linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Miaohe Lin <linmiaohe@huawei.com>
To: kernel test robot <lkp@intel.com>
Cc: <llvm@lists.linux.dev>, <kbuild-all@lists.01.org>,
	<linux-kernel@vger.kernel.org>,
	Johannes Weiner <hannes@cmpxchg.org>,
	Naoya Horiguchi <naoya.horiguchi@nec.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	Linux Memory Management List <linux-mm@kvack.org>
Subject: Re: [hnaz-mm:master 214/410] mm/madvise.c:632:8: error: call to undeclared function 'is_swapin_error_entry'; ISO C99 and later do not support implicit function declarations
Date: Thu, 28 Apr 2022 18:47:47 +0800	[thread overview]
Message-ID: <6b55693d-24ab-a0bb-5289-8ca917680088@huawei.com> (raw)
In-Reply-To: <b45ac0cd-ba02-c5ab-a91b-1289a7c05120@huawei.com>

[-- Attachment #1: Type: text/plain, Size: 7041 bytes --]

On 2022/4/28 15:48, Miaohe Lin wrote:
> On 2022/4/28 14:54, kernel test robot wrote:
>> tree:   https://github.com/hnaz/linux-mm master
>> head:   bf4803abaa3e9d2fa207c0675a2d2abf0fd44f66
>> commit: fe343091405131f91dcb9230ae7649bb1ee7fe9a [214/410] mm/madvise: free hwpoison and swapin error entry in madvise_free_pte_range
>> config: i386-randconfig-a005-20220425 (https://download.01.org/0day-ci/archive/20220428/202204281452.cEOGcf4p-lkp@intel.com/config)
>> compiler: clang version 15.0.0 (https://github.com/llvm/llvm-project 1cddcfdc3c683b393df1a5c9063252eb60e52818)
>> reproduce (this is a W=1 build):
>>         wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
>>         chmod +x ~/bin/make.cross
>>         # https://github.com/hnaz/linux-mm/commit/fe343091405131f91dcb9230ae7649bb1ee7fe9a
>>         git remote add hnaz-mm https://github.com/hnaz/linux-mm
>>         git fetch --no-tags hnaz-mm master
>>         git checkout fe343091405131f91dcb9230ae7649bb1ee7fe9a
>>         # save the config file
>>         mkdir build_dir && cp config build_dir/.config
>>         COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross W=1 O=build_dir ARCH=i386 SHELL=/bin/bash
>>
>> If you fix the issue, kindly add following tag as appropriate
>> Reported-by: kernel test robot <lkp@intel.com>
>>
>> Note: the hnaz-mm/master HEAD bf4803abaa3e9d2fa207c0675a2d2abf0fd44f66 builds fine.
>>       It only hurts bisectability.
>>
>> All errors (new ones prefixed by >>):
>>
>>>> mm/madvise.c:632:8: error: call to undeclared function 'is_swapin_error_entry'; ISO C99 and later do not support implicit function declarations [-Wimplicit-function-declaration]
>>                                       is_swapin_error_entry(entry)) {
>>                                       ^
>>    1 error generated.
> 
> I browsed the config file and found that CONFIG_SWAP is not set. But when I test it on my env, it works.
> And I'm somewhat confused now:
> 
> is_swapin_error_entry and pte_to_swp_entry are both defined in include/linux/swapops.h. But pte_to_swp_entry
> can be used normally but is_swapin_error_entry breaks the compile.
> 
> Will take a more close look later. Thanks for report!

I'm sorry but I can't reproduce it in my env. It's very kind of you if you can help test the attached patch.
Or should I send this patch directly?

Thanks!

> 
>>
>>
>> vim +/is_swapin_error_entry +632 mm/madvise.c
>>
>>    587	
>>    588	static int madvise_free_pte_range(pmd_t *pmd, unsigned long addr,
>>    589					unsigned long end, struct mm_walk *walk)
>>    590	
>>    591	{
>>    592		struct mmu_gather *tlb = walk->private;
>>    593		struct mm_struct *mm = tlb->mm;
>>    594		struct vm_area_struct *vma = walk->vma;
>>    595		spinlock_t *ptl;
>>    596		pte_t *orig_pte, *pte, ptent;
>>    597		struct page *page;
>>    598		int nr_swap = 0;
>>    599		unsigned long next;
>>    600	
>>    601		next = pmd_addr_end(addr, end);
>>    602		if (pmd_trans_huge(*pmd))
>>    603			if (madvise_free_huge_pmd(tlb, vma, pmd, addr, next))
>>    604				goto next;
>>    605	
>>    606		if (pmd_trans_unstable(pmd))
>>    607			return 0;
>>    608	
>>    609		tlb_change_page_size(tlb, PAGE_SIZE);
>>    610		orig_pte = pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
>>    611		flush_tlb_batched_pending(mm);
>>    612		arch_enter_lazy_mmu_mode();
>>    613		for (; addr != end; pte++, addr += PAGE_SIZE) {
>>    614			ptent = *pte;
>>    615	
>>    616			if (pte_none(ptent))
>>    617				continue;
>>    618			/*
>>    619			 * If the pte has swp_entry, just clear page table to
>>    620			 * prevent swap-in which is more expensive rather than
>>    621			 * (page allocation + zeroing).
>>    622			 */
>>    623			if (!pte_present(ptent)) {
>>    624				swp_entry_t entry;
>>    625	
>>    626				entry = pte_to_swp_entry(ptent);
>>    627				if (!non_swap_entry(entry)) {
>>    628					nr_swap--;
>>    629					free_swap_and_cache(entry);
>>    630					pte_clear_not_present_full(mm, addr, pte, tlb->fullmm);
>>    631				} else if (is_hwpoison_entry(entry) ||
>>  > 632					   is_swapin_error_entry(entry)) {
>>    633					pte_clear_not_present_full(mm, addr, pte, tlb->fullmm);
>>    634				}
>>    635				continue;
>>    636			}
>>    637	
>>    638			page = vm_normal_page(vma, addr, ptent);
>>    639			if (!page)
>>    640				continue;
>>    641	
>>    642			/*
>>    643			 * If pmd isn't transhuge but the page is THP and
>>    644			 * is owned by only this process, split it and
>>    645			 * deactivate all pages.
>>    646			 */
>>    647			if (PageTransCompound(page)) {
>>    648				if (page_mapcount(page) != 1)
>>    649					goto out;
>>    650				get_page(page);
>>    651				if (!trylock_page(page)) {
>>    652					put_page(page);
>>    653					goto out;
>>    654				}
>>    655				pte_unmap_unlock(orig_pte, ptl);
>>    656				if (split_huge_page(page)) {
>>    657					unlock_page(page);
>>    658					put_page(page);
>>    659					orig_pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
>>    660					goto out;
>>    661				}
>>    662				unlock_page(page);
>>    663				put_page(page);
>>    664				orig_pte = pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
>>    665				pte--;
>>    666				addr -= PAGE_SIZE;
>>    667				continue;
>>    668			}
>>    669	
>>    670			VM_BUG_ON_PAGE(PageTransCompound(page), page);
>>    671	
>>    672			if (PageSwapCache(page) || PageDirty(page)) {
>>    673				if (!trylock_page(page))
>>    674					continue;
>>    675				/*
>>    676				 * If page is shared with others, we couldn't clear
>>    677				 * PG_dirty of the page.
>>    678				 */
>>    679				if (page_mapcount(page) != 1) {
>>    680					unlock_page(page);
>>    681					continue;
>>    682				}
>>    683	
>>    684				if (PageSwapCache(page) && !try_to_free_swap(page)) {
>>    685					unlock_page(page);
>>    686					continue;
>>    687				}
>>    688	
>>    689				ClearPageDirty(page);
>>    690				unlock_page(page);
>>    691			}
>>    692	
>>    693			if (pte_young(ptent) || pte_dirty(ptent)) {
>>    694				/*
>>    695				 * Some of architecture(ex, PPC) don't update TLB
>>    696				 * with set_pte_at and tlb_remove_tlb_entry so for
>>    697				 * the portability, remap the pte with old|clean
>>    698				 * after pte clearing.
>>    699				 */
>>    700				ptent = ptep_get_and_clear_full(mm, addr, pte,
>>    701								tlb->fullmm);
>>    702	
>>    703				ptent = pte_mkold(ptent);
>>    704				ptent = pte_mkclean(ptent);
>>    705				set_pte_at(mm, addr, pte, ptent);
>>    706				tlb_remove_tlb_entry(tlb, pte, addr);
>>    707			}
>>    708			mark_page_lazyfree(page);
>>    709		}
>>    710	out:
>>    711		if (nr_swap) {
>>    712			if (current->mm == mm)
>>    713				sync_mm_rss(mm);
>>    714	
>>    715			add_mm_counter(mm, MM_SWAPENTS, nr_swap);
>>    716		}
>>    717		arch_leave_lazy_mmu_mode();
>>    718		pte_unmap_unlock(orig_pte, ptl);
>>    719		cond_resched();
>>    720	next:
>>    721		return 0;
>>    722	}
>>    723	
>>
> 


[-- Attachment #2: 0001-mm-swap-fix-implicit-function-declarations-of-is_swa.patch --]
[-- Type: text/plain, Size: 2037 bytes --]

From c7b3d6e0a15fd4e85a8f7333abb2220a54c052e5 Mon Sep 17 00:00:00 2001
From: Miaohe Lin <linmiaohe@huawei.com>
Date: Thu, 28 Apr 2022 17:21:55 +0800
Subject: [PATCH] mm/swap: fix implicit function declarations of
 is_swapin_error_entry

Kernel test robot reported the error of calling to undeclared function
is_swapin_error_entry. Fix it by wrapping up is_swapin_error_entry with
CONFIG_SWAP.

Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
---
 include/linux/swap.h    | 4 ++++
 include/linux/swapops.h | 7 +++++++
 2 files changed, 11 insertions(+)

diff --git a/include/linux/swap.h b/include/linux/swap.h
index b82c196d8867..f3d44252e477 100644
--- a/include/linux/swap.h
+++ b/include/linux/swap.h
@@ -55,10 +55,14 @@ static inline int current_is_kswapd(void)
  * actions on faults.
  */
 
+#ifdef CONFIG_SWAP
 #define SWP_SWAPIN_ERROR_NUM 1
 #define SWP_SWAPIN_ERROR     (MAX_SWAPFILES + SWP_HWPOISON_NUM + \
 			     SWP_MIGRATION_NUM + SWP_DEVICE_NUM + \
 			     SWP_PTE_MARKER_NUM)
+#else
+#define SWP_SWAPIN_ERROR_NUM 0
+#endif
 /*
  * PTE markers are used to persist information onto PTEs that are mapped with
  * file-backed memories.  As its name "PTE" hints, it should only be applied to
diff --git a/include/linux/swapops.h b/include/linux/swapops.h
index ffc9f19ab6fb..af6e6fb4b8e6 100644
--- a/include/linux/swapops.h
+++ b/include/linux/swapops.h
@@ -108,6 +108,7 @@ static inline void *swp_to_radix_entry(swp_entry_t entry)
 	return xa_mk_value(entry.val);
 }
 
+#ifdef CONFIG_SWAP
 static inline swp_entry_t make_swapin_error_entry(struct page *page)
 {
 	return swp_entry(SWP_SWAPIN_ERROR, page_to_pfn(page));
@@ -117,6 +118,12 @@ static inline int is_swapin_error_entry(swp_entry_t entry)
 {
 	return swp_type(entry) == SWP_SWAPIN_ERROR;
 }
+#else
+static inline int is_swapin_error_entry(swp_entry_t entry)
+{
+	return 0;
+}
+#endif
 
 #if IS_ENABLED(CONFIG_DEVICE_PRIVATE)
 static inline swp_entry_t make_readable_device_private_entry(pgoff_t offset)
-- 
2.23.0


      reply	other threads:[~2022-04-28 10:47 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-04-28  6:54 kernel test robot
2022-04-28  7:48 ` Miaohe Lin
2022-04-28 10:47   ` Miaohe Lin [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=6b55693d-24ab-a0bb-5289-8ca917680088@huawei.com \
    --to=linmiaohe@huawei.com \
    --cc=akpm@linux-foundation.org \
    --cc=hannes@cmpxchg.org \
    --cc=kbuild-all@lists.01.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=lkp@intel.com \
    --cc=llvm@lists.linux.dev \
    --cc=naoya.horiguchi@nec.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox