From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1032843AbdAJM06 (ORCPT ); Tue, 10 Jan 2017 07:26:58 -0500 Received: from mx2.suse.de ([195.135.220.15]:35790 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754932AbdAJM05 (ORCPT ); Tue, 10 Jan 2017 07:26:57 -0500 Date: Tue, 10 Jan 2017 13:26:54 +0100 From: Michal Hocko To: "Kirill A. Shutemov" Cc: Keno Fischer , linux-kernel@vger.kernel.org, linux-mm@kvack.org, torvalds@linux-foundation.org, gthelen@google.com, npiggin@gmail.com, w@1wt.eu, oleg@redhat.com, keescook@chromium.org, luto@kernel.org, hughd@google.com Subject: Re: [PATCH v2] mm: Respect FOLL_FORCE/FOLL_COW for thp Message-ID: <20170110122654.GE28025@dhcp22.suse.cz> References: <20170106015025.GA38411@juliacomputing.com> <20170106081844.GA4454@node.shutemov.name> <20170110092909.GA28025@dhcp22.suse.cz> <20170110122045.GA2058@node.shutemov.name> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20170110122045.GA2058@node.shutemov.name> User-Agent: Mutt/1.6.0 (2016-04-01) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue 10-01-17 15:20:45, Kirill A. Shutemov wrote: > On Tue, Jan 10, 2017 at 10:29:10AM +0100, Michal Hocko wrote: > > On Fri 06-01-17 11:18:44, Kirill A. Shutemov wrote: > > > On Thu, Jan 05, 2017 at 08:50:25PM -0500, Keno Fischer wrote: > > > > In 19be0eaff ("mm: remove gup_flags FOLL_WRITE games from __get_user_pages()"), > > > > the mm code was changed from unsetting FOLL_WRITE after a COW was resolved to > > > > setting the (newly introduced) FOLL_COW instead. Simultaneously, the check in > > > > gup.c was updated to still allow writes with FOLL_FORCE set if FOLL_COW had > > > > also been set. However, a similar check in huge_memory.c was forgotten. As a > > > > result, remote memory writes to ro regions of memory backed by transparent huge > > > > pages cause an infinite loop in the kernel (handle_mm_fault sets FOLL_COW and > > > > returns 0 causing a retry, but follow_trans_huge_pmd bails out immidiately > > > > because `(flags & FOLL_WRITE) && !pmd_write(*pmd)` is true. While in this > > > > state the process is stil SIGKILLable, but little else works (e.g. no ptrace > > > > attach, no other signals). This is easily reproduced with the following > > > > code (assuming thp are set to always): > > > > > > > > #include > > > > #include > > > > #include > > > > #include > > > > #include > > > > #include > > > > #include > > > > #include > > > > #include > > > > #include > > > > > > > > #define TEST_SIZE 5 * 1024 * 1024 > > > > > > > > int main(void) { > > > > int status; > > > > pid_t child; > > > > int fd = open("/proc/self/mem", O_RDWR); > > > > void *addr = mmap(NULL, TEST_SIZE, PROT_READ, > > > > MAP_ANONYMOUS | MAP_PRIVATE, 0, 0); > > > > assert(addr != MAP_FAILED); > > > > pid_t parent_pid = getpid(); > > > > if ((child = fork()) == 0) { > > > > void *addr2 = mmap(NULL, TEST_SIZE, PROT_READ | PROT_WRITE, > > > > MAP_ANONYMOUS | MAP_PRIVATE, 0, 0); > > > > assert(addr2 != MAP_FAILED); > > > > memset(addr2, 'a', TEST_SIZE); > > > > pwrite(fd, addr2, TEST_SIZE, (uintptr_t)addr); > > > > return 0; > > > > } > > > > assert(child == waitpid(child, &status, 0)); > > > > assert(WIFEXITED(status) && WEXITSTATUS(status) == 0); > > > > return 0; > > > > } > > > > > > > > Fix this by updating follow_trans_huge_pmd in huge_memory.c analogously to > > > > the update in gup.c in the original commit. The same pattern exists in > > > > follow_devmap_pmd. However, we should not be able to reach that check > > > > with FOLL_COW set, so add WARN_ONCE to make sure we notice if we ever > > > > do. > > > > > > > > Signed-off-by: Keno Fischer > > > > > > Cc: stable@ ? > > > > Yes, please! I am just wondering how far do we have to go. I was under > > impression that we split THP in the past in the gup path but I cannot > > find the respective code now. Many things have changed after your > > refcount rework. Could you clarify this part Kirill, please? > > No, we didn't split THP before, unless it's asked specifically with > FOLL_SPLIT. Otherwise we just pin whole huge page. Yeah, I've tried to find the FOLL_SPLIT but couldn't... > I think we need to port it all active stable trees as we do with > 19be0eaff. The race was there since beginning of THP, I believe. thanks for double checking! -- Michal Hocko SUSE Labs