_COMMAND: %d", op);
>>> +		return -EINVAL;
>>> +	}
>>> +
>>> +	if (as_nr < 0)
>>> +		return 0;
>>> +
>>
>> Hi Steve,
>>
>> Thanks for pushing this patch. I was planning to address Daniel's
>> comment next week.
>>
>> One small nit, would it be better to move the (as_nr < 0) check just
>> after the lockdep_assert_held() (above the switch case)?
> 
> I'm not sure it makes much difference, but there was a minor reason for
> my ordering:
> 
> By having it after the switch statement then if someone adds a call with
> an invalid op value it will always fail (with a warning). Whereas if we
> move the (as_nr < 0) check earlier then there's a chance they won't
> notice if their testing doesn't trigger that case.
> 
> Obviously there might be a (theoretical) performance impact, but I don't
> think the one extra check would be noticeable - this isn't exactly a
> major fast path. Is there something else I've missed which would justify
> switching it around?
> 

Hi Steve,

Thanks for the explanation. I agree it would be better to have the
switch case first to check for an invalid op. LGTM.

Reviewed-by: Karunika Choo <karunika.choo@....com>

Kind regards,
Karunika

> Thanks,
> Steve
> 
>> Looks good to me otherwise.
>>
>> Kind regards,
>> Karunika
>>
>>> +	/*
>>> +	 * If the AS number is greater than zero, then we can be sure
>>> +	 * the device is up and running, so we don't need to explicitly
>>> +	 * power it up
>>> +	 */
>>> +
>>> +	lock_region(ptdev, as_nr, iova, size);
>>>  
>>>  	ret = wait_ready(ptdev, as_nr);
>>>  	if (ret)
>>> @@ -598,33 +620,6 @@ static int mmu_hw_do_flush_on_gpu_ctrl(struct panthor_device *ptdev, int as_nr,
>>>  	return wait_ready(ptdev, as_nr);
>>>  }
>>>  
>>> -static int mmu_hw_do_operation_locked(struct panthor_device *ptdev, int as_nr,
>>> -				      u64 iova, u64 size, u32 op)
>>> -{
>>> -	lockdep_assert_held(&ptdev->mmu->as.slots_lock);
>>> -
>>> -	if (as_nr < 0)
>>> -		return 0;
>>> -
>>> -	/*
>>> -	 * If the AS number is greater than zero, then we can be sure
>>> -	 * the device is up and running, so we don't need to explicitly
>>> -	 * power it up
>>> -	 */
>>> -
>>> -	if (op != AS_COMMAND_UNLOCK)
>>> -		lock_region(ptdev, as_nr, iova, size);
>>> -
>>> -	if (op == AS_COMMAND_FLUSH_MEM || op == AS_COMMAND_FLUSH_PT)
>>> -		return mmu_hw_do_flush_on_gpu_ctrl(ptdev, as_nr, op);
>>> -
>>> -	/* Run the MMU operation */
>>> -	write_cmd(ptdev, as_nr, op);
>>> -
>>> -	/* Wait for the flush to complete */
>>> -	return wait_ready(ptdev, as_nr);
>>> -}
>>> -
>>>  static int mmu_hw_do_operation(struct panthor_vm *vm,
>>>  			       u64 iova, u64 size, u32 op)
>>>  {
>>
>