Skip to contentSkip to conversation

fs/userfaultfd.c:337:13: error: implicit declaration of function 'pte_none_mostly'

PATCH kernel test robot 2 participants 2 messages Apr 23, 2022

Subsystems (2)
lore
Share & tools
lore

kernel test robot

author build bot
#0 lore raw
tree:   https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
head:   e7d6987e09a328d4a949701db40ef63fbb970670
commit: 70a8da53f7cb7368776f9d2790ecbb4f8bc53ba5 [5085/5845] mm: teach core mm about pte markers
config: x86_64-randconfig-a002-20210928 (https://download.01.org/0day-ci/archive/20220424/[email protected]/config)
compiler: gcc-11 (Debian 11.2.0-20) 11.2.0
reproduce (this is a W=1 build):
        # https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=70a8da53f7cb7368776f9d2790ecbb4f8bc53ba5
        git remote add linux-next https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git
        git fetch --no-tags linux-next master
        git checkout 70a8da53f7cb7368776f9d2790ecbb4f8bc53ba5
        # save the config file
        mkdir build_dir && cp config build_dir/.config
        make W=1 O=build_dir ARCH=x86_64 SHELL=/bin/bash

If you fix the issue, kindly add following tag as appropriate

Note: the linux-next/master HEAD e7d6987e09a328d4a949701db40ef63fbb970670 builds fine.
It may have been fixed somewhere.

All errors (new ones prefixed by >>):

fs/userfaultfd.c: In function 'userfaultfd_must_wait':

>> fs/userfaultfd.c:337:13: error: implicit declaration of function 'pte_none_mostly' [-Werror=implicit-function-declaration]
     337 |         if (pte_none_mostly(*pte))
         |             ^~~~~~~~~~~~~~~

cc1: some warnings being treated as errors

vim +/pte_none_mostly +337 fs/userfaultfd.c

272
273 /*
274 * Verify the pagetables are still not ok after having reigstered into
275 * the fault_pending_wqh to avoid userland having to UFFDIO_WAKE any
276 * userfault that has already been resolved, if userfaultfd_read and
277 * UFFDIO_COPY|ZEROPAGE are being run simultaneously on two different
278 * threads.
279 */
280 static inline bool userfaultfd_must_wait(struct userfaultfd_ctx *ctx,
281 unsigned long address,
282 unsigned long flags,
283 unsigned long reason)
284 {
285 struct mm_struct *mm = ctx->mm;
286 pgd_t *pgd;
287 p4d_t *p4d;
288 pud_t *pud;
289 pmd_t *pmd, _pmd;
290 pte_t *pte;
291 bool ret = true;
292
293 mmap_assert_locked(mm);
294
295 pgd = pgd_offset(mm, address);
296 if (!pgd_present(*pgd))
297 goto out;
298 p4d = p4d_offset(pgd, address);
299 if (!p4d_present(*p4d))
300 goto out;
301 pud = pud_offset(p4d, address);
302 if (!pud_present(*pud))
303 goto out;
304 pmd = pmd_offset(pud, address);
305 /*
306 * READ_ONCE must function as a barrier with narrower scope
307 * and it must be equivalent to:
308 * _pmd = *pmd; barrier();
309 *
310 * This is to deal with the instability (as in
311 * pmd_trans_unstable) of the pmd.
312 */
313 _pmd = READ_ONCE(*pmd);
314 if (pmd_none(_pmd))
315 goto out;
316
317 ret = false;
318 if (!pmd_present(_pmd))
319 goto out;
320
321 if (pmd_trans_huge(_pmd)) {
322 if (!pmd_write(_pmd) && (reason & VM_UFFD_WP))
323 ret = true;
324 goto out;
325 }
326
327 /*
328 * the pmd is stable (as in !pmd_trans_unstable) so we can re-read it
329 * and use the standard pte_offset_map() instead of parsing _pmd.
330 */
331 pte = pte_offset_map(pmd, address);
332 /*
333 * Lockless access: we're in a wait_event so it's ok if it
334 * changes under us. PTE markers should be handled the same as none
335 * ptes here.
336 */

 > 337		if (pte_none_mostly(*pte))

338 ret = true;
339 if (!pte_write(*pte) && (reason & VM_UFFD_WP))
340 ret = true;
341 pte_unmap(pte);
342
343 out:
344 return ret;
345 }
346

Signature · 2 lines
0-DAY CI Kernel Test Service
https://01.org/lkp
#1 lore raw
Quoted 20 lines · kernel test robot
tree:   https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
head:   e7d6987e09a328d4a949701db40ef63fbb970670
commit: 70a8da53f7cb7368776f9d2790ecbb4f8bc53ba5 [5085/5845] mm: teach core mm about pte markers
config: x86_64-randconfig-a002-20210928 (https://download.01.org/0day-ci/archive/20220424/202204240320.TGDuGcsL-lkp(a)intel.com/config)
compiler: gcc-11 (Debian 11.2.0-20) 11.2.0
reproduce (this is a W=1 build):
        # https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=70a8da53f7cb7368776f9d2790ecbb4f8bc53ba5
        git remote add linux-next https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git
        git fetch --no-tags linux-next master
        git checkout 70a8da53f7cb7368776f9d2790ecbb4f8bc53ba5
        # save the config file
        mkdir build_dir && cp config build_dir/.config
        make W=1 O=build_dir ARCH=x86_64 SHELL=/bin/bash

If you fix the issue, kindly add following tag as appropriate

Note: the linux-next/master HEAD e7d6987e09a328d4a949701db40ef63fbb970670 builds fine.
It may have been fixed somewhere.

Quoted 7 lines

All errors (new ones prefixed by >>):

fs/userfaultfd.c: In function 'userfaultfd_must_wait':

Quoted 1 line
Quoted 1 line

fs/userfaultfd.c:337:13: error: implicit declaration of function 'pte_none_mostly' [-Werror=implicit-function-declaration]

     337 |         if (pte_none_mostly(*pte))
         |             ^~~~~~~~~~~~~~~

cc1: some warnings being treated as errors

Ah, userfaultfd.c may need explicit inclusion for swapops.h now.. Expected fix attached to be squashed into "mm: Teach core mm about pte markers", I think.

Thanks,

Signature · 1 line
Peter Xu

From 009572e59f7d72c5085676b0506f0b269f9ef334 Mon Sep 17 00:00:00 2001
From: Peter Xu <[email protected]>
Date: Sat, 23 Apr 2022 16:39:15 -0400
Subject: [PATCH] fixup! mm: Teach core mm about pte markers
Content-type: text/plain

---
 fs/userfaultfd.c | 1 +
 1 file changed, 1 insertion(+)
fs/userfaultfd.c +1 -0
index fb45522a2b44..e943370107d0 100644
@@ -29,6 +29,7 @@
 #include <linux/ioctl.h>
 #include <linux/security.h>
 #include <linux/hugetlb.h>
+#include <linux/swapops.h>
 
 int sysctl_unprivileged_userfaultfd __read_mostly;
 
Signature · 1 line
2.32.0