From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Date: Tue, 31 May 2022 09:59:37 +0200 From: Heiko Carstens Subject: Re: [PATCH v5] mm: Avoid unnecessary page fault retires on shared memory types Message-ID: References: <20220530183450.42886-1-peterx@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20220530183450.42886-1-peterx@redhat.com> To: Peter Xu Cc: linux-kernel@vger.kernel.org, linux-mm@kvack.org, Ingo Molnar , Vasily Gorbik , linuxppc-dev@lists.ozlabs.org, Stafford Horne , "David S . Miller" , Johannes Berg , Brian Cain , x86@kernel.org, linux-parisc@vger.kernel.org, Richard Henderson , Christian Borntraeger , Richard Weinberger , linux-hexagon@vger.kernel.org, Benjamin Herrenschmidt , Thomas Gleixner , Janosch Frank , Albert Ou , Anton Ivanov , Dave Hansen , linux-ia64@vger.kernel.org, Borislav Petkov , Russell King , Sven Schnelle , Andrea Arcangeli , "James E . J . Bottomley" , Al Viro , Alexander Gordeev , Peter Zijlstra , Alistair Popple , Jonas Bonn , sparclinux@vger.kernel.org, linux-csky@vger.kernel.org, Will Deacon , Andrew Morton , Vlastimil Babka , linux-um@lists.infradead.org, Michal Simek , Matt Turner , linux-m68k@lists.linux-m68k.org, Paul Mackerras , linux-xtensa@linux-xtensa.org, Geert Uytterhoeven , David Hildenbrand , openrisc@lists.librecores.org, linux-arm-kernel@lists.infradead.org, Nicholas Piggin , Palmer Dabbelt , Michael Ellerman , Stefan Kristiansson , Paul Walmsley , Ivan Kokshaysky , Chris Zankel , Hugh Dickins , Vineet Gupta , Dinh Nguyen , Catalin Marinas , Rich Felker , "H . Peter Anvin" , linux-s390@vger.kernel.org, linux-sh@vger.kernel.org, linux-riscv@lists.infradead.org, Johannes Weiner , Andy Lutomirski , Max Filippov , Guo Ren , Thomas Bogendoerfer , linux-snps-arc@lists.infradead.org, linux-alpha@vger.kernel.org, linux-mips@vger.kernel.org, Helge Deller , Yoshinori Sato , Ingo Molnar List-ID: On Mon, May 30, 2022 at 02:34:50PM -0400, Peter Xu wrote: > I observed that for each of the shared file-backed page faults, we're very > likely to retry one more time for the 1st write fault upon no page. It's > because we'll need to release the mmap lock for dirty rate limit purpose > with balance_dirty_pages_ratelimited() (in fault_dirty_shared_page()). > > Then after that throttling we return VM_FAULT_RETRY. > > We did that probably because VM_FAULT_RETRY is the only way we can return > to the fault handler at that time telling it we've released the mmap lock. > > However that's not ideal because it's very likely the fault does not need > to be retried at all since the pgtable was well installed before the > throttling, so the next continuous fault (including taking mmap read lock, > walk the pgtable, etc.) could be in most cases unnecessary. > > It's not only slowing down page faults for shared file-backed, but also add > more mmap lock contention which is in most cases not needed at all. > > To observe this, one could try to write to some shmem page and look at > "pgfault" value in /proc/vmstat, then we should expect 2 counts for each > shmem write simply because we retried, and vm event "pgfault" will capture > that. > > To make it more efficient, add a new VM_FAULT_COMPLETED return code just to > show that we've completed the whole fault and released the lock. It's also > a hint that we should very possibly not need another fault immediately on > this page because we've just completed it. > > This patch provides a ~12% perf boost on my aarch64 test VM with a simple > program sequentially dirtying 400MB shmem file being mmap()ed and these are > the time it needs: > > Before: 650.980 ms (+-1.94%) > After: 569.396 ms (+-1.38%) > > I believe it could help more than that. > > We need some special care on GUP and the s390 pgfault handler (for gmap > code before returning from pgfault), the rest changes in the page fault > handlers should be relatively straightforward. > > Another thing to mention is that mm_account_fault() does take this new > fault as a generic fault to be accounted, unlike VM_FAULT_RETRY. > > I explicitly didn't touch hmm_vma_fault() and break_ksm() because they do > not handle VM_FAULT_RETRY even with existing code, so I'm literally keeping > them as-is. > > Acked-by: Geert Uytterhoeven > Acked-by: Peter Zijlstra (Intel) > Acked-by: Johannes Weiner > Acked-by: Vineet Gupta > Acked-by: Guo Ren > Acked-by: Max Filippov > Acked-by: Christian Borntraeger > Acked-by: Michael Ellerman (powerpc) > Acked-by: Catalin Marinas > Reviewed-by: Alistair Popple > Reviewed-by: Ingo Molnar > Signed-off-by: Peter Xu > --- ... > arch/s390/mm/fault.c | 12 ++++++++++++ > diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c > index e173b6187ad5..973dcd05c293 100644 > --- a/arch/s390/mm/fault.c > +++ b/arch/s390/mm/fault.c > @@ -433,6 +433,17 @@ static inline vm_fault_t do_exception(struct pt_regs *regs, int access) > goto out_up; > goto out; > } > + > + /* The fault is fully completed (including releasing mmap lock) */ > + if (fault & VM_FAULT_COMPLETED) { > + if (gmap) { > + mmap_read_lock(mm); > + goto out_gmap; > + } > + fault = 0; > + goto out; > + } > + > if (unlikely(fault & VM_FAULT_ERROR)) > goto out_up; > > @@ -452,6 +463,7 @@ static inline vm_fault_t do_exception(struct pt_regs *regs, int access) > mmap_read_lock(mm); > goto retry; > } > +out_gmap: > if (IS_ENABLED(CONFIG_PGSTE) && gmap) { > address = __gmap_link(gmap, current->thread.gmap_addr, > address); FWIW: Acked-by: Heiko Carstens