Re: [PATCH hmm 0/6] Small hmm_range_fault() cleanups

From: Ralph Campbell <rcampbell@nvidia.com>
To: Jason Gunthorpe <jgg@ziepe.ca>,
	Jerome Glisse <jglisse@redhat.com>, <Felix.Kuehling@amd.com>
Cc: Philip Yang <Philip.Yang@amd.com>,
	John Hubbard <jhubbard@nvidia.com>,
	amd-gfx@lists.freedesktop.org, linux-mm@kvack.org,
	Jason Gunthorpe <jgg@mellanox.com>,
	dri-devel@lists.freedesktop.org, Christoph Hellwig <hch@lst.de>
Subject: Re: [PATCH hmm 0/6] Small hmm_range_fault() cleanups
Date: Fri, 20 Mar 2020 11:51:47 -0700	[thread overview]
Message-ID: <030773be-3ea0-033e-051e-a3c0a13ef709@nvidia.com> (raw)
In-Reply-To: <20200320164905.21722-1-jgg@ziepe.ca>

On 3/20/20 9:48 AM, Jason Gunthorpe wrote:
> From: Jason Gunthorpe <jgg@mellanox.com>
> 
> I've had these in my work queue for a bit, nothing profound here, just some
> small edits for clarity.

The hmm tester changes are clear enough but I'm having a bit of trouble figuring out
what this series applies cleanly to since I'm trying to apply it on top of the
other patches you and Christoph have sent out.
Is there a private git tree/branch where everything is applied?

> Ralph's hmm tester will need a small diff to work after this - which
> illustrates how setting default_flags == 0 is the same as what was called
> SNAPSHOT:
> 
> diff --git a/lib/test_hmm.c b/lib/test_hmm.c
> index 6ca953926dc13f..5f31f5b3e64cb9 100644
> --- a/lib/test_hmm.c
> +++ b/lib/test_hmm.c
> @@ -300,7 +300,7 @@ static int dmirror_range_fault(struct dmirror *dmirror,
>   
>   		range->notifier_seq = mmu_interval_read_begin(range->notifier);
>   		down_read(&mm->mmap_sem);
> -		count = hmm_range_fault(range, 0);
> +		count = hmm_range_fault(range);
>   		up_read(&mm->mmap_sem);
>   		if (count <= 0) {
>   			if (count == 0 || count == -EBUSY)
> @@ -337,8 +337,7 @@ static int dmirror_fault(struct dmirror *dmirror, unsigned long start,
>   		.flags = dmirror_hmm_flags,
>   		.values = dmirror_hmm_values,
>   		.pfn_shift = DPT_SHIFT,
> -		.pfn_flags_mask = ~(dmirror_hmm_flags[HMM_PFN_VALID] |
> -				    dmirror_hmm_flags[HMM_PFN_WRITE]),
> +		.pfn_flags_mask = 0,
>   		.default_flags = dmirror_hmm_flags[HMM_PFN_VALID] |
>   				(write ? dmirror_hmm_flags[HMM_PFN_WRITE] : 0),
>   		.dev_private_owner = dmirror->mdevice,
> @@ -872,7 +871,7 @@ static int dmirror_range_snapshot(struct dmirror *dmirror,
>   		range->notifier_seq = mmu_interval_read_begin(range->notifier);
>   
>   		down_read(&mm->mmap_sem);
> -		count = hmm_range_fault(range, HMM_FAULT_SNAPSHOT);
> +		count = hmm_range_fault(range);
>   		up_read(&mm->mmap_sem);
>   		if (count <= 0) {
>   			if (count == 0 || count == -EBUSY)
> @@ -916,7 +915,7 @@ static int dmirror_snapshot(struct dmirror *dmirror,
>   		.flags = dmirror_hmm_flags,
>   		.values = dmirror_hmm_values,
>   		.pfn_shift = DPT_SHIFT,
> -		.pfn_flags_mask = ~0ULL,
> +		.pfn_flags_mask = 0,
>   		.dev_private_owner = dmirror->mdevice,
>   	};
>   	int ret = 0;
> 
> Jason Gunthorpe (6):
>    mm/hmm: remove pgmap checking for devmap pages
>    mm/hmm: return the fault type from hmm_pte_need_fault()
>    mm/hmm: remove unused code and tidy comments
>    mm/hmm: remove HMM_FAULT_SNAPSHOT
>    mm/hmm: remove the CONFIG_TRANSPARENT_HUGEPAGE #ifdef
>    mm/hmm: use device_private_entry_to_pfn()
> 
>   Documentation/vm/hmm.rst                |  12 +-
>   drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c |   2 +-
>   drivers/gpu/drm/nouveau/nouveau_svm.c   |   2 +-
>   include/linux/hmm.h                     |  55 +-----
>   mm/hmm.c                                | 238 +++++++++---------------
>   5 files changed, 98 insertions(+), 211 deletions(-)
> 
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx