On 1/30/24 17:04, Alexandru Elisei wrote: > Hi, > > On Tue, Jan 30, 2024 at 03:25:20PM +0530, Anshuman Khandual wrote: >> >> On 1/25/24 22:12, Alexandru Elisei wrote: >>> arm64 uses VM_HIGH_ARCH_0 and VM_HIGH_ARCH_1 for enabling MTE for a VMA. >>> When VM_HIGH_ARCH_0, which arm64 renames to VM_MTE, is set for a VMA, and >>> the gfp flag __GFP_ZERO is present, the __GFP_ZEROTAGS gfp flag also gets >>> set in vma_alloc_zeroed_movable_folio(). >>> >>> Expand this to be more generic by adding an arch hook that modifes the gfp >>> flags for an allocation when the VMA is known. >>> >>> Note that __GFP_ZEROTAGS is ignored by the page allocator unless __GFP_ZERO >>> is also set; from that point of view, the current behaviour is unchanged, >>> even though the arm64 flag is set in more places. When arm64 will have >>> support to reuse the tag storage for data allocation, the uses of the >>> __GFP_ZEROTAGS flag will be expanded to instruct the page allocator to try >>> to reserve the corresponding tag storage for the pages being allocated. >> Right but how will pushing __GFP_ZEROTAGS addition into gfp_t flags further >> down via a new arch call back i.e arch_calc_vma_gfp() while still maintaining >> (vma->vm_flags & VM_MTE) conditionality improve the current scenario. Because > I'm afraid I don't follow you. I was just asking whether the overall scope of __GFP_ZEROTAGS flag is being increased to cover more core MM paths through this patch. I think you have already answered that below. > >> the page allocator could have still analyzed alloc flags for __GFP_ZEROTAGS >> for any additional stuff. >> >> OR this just adds some new core MM paths to get __GFP_ZEROTAGS which was not >> the case earlier via this call back. > Before this patch: vma_alloc_zeroed_movable_folio() sets __GFP_ZEROTAGS. > After this patch: vma_alloc_folio() sets __GFP_ZEROTAGS. Understood. > > This patch is about adding __GFP_ZEROTAGS for more callers. Right, I guess that is the real motivation for this patch. But just wondering does this cover all possible anon fault paths for converting given vma_flag's VM_MTE flag into page alloc flag __GFP_ZEROTAGS ? Aren't there any other file besides (mm/shmem.c) which needs to be changed to include arch_calc_vma_gfp() ?
