linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: "Uladzislau Rezki (Sony)" <urezki@gmail.com>
To: linux-mm@kvack.org, Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@kernel.org>, Baoquan He <bhe@redhat.com>,
	LKML <linux-kernel@vger.kernel.org>,
	Uladzislau Rezki <urezki@gmail.com>,
	Michal Hocko <mhocko@suse.com>
Subject: [PATCH v3 05/10] mm/vmalloc: Handle non-blocking GFP in __vmalloc_area_node()
Date: Wed,  1 Oct 2025 21:26:41 +0200	[thread overview]
Message-ID: <20251001192647.195204-6-urezki@gmail.com> (raw)
In-Reply-To: <20251001192647.195204-1-urezki@gmail.com>

Make __vmalloc_area_node() respect non-blocking GFP masks such
as GFP_ATOMIC and GFP_NOWAIT.

- Add memalloc_apply_gfp_scope()/memalloc_restore_scope()
  helpers to apply a proper scope.
- Apply memalloc_apply_gfp_scope()/memalloc_restore_scope()
  around vmap_pages_range() for page table setup.
- Set "nofail" to false if a non-blocking mask is used, as
  they are mutually exclusive.

This is particularly important for page table allocations that
internally use GFP_PGTABLE_KERNEL, which may sleep unless such
scope restrictions are applied. For example:

<snip>
__pte_alloc_kernel()
  pte_alloc_one_kernel(&init_mm);
    pagetable_alloc_noprof(GFP_PGTABLE_KERNEL & ~__GFP_HIGHMEM, 0);
<snip>

Note: in most cases, PTE entries are established only up to the
level required by current vmap space usage, meaning the page tables
are typically fully populated during the mapping process.

Acked-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: Baoquan He <bhe@redhat.com>
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
---
 include/linux/vmalloc.h |  2 ++
 mm/vmalloc.c            | 52 +++++++++++++++++++++++++++++++++--------
 2 files changed, 44 insertions(+), 10 deletions(-)

diff --git a/include/linux/vmalloc.h b/include/linux/vmalloc.h
index 1e43181369f1..e8e94f90d686 100644
--- a/include/linux/vmalloc.h
+++ b/include/linux/vmalloc.h
@@ -332,4 +332,6 @@ bool vmalloc_dump_obj(void *object);
 static inline bool vmalloc_dump_obj(void *object) { return false; }
 #endif
 
+unsigned int memalloc_apply_gfp_scope(gfp_t gfp_mask);
+void memalloc_restore_scope(unsigned int flags);
 #endif /* _LINUX_VMALLOC_H */
diff --git a/mm/vmalloc.c b/mm/vmalloc.c
index 9e29dd767c41..d8bcd87239b5 100644
--- a/mm/vmalloc.c
+++ b/mm/vmalloc.c
@@ -3716,6 +3716,42 @@ static void defer_vm_area_cleanup(struct vm_struct *area)
 		schedule_work(&cleanup_vm_area);
 }
 
+/*
+ * Page tables allocations ignore external GFP. Enforces it by
+ * the memalloc scope API. It is used by vmalloc internals and
+ * KASAN shadow population only.
+ *
+ * GFP to scope mapping:
+ *
+ * non-blocking (no __GFP_DIRECT_RECLAIM) - memalloc_noreclaim_save()
+ * GFP_NOFS - memalloc_nofs_save()
+ * GFP_NOIO - memalloc_noio_save()
+ *
+ * Returns a flag cookie to pair with restore.
+ */
+unsigned int
+memalloc_apply_gfp_scope(gfp_t gfp_mask)
+{
+	unsigned int flags = 0;
+
+	if (!gfpflags_allow_blocking(gfp_mask))
+		flags = memalloc_noreclaim_save();
+	else if ((gfp_mask & (__GFP_FS | __GFP_IO)) == __GFP_IO)
+		flags = memalloc_nofs_save();
+	else if ((gfp_mask & (__GFP_FS | __GFP_IO)) == 0)
+		flags = memalloc_noio_save();
+
+	/* 0 - no scope applied. */
+	return flags;
+}
+
+void
+memalloc_restore_scope(unsigned int flags)
+{
+	if (flags)
+		memalloc_flags_restore(flags);
+}
+
 static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask,
 				 pgprot_t prot, unsigned int page_shift,
 				 int node)
@@ -3732,6 +3768,10 @@ static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask,
 
 	array_size = (unsigned long)nr_small_pages * sizeof(struct page *);
 
+	/* __GFP_NOFAIL and "noblock" flags are mutually exclusive. */
+	if (!gfpflags_allow_blocking(gfp_mask))
+		nofail = false;
+
 	if (!(gfp_mask & (GFP_DMA | GFP_DMA32)))
 		gfp_mask |= __GFP_HIGHMEM;
 
@@ -3797,22 +3837,14 @@ static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask,
 	 * page tables allocations ignore external gfp mask, enforce it
 	 * by the scope API
 	 */
-	if ((gfp_mask & (__GFP_FS | __GFP_IO)) == __GFP_IO)
-		flags = memalloc_nofs_save();
-	else if ((gfp_mask & (__GFP_FS | __GFP_IO)) == 0)
-		flags = memalloc_noio_save();
-
+	flags = memalloc_apply_gfp_scope(gfp_mask);
 	do {
 		ret = vmap_pages_range(addr, addr + size, prot, area->pages,
 			page_shift);
 		if (nofail && (ret < 0))
 			schedule_timeout_uninterruptible(1);
 	} while (nofail && (ret < 0));
-
-	if ((gfp_mask & (__GFP_FS | __GFP_IO)) == __GFP_IO)
-		memalloc_nofs_restore(flags);
-	else if ((gfp_mask & (__GFP_FS | __GFP_IO)) == 0)
-		memalloc_noio_restore(flags);
+	memalloc_restore_scope(flags);
 
 	if (ret < 0) {
 		warn_alloc(gfp_mask, NULL,
-- 
2.47.3



  parent reply	other threads:[~2025-10-01 19:27 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-10-01 19:26 [PATCH v3 00/10] __vmalloc()/kvmalloc() and no-block support(v3) Uladzislau Rezki (Sony)
2025-10-01 19:26 ` [PATCH v3 01/10] lib/test_vmalloc: add no_block_alloc_test case Uladzislau Rezki (Sony)
2025-10-01 19:26 ` [PATCH v3 02/10] lib/test_vmalloc: Remove xfail condition check Uladzislau Rezki (Sony)
2025-10-01 19:26 ` [PATCH v3 03/10] mm/vmalloc: Support non-blocking GFP flags in alloc_vmap_area() Uladzislau Rezki (Sony)
2025-10-01 19:26 ` [PATCH v3 04/10] mm/vmalloc: Defer freeing partly initialized vm_struct Uladzislau Rezki (Sony)
2025-10-01 19:26 ` Uladzislau Rezki (Sony) [this message]
2025-10-01 19:26 ` [PATCH v3 06/10] mm/kasan: Support non-blocking GFP in kasan_populate_vmalloc() Uladzislau Rezki (Sony)
2025-10-01 19:26 ` [PATCH v3 07/10] kmsan: Remove hard-coded GFP_KERNEL flags Uladzislau Rezki (Sony)
2025-10-01 19:26 ` [PATCH v3 08/10] mm: Skip might_alloc() warnings when PF_MEMALLOC is set Uladzislau Rezki (Sony)
2025-10-04  3:53   ` Baoquan He
2025-10-01 19:26 ` [PATCH v3 09/10] mm/vmalloc: Update __vmalloc_node_range() documentation Uladzislau Rezki (Sony)
2025-10-04  4:11   ` Baoquan He
2025-10-04  5:02     ` Baoquan He
2025-10-06 10:06       ` Uladzislau Rezki
2025-10-06 13:17         ` Baoquan He
2025-10-07  9:42           ` Uladzislau Rezki
2025-10-07  9:53             ` Baoquan He
2025-10-01 19:26 ` [PATCH v3 10/10] mm: kvmalloc: Add non-blocking support for vmalloc Uladzislau Rezki (Sony)
2025-10-03  6:30   ` Michal Hocko
2025-10-04  4:37   ` Baoquan He

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=20251001192647.195204-6-urezki@gmail.com \
    --to=urezki@gmail.com \
    --cc=akpm@linux-foundation.org \
    --cc=bhe@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mhocko@kernel.org \
    --cc=mhocko@suse.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