| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <CAFPAmTRB9vvoNtGbnfLt9PX3jchxhs63F4WbPdi4bjYBrHgQXg@mail.gmail.com>
Date: Sat, 12 Nov 2011 18:01:10 -0500
From: "kautuk.c @samsung.com" <consul.kautuk@...il.com>
To: Russell King <linux@....linux.org.uk>,
Catalin Marinas <catalin.marinas@....com>,
Will Deacon <will.deacon@....com>, Ingo Molnar <mingo@...e.hu>,
Peter Zijlstra <a.p.zijlstra@...llo.nl>
Cc: linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/1] arch/arm/mm/fault.c: Porting OOM changes into __do_page_fault
Sorry, please ignore this patch.
There is one error in it which I discovered through code review:
lockdep will detect a lock imbalance in the up_read() inside do_page_fault
if __do_page_fault returns with VM_FAULT_RETRY set in the fault variable.
This will happen in the fatal signal handling case of __do_page_fault.
I'll post another patch which takes care of this.
On Sat, Nov 12, 2011 at 1:33 PM, Kautuk Consul <consul.kautuk@...il.com> wrote:
> Commits d065bd810b6deb67d4897a14bfe21f8eb526ba99 and
> 37b23e0525d393d48a7d59f870b3bc061a30ccdb introduced changes into
> the x86 pagefault handler for makeing the page fault handler
> retryable as well as killable.
>
> These changes reduce the mmap_sem hold time(for x86), which is crucial
> during OOM killer invocation.
>
> Porting these changes to ARM.
>
> Without these changes, my ARM board encounters many hang and livelock
> scenarios.
> After applying this patch, OOM feature performance improves according to
> my testing.
>
> Motivation for porting these changes:
> ------------------------------------
> Embedded devices such as SMART TVs and SMART phones in the near future
> will have the capability to download and run apps from the internet.
> Due to this, the device user might run some malignant app that
> allocates too much memory.
> In that case, OOM killer performance is very important so that the
> device can free up memory for other apps as well as the kernel.
>
> Signed-off-by: Kautuk Consul <consul.kautuk@...il.com>
> ---
> arch/arm/mm/fault.c | 57 ++++++++++++++++++++++++++++++++++++++------------
> 1 files changed, 43 insertions(+), 14 deletions(-)
>
> diff --git a/arch/arm/mm/fault.c b/arch/arm/mm/fault.c
> index aa33949..f251ec1 100644
> --- a/arch/arm/mm/fault.c
> +++ b/arch/arm/mm/fault.c
> @@ -231,11 +231,15 @@ static inline bool access_error(unsigned int fsr, struct vm_area_struct *vma)
>
> static int __kprobes
> __do_page_fault(struct mm_struct *mm, unsigned long addr, unsigned int fsr,
> - struct task_struct *tsk)
> + struct pt_regs *regs, struct task_struct *tsk)
> {
> struct vm_area_struct *vma;
> int fault;
> + int write = fsr & FSR_WRITE;
> + unsigned int flags = FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE |
> + (write ? FAULT_FLAG_WRITE : 0);
>
> +retry:
> vma = find_vma(mm, addr);
> fault = VM_FAULT_BADMAP;
> if (unlikely(!vma))
> @@ -257,13 +261,44 @@ good_area:
> * If for any reason at all we couldn't handle the fault, make
> * sure we exit gracefully rather than endlessly redo the fault.
> */
> - fault = handle_mm_fault(mm, vma, addr & PAGE_MASK, (fsr & FSR_WRITE) ? FAULT_FLAG_WRITE : 0);
> - if (unlikely(fault & VM_FAULT_ERROR))
> + fault = handle_mm_fault(mm, vma, addr & PAGE_MASK, flags);
> +
> + if (unlikely((fault & VM_FAULT_ERROR)))
> return fault;
> - if (fault & VM_FAULT_MAJOR)
> - tsk->maj_flt++;
> - else
> - tsk->min_flt++;
> +
> + if ((fault & VM_FAULT_RETRY) && fatal_signal_pending(current))
> + return fault;
> +
> + /*
> + * Major/minor page fault accounting is only done on the
> + * initial attempt. If we go through a retry, it is extremely
> + * likely that the page will be found in page cache at that point.
> + */
> + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, addr);
> + if (flags & FAULT_FLAG_ALLOW_RETRY) {
> + if (fault & VM_FAULT_MAJOR) {
> + tsk->maj_flt++;
> + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1,
> + regs, addr);
> + } else {
> + tsk->min_flt++;
> + perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1,
> + regs, addr);
> + }
> + if (fault & VM_FAULT_RETRY) {
> + /* Clear FAULT_FLAG_ALLOW_RETRY to avoid any risk
> + * of starvation. */
> + flags &= ~FAULT_FLAG_ALLOW_RETRY;
> +
> + /* Acquire the mmap_sem again before retrying this
> + * pagefault. This would have been released by
> + * __lock_page_or_retry() in mm/filemap.c. */
> + down_read(&mm->mmap_sem);
> +
> + goto retry;
> + }
> + }
> +
> return fault;
>
> check_stack:
> @@ -320,15 +355,9 @@ do_page_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
> #endif
> }
>
> - fault = __do_page_fault(mm, addr, fsr, tsk);
> + fault = __do_page_fault(mm, addr, fsr, regs, tsk);
> up_read(&mm->mmap_sem);
>
> - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, addr);
> - if (fault & VM_FAULT_MAJOR)
> - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, regs, addr);
> - else if (fault & VM_FAULT_MINOR)
> - perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, regs, addr);
> -
> /*
> * Handle the "normal" case first - VM_FAULT_MAJOR / VM_FAULT_MINOR
> */
> --
> 1.7.5.4
>
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists