From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail190.messagelabs.com (mail190.messagelabs.com [216.82.249.51]) by kanga.kvack.org (Postfix) with ESMTP id AD4BC8D0039 for ; Thu, 3 Feb 2011 16:22:34 -0500 (EST) Received: from hpaq13.eem.corp.google.com (hpaq13.eem.corp.google.com [172.25.149.13]) by smtp-out.google.com with ESMTP id p13LMVek021873 for ; Thu, 3 Feb 2011 13:22:31 -0800 Received: from pvg13 (pvg13.prod.google.com [10.241.210.141]) by hpaq13.eem.corp.google.com with ESMTP id p13LM4p6010399 (version=TLSv1/SSLv3 cipher=RC4-MD5 bits=128 verify=NOT) for ; Thu, 3 Feb 2011 13:22:29 -0800 Received: by pvg13 with SMTP id 13so351815pvg.38 for ; Thu, 03 Feb 2011 13:22:29 -0800 (PST) Date: Thu, 3 Feb 2011 13:22:26 -0800 (PST) From: David Rientjes Subject: Re: [RFC][PATCH 3/6] break out smaps_pte_entry() from smaps_pte_range() In-Reply-To: <20110201003401.95CFBFA6@kernel> Message-ID: References: <20110201003357.D6F0BE0D@kernel> <20110201003401.95CFBFA6@kernel> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-mm@kvack.org List-ID: To: Dave Hansen Cc: linux-kernel@vger.kernel.org, linux-mm@kvack.org, Michael J Wolf , Andrea Arcangeli On Mon, 31 Jan 2011, Dave Hansen wrote: > > We will use smaps_pte_entry() in a moment to handle both small > and transparent large pages. But, we must break it out of > smaps_pte_range() first. > The extraction from smaps_pte_range() looks good. What's the performance impact on very frequent consumers of /proc/pid/smaps, though, as the result of the calls throughout the iteration if smaps_pte_entry() doesn't get inlined (supposedly because you'll be reusing the extracted function again elsewhere)? > > --- > > linux-2.6.git-dave/fs/proc/task_mmu.c | 85 ++++++++++++++++++---------------- > 1 file changed, 46 insertions(+), 39 deletions(-) > > diff -puN fs/proc/task_mmu.c~break-out-smaps_pte_entry fs/proc/task_mmu.c > --- linux-2.6.git/fs/proc/task_mmu.c~break-out-smaps_pte_entry 2011-01-27 11:03:06.761548697 -0800 > +++ linux-2.6.git-dave/fs/proc/task_mmu.c 2011-01-27 11:03:06.773548685 -0800 > @@ -333,56 +333,63 @@ struct mem_size_stats { > u64 pss; > }; > > -static int smaps_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end, > - struct mm_walk *walk) > + > +static void smaps_pte_entry(pte_t ptent, unsigned long addr, > + struct mm_walk *walk) > { > struct mem_size_stats *mss = walk->private; > struct vm_area_struct *vma = mss->vma; > - pte_t *pte, ptent; > - spinlock_t *ptl; > struct page *page; > int mapcount; > > - split_huge_page_pmd(walk->mm, pmd); > - > - pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl); > - for (; addr != end; pte++, addr += PAGE_SIZE) { > - ptent = *pte; > + if (is_swap_pte(ptent)) { > + mss->swap += PAGE_SIZE; > + return; > + } > > - if (is_swap_pte(ptent)) { > - mss->swap += PAGE_SIZE; > - continue; > - } > + if (!pte_present(ptent)) > + return; > > - if (!pte_present(ptent)) > - continue; > + page = vm_normal_page(vma, addr, ptent); > + if (!page) > + return; > + > + if (PageAnon(page)) > + mss->anonymous += PAGE_SIZE; > + > + mss->resident += PAGE_SIZE; > + /* Accumulate the size in pages that have been accessed. */ > + if (pte_young(ptent) || PageReferenced(page)) > + mss->referenced += PAGE_SIZE; > + mapcount = page_mapcount(page); > + if (mapcount >= 2) { > + if (pte_dirty(ptent) || PageDirty(page)) > + mss->shared_dirty += PAGE_SIZE; > + else > + mss->shared_clean += PAGE_SIZE; > + mss->pss += (PAGE_SIZE << PSS_SHIFT) / mapcount; > + } else { > + if (pte_dirty(ptent) || PageDirty(page)) > + mss->private_dirty += PAGE_SIZE; > + else > + mss->private_clean += PAGE_SIZE; > + mss->pss += (PAGE_SIZE << PSS_SHIFT); > + } > +} > > - page = vm_normal_page(vma, addr, ptent); > - if (!page) > - continue; > +static int smaps_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end, > + struct mm_walk *walk) > +{ > + struct mem_size_stats *mss = walk->private; > + struct vm_area_struct *vma = mss->vma; > + pte_t *pte; > + spinlock_t *ptl; > > - if (PageAnon(page)) > - mss->anonymous += PAGE_SIZE; > + split_huge_page_pmd(walk->mm, pmd); > > - mss->resident += PAGE_SIZE; > - /* Accumulate the size in pages that have been accessed. */ > - if (pte_young(ptent) || PageReferenced(page)) > - mss->referenced += PAGE_SIZE; > - mapcount = page_mapcount(page); > - if (mapcount >= 2) { > - if (pte_dirty(ptent) || PageDirty(page)) > - mss->shared_dirty += PAGE_SIZE; > - else > - mss->shared_clean += PAGE_SIZE; > - mss->pss += (PAGE_SIZE << PSS_SHIFT) / mapcount; > - } else { > - if (pte_dirty(ptent) || PageDirty(page)) > - mss->private_dirty += PAGE_SIZE; > - else > - mss->private_clean += PAGE_SIZE; > - mss->pss += (PAGE_SIZE << PSS_SHIFT); > - } > - } > + pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl); > + for (; addr != end; pte++, addr += PAGE_SIZE) > + smaps_pte_entry(*pte, addr, walk); > pte_unmap_unlock(pte - 1, ptl); > cond_resched(); > return 0; > diff -puN mm/huge_memory.c~break-out-smaps_pte_entry mm/huge_memory.c > _ Is there a missing change to mm/huge_memory.c? -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/ Don't email: email@kvack.org