From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752100Ab1GSFgE (ORCPT ); Tue, 19 Jul 2011 01:36:04 -0400 Received: from mail-qw0-f46.google.com ([209.85.216.46]:42831 "EHLO mail-qw0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751021Ab1GSFgC (ORCPT ); Tue, 19 Jul 2011 01:36:02 -0400 Message-ID: <4E25185F.1050605@gmail.com> Date: Tue, 19 Jul 2011 13:38:39 +0800 From: Shan Hai User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2.17) Gecko/20110424 Thunderbird/3.1.10 MIME-Version: 1.0 To: Benjamin Herrenschmidt CC: Peter Zijlstra , Peter Zijlstra , paulus@samba.org, tglx@linutronix.de, walken@google.com, dhowells@redhat.com, cmetcalf@tilera.com, tony.luck@intel.com, akpm@linux-foundation.org, linuxppc-dev@lists.ozlabs.org, linux-kernel@vger.kernel.org Subject: Re: [RFC/PATCH] mm/futex: Fix futex writes on archs with SW tracking of dirty & young References: <1310717238-13857-1-git-send-email-haishan.bai@gmail.com> <1310717238-13857-2-git-send-email-haishan.bai@gmail.com> <1310725418.2586.309.camel@twins> <4E21A526.8010904@gmail.com> <1310860194.25044.17.camel@pasglop> <4b337921-d430-4b63-bc36-ad31753cf800@email.android.com> <1310912990.25044.203.camel@pasglop> <1310944453.25044.262.camel@pasglop> <1310961691.25044.274.camel@pasglop> <4E23D728.7090406@gmail.com> <1310972462.25044.292.camel@pasglop> <4E23E02C.8090401@gmail.com> <1310974591.25044.298.camel@pasglop> <4E24FA51.70602@gmail.com> <1311049762.25044.392.camel@pasglop> <4E251365.9090004@gmail.com> <1311053063.25044.397.camel@pasglop> In-Reply-To: <1311053063.25044.397.camel@pasglop> Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 07/19/2011 01:24 PM, Benjamin Herrenschmidt wrote: > On Tue, 2011-07-19 at 13:17 +0800, Shan Hai wrote: > >> The patch works, but I have certain confusions, >> - Do we want to handle_mm_fault on each futex_lock_pi >> even though in most cases there is no write permission >> fixup's needed? > Don't we only ever call this when futex_atomic_op_inuser() failed ? > Which means a fixup -is- needed .... The fast path is still there. > What you said is another path, that is futex_wake_op(), but what about futex_lock_pi in which my test case failed? your patch will call handle_mm_fault on every futex contention in the futex_lock_pi path. futex_lock_pi() ret = futex_lock_pi_atomic(uaddr, hb, &q.key, &q.pi_state, current, 0); case -EFAULT: goto uaddr_faulted; ... uaddr_faulted: ret = fault_in_user_writeable(uaddr); >> - How about let the archs do their own write permission >> fixup as what I did in my original > Why ? This is generic and will fix all archs at once with generic code > which is a significant improvement in my book and a lot more > maintainable :-) > If the overhead in the futex_lock_pi path is not considerable yes fix it up generally is nice :-) >> "[PATCH 1/1] Fixup write permission of TLB on powerpc e500 core"? >> (I will fix the stupid errors in my original patch if the concept >> is acceptable) >> in this way we could decrease the overhead of handle_mm_fault >> in the path which does not need write permission fixup. > Which overhead ? gup does handle_mm_fault() as well if needed. it does it *if needed*, and this requirement is rare in my opinion. Thanks Shan Hai > What I do is I replace what is arguably an abuse of gup() in the case > where a fixup -is- needed with a dedicated function designed to perform > the said fixup ... and do it properly which gup() didn't :-) > > Cheers, > Ben. > >> Thanks >> Shan Hai >>> Cheers, >>> Ben. >>> >>> diff --git a/include/linux/mm.h b/include/linux/mm.h >>> index 9670f71..1036614 100644 >>> --- a/include/linux/mm.h >>> +++ b/include/linux/mm.h >>> @@ -985,6 +985,8 @@ int get_user_pages(struct task_struct *tsk, struct mm_struct *mm, >>> int get_user_pages_fast(unsigned long start, int nr_pages, int write, >>> struct page **pages); >>> struct page *get_dump_page(unsigned long addr); >>> +extern int fixup_user_fault(struct task_struct *tsk, struct mm_struct *mm, >>> + unsigned long address, unsigned int fault_flags); >>> >>> extern int try_to_release_page(struct page * page, gfp_t gfp_mask); >>> extern void do_invalidatepage(struct page *page, unsigned long offset); >>> diff --git a/kernel/futex.c b/kernel/futex.c >>> index fe28dc2..7a0a4ed 100644 >>> --- a/kernel/futex.c >>> +++ b/kernel/futex.c >>> @@ -355,8 +355,8 @@ static int fault_in_user_writeable(u32 __user *uaddr) >>> int ret; >>> >>> down_read(&mm->mmap_sem); >>> - ret = get_user_pages(current, mm, (unsigned long)uaddr, >>> - 1, 1, 0, NULL, NULL); >>> + ret = fixup_user_fault(current, mm, (unsigned long)uaddr, >>> + FAULT_FLAG_WRITE); >>> up_read(&mm->mmap_sem); >>> >>> return ret< 0 ? ret : 0; >>> diff --git a/mm/memory.c b/mm/memory.c >>> index 40b7531..b967fb0 100644 >>> --- a/mm/memory.c >>> +++ b/mm/memory.c >>> @@ -1815,7 +1815,64 @@ next_page: >>> } >>> EXPORT_SYMBOL(__get_user_pages); >>> >>> -/** >>> +/* >>> + * fixup_user_fault() - manually resolve a user page fault >>> + * @tsk: the task_struct to use for page fault accounting, or >>> + * NULL if faults are not to be recorded. >>> + * @mm: mm_struct of target mm >>> + * @address: user address >>> + * @fault_flags:flags to pass down to handle_mm_fault() >>> + * >>> + * This is meant to be called in the specific scenario where for >>> + * locking reasons we try to access user memory in atomic context >>> + * (within a pagefault_disable() section), this returns -EFAULT, >>> + * and we want to resolve the user fault before trying again. >>> + * >>> + * Typically this is meant to be used by the futex code. >>> + * >>> + * The main difference with get_user_pages() is that this function >>> + * will unconditionally call handle_mm_fault() which will in turn >>> + * perform all the necessary SW fixup of the dirty and young bits >>> + * in the PTE, while handle_mm_fault() only guarantees to update >>> + * these in the struct page. >>> + * >>> + * This is important for some architectures where those bits also >>> + * gate the access permission to the page because their are >>> + * maintained in software. On such architecture, gup() will not >>> + * be enough to make a subsequent access succeed. >>> + * >>> + * This should be called with the mm_sem held for read. >>> + */ >>> +int fixup_user_fault(struct task_struct *tsk, struct mm_struct *mm, >>> + unsigned long address, unsigned int fault_flags) >>> +{ >>> + struct vm_area_struct *vma; >>> + int ret; >>> + >>> + vma = find_extend_vma(mm, address); >>> + if (!vma || address< vma->vm_start) >>> + return -EFAULT; >>> + >>> + ret = handle_mm_fault(mm, vma, address, fault_flags); >>> + if (ret& VM_FAULT_ERROR) { >>> + if (ret& VM_FAULT_OOM) >>> + return -ENOMEM; >>> + if (ret& (VM_FAULT_HWPOISON | VM_FAULT_HWPOISON_LARGE)) >>> + return -EHWPOISON; >>> + if (ret& VM_FAULT_SIGBUS) >>> + return -EFAULT; >>> + BUG(); >>> + } >>> + if (tsk) { >>> + if (ret& VM_FAULT_MAJOR) >>> + tsk->maj_flt++; >>> + else >>> + tsk->min_flt++; >>> + } >>> + return 0; >>> +} >>> + >>> +/* >>> * get_user_pages() - pin user pages in memory >>> * @tsk: the task_struct to use for page fault accounting, or >>> * NULL if faults are not to be recorded. >>> >>> >> -- >> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in >> the body of a message to majordomo@vger.kernel.org >> More majordomo info at http://vger.kernel.org/majordomo-info.html >> Please read the FAQ at http://www.tux.org/lkml/ > From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-qw0-f51.google.com (mail-qw0-f51.google.com [209.85.216.51]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (Client CN "smtp.gmail.com", Issuer "Google Internet Authority" (verified OK)) by ozlabs.org (Postfix) with ESMTPS id 088F6B6F75 for ; Tue, 19 Jul 2011 15:36:05 +1000 (EST) Received: by qwf7 with SMTP id 7so2040992qwf.38 for ; Mon, 18 Jul 2011 22:36:01 -0700 (PDT) Message-ID: <4E25185F.1050605@gmail.com> Date: Tue, 19 Jul 2011 13:38:39 +0800 From: Shan Hai MIME-Version: 1.0 To: Benjamin Herrenschmidt Subject: Re: [RFC/PATCH] mm/futex: Fix futex writes on archs with SW tracking of dirty & young References: <1310717238-13857-1-git-send-email-haishan.bai@gmail.com> <1310717238-13857-2-git-send-email-haishan.bai@gmail.com> <1310725418.2586.309.camel@twins> <4E21A526.8010904@gmail.com> <1310860194.25044.17.camel@pasglop> <4b337921-d430-4b63-bc36-ad31753cf800@email.android.com> <1310912990.25044.203.camel@pasglop> <1310944453.25044.262.camel@pasglop> <1310961691.25044.274.camel@pasglop> <4E23D728.7090406@gmail.com> <1310972462.25044.292.camel@pasglop> <4E23E02C.8090401@gmail.com> <1310974591.25044.298.camel@pasglop> <4E24FA51.70602@gmail.com> <1311049762.25044.392.camel@pasglop> <4E251365.9090004@gmail.com> <1311053063.25044.397.camel@pasglop> In-Reply-To: <1311053063.25044.397.camel@pasglop> Content-Type: text/plain; charset=UTF-8; format=flowed Cc: tony.luck@intel.com, Peter Zijlstra , Peter Zijlstra , linux-kernel@vger.kernel.org, cmetcalf@tilera.com, dhowells@redhat.com, paulus@samba.org, tglx@linutronix.de, walken@google.com, linuxppc-dev@lists.ozlabs.org, akpm@linux-foundation.org List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On 07/19/2011 01:24 PM, Benjamin Herrenschmidt wrote: > On Tue, 2011-07-19 at 13:17 +0800, Shan Hai wrote: > >> The patch works, but I have certain confusions, >> - Do we want to handle_mm_fault on each futex_lock_pi >> even though in most cases there is no write permission >> fixup's needed? > Don't we only ever call this when futex_atomic_op_inuser() failed ? > Which means a fixup -is- needed .... The fast path is still there. > What you said is another path, that is futex_wake_op(), but what about futex_lock_pi in which my test case failed? your patch will call handle_mm_fault on every futex contention in the futex_lock_pi path. futex_lock_pi() ret = futex_lock_pi_atomic(uaddr, hb, &q.key, &q.pi_state, current, 0); case -EFAULT: goto uaddr_faulted; ... uaddr_faulted: ret = fault_in_user_writeable(uaddr); >> - How about let the archs do their own write permission >> fixup as what I did in my original > Why ? This is generic and will fix all archs at once with generic code > which is a significant improvement in my book and a lot more > maintainable :-) > If the overhead in the futex_lock_pi path is not considerable yes fix it up generally is nice :-) >> "[PATCH 1/1] Fixup write permission of TLB on powerpc e500 core"? >> (I will fix the stupid errors in my original patch if the concept >> is acceptable) >> in this way we could decrease the overhead of handle_mm_fault >> in the path which does not need write permission fixup. > Which overhead ? gup does handle_mm_fault() as well if needed. it does it *if needed*, and this requirement is rare in my opinion. Thanks Shan Hai > What I do is I replace what is arguably an abuse of gup() in the case > where a fixup -is- needed with a dedicated function designed to perform > the said fixup ... and do it properly which gup() didn't :-) > > Cheers, > Ben. > >> Thanks >> Shan Hai >>> Cheers, >>> Ben. >>> >>> diff --git a/include/linux/mm.h b/include/linux/mm.h >>> index 9670f71..1036614 100644 >>> --- a/include/linux/mm.h >>> +++ b/include/linux/mm.h >>> @@ -985,6 +985,8 @@ int get_user_pages(struct task_struct *tsk, struct mm_struct *mm, >>> int get_user_pages_fast(unsigned long start, int nr_pages, int write, >>> struct page **pages); >>> struct page *get_dump_page(unsigned long addr); >>> +extern int fixup_user_fault(struct task_struct *tsk, struct mm_struct *mm, >>> + unsigned long address, unsigned int fault_flags); >>> >>> extern int try_to_release_page(struct page * page, gfp_t gfp_mask); >>> extern void do_invalidatepage(struct page *page, unsigned long offset); >>> diff --git a/kernel/futex.c b/kernel/futex.c >>> index fe28dc2..7a0a4ed 100644 >>> --- a/kernel/futex.c >>> +++ b/kernel/futex.c >>> @@ -355,8 +355,8 @@ static int fault_in_user_writeable(u32 __user *uaddr) >>> int ret; >>> >>> down_read(&mm->mmap_sem); >>> - ret = get_user_pages(current, mm, (unsigned long)uaddr, >>> - 1, 1, 0, NULL, NULL); >>> + ret = fixup_user_fault(current, mm, (unsigned long)uaddr, >>> + FAULT_FLAG_WRITE); >>> up_read(&mm->mmap_sem); >>> >>> return ret< 0 ? ret : 0; >>> diff --git a/mm/memory.c b/mm/memory.c >>> index 40b7531..b967fb0 100644 >>> --- a/mm/memory.c >>> +++ b/mm/memory.c >>> @@ -1815,7 +1815,64 @@ next_page: >>> } >>> EXPORT_SYMBOL(__get_user_pages); >>> >>> -/** >>> +/* >>> + * fixup_user_fault() - manually resolve a user page fault >>> + * @tsk: the task_struct to use for page fault accounting, or >>> + * NULL if faults are not to be recorded. >>> + * @mm: mm_struct of target mm >>> + * @address: user address >>> + * @fault_flags:flags to pass down to handle_mm_fault() >>> + * >>> + * This is meant to be called in the specific scenario where for >>> + * locking reasons we try to access user memory in atomic context >>> + * (within a pagefault_disable() section), this returns -EFAULT, >>> + * and we want to resolve the user fault before trying again. >>> + * >>> + * Typically this is meant to be used by the futex code. >>> + * >>> + * The main difference with get_user_pages() is that this function >>> + * will unconditionally call handle_mm_fault() which will in turn >>> + * perform all the necessary SW fixup of the dirty and young bits >>> + * in the PTE, while handle_mm_fault() only guarantees to update >>> + * these in the struct page. >>> + * >>> + * This is important for some architectures where those bits also >>> + * gate the access permission to the page because their are >>> + * maintained in software. On such architecture, gup() will not >>> + * be enough to make a subsequent access succeed. >>> + * >>> + * This should be called with the mm_sem held for read. >>> + */ >>> +int fixup_user_fault(struct task_struct *tsk, struct mm_struct *mm, >>> + unsigned long address, unsigned int fault_flags) >>> +{ >>> + struct vm_area_struct *vma; >>> + int ret; >>> + >>> + vma = find_extend_vma(mm, address); >>> + if (!vma || address< vma->vm_start) >>> + return -EFAULT; >>> + >>> + ret = handle_mm_fault(mm, vma, address, fault_flags); >>> + if (ret& VM_FAULT_ERROR) { >>> + if (ret& VM_FAULT_OOM) >>> + return -ENOMEM; >>> + if (ret& (VM_FAULT_HWPOISON | VM_FAULT_HWPOISON_LARGE)) >>> + return -EHWPOISON; >>> + if (ret& VM_FAULT_SIGBUS) >>> + return -EFAULT; >>> + BUG(); >>> + } >>> + if (tsk) { >>> + if (ret& VM_FAULT_MAJOR) >>> + tsk->maj_flt++; >>> + else >>> + tsk->min_flt++; >>> + } >>> + return 0; >>> +} >>> + >>> +/* >>> * get_user_pages() - pin user pages in memory >>> * @tsk: the task_struct to use for page fault accounting, or >>> * NULL if faults are not to be recorded. >>> >>> >> -- >> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in >> the body of a message to majordomo@vger.kernel.org >> More majordomo info at http://vger.kernel.org/majordomo-info.html >> Please read the FAQ at http://www.tux.org/lkml/ >