On Tue, Apr 26, 2022 at 03:06:35PM +0000, Liam Howlett wrote: > From: "Liam R. Howlett" <Liam.Howlett@xxxxxxxxxx> > > Avoid allocating a new VMA when it a vma modification can occur. When a > brk() can expand or contract a VMA, then the single store operation will > only modify one index of the maple tree instead of causing a node to split > or coalesce. This avoids unnecessary allocations/frees of maple tree > nodes and VMAs. > > Move some limit & flag verifications out of the do_brk_flags() function to > use only relevant checks in the code path of bkr() and vm_brk_flags(). > > Set the vma to check if it can expand in vm_brk_flags() if extra criteria > are met. > > Drop userfaultfd from do_brk_flags() path and only use it in > vm_brk_flags() path since that is the only place a munmap will happen. > > Signed-off-by: Liam R. Howlett <Liam.Howlett@xxxxxxxxxx> A build failure seen when building ppc64:corenet64_smp_defconfig also biscects to this patch. mm/mmap.c: In function 'do_brk_flags': mm/mmap.c:2908:17: error: implicit declaration of function 'khugepaged_enter_vma_merge'; did you mean 'khugepaged_enter_vma'? It appears that this is later fixed, but it hurts bisectability (and prevents me from finding the actual build failure in linux-next when trying to build corenet64_smp_defconfig). While looking into the patch, I noticed the following. [ ... ] > @@ -2773,30 +2888,52 @@ static int do_brk_flags(unsigned long addr, unsigned long len, > if (security_vm_enough_memory_mm(mm, len >> PAGE_SHIFT)) > return -ENOMEM; > > out: > perf_event_mmap(vma); > mm->total_vm += len >> PAGE_SHIFT; > @@ -2807,18 +2944,21 @@ static int do_brk_flags(unsigned long addr, unsigned long len, > validate_mm_mt(mm); > return 0; > > -no_vma_link: > vm_area_free(vma); That really looks wrong. vm_area_free(vma) can not be reached after this patch has been applied. Guenter > +vma_alloc_fail: > + vm_unacct_memory(len >> PAGE_SHIFT); > return -ENOMEM; > } >
