From nobody Sat Feb 7 20:39:39 2026 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-1.web.codeaurora.org [10.30.226.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 1CD432DCF58 for ; Mon, 19 Jan 2026 22:07:16 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=10.30.226.201 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1768860437; cv=none; b=ne80pr4E4UDYkmYSe/E5SjmsW4NvH0uGNrNQcXYXoWuF0H9NCc5Zriiu2ow9kW0LRwueC9D5uRz4raBseHs/auUH2Nf9FcfDmakpdejBesJsCWwbsCYE+OvPJ/TJdd/K0Cd/oibScXnCbNZ234a57VygfazLUHk4p2Gkrcd2U3I= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1768860437; c=relaxed/simple; bh=6S73vOd5UI/VdrU0A44F+umXZrxl6lMbP+XMOUHhvwk=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=S2YwPBOefq9ILTcEeXYAr7+dMtv7yXd4wFq0vxwFo6GQlSI8/XOFPOp4NgE9Joh2hqx0vUl4NI/nt4q88SatFdH3z9hNzQf7++ZHPTIUeHYNy9AmhQtIoBPny62tAYRnSBk1cOwzMTiGgaMpJLkMRBNded1kimRGSGSDvCVG3xo= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b=kLvk1sih; arc=none smtp.client-ip=10.30.226.201 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b="kLvk1sih" Received: by smtp.kernel.org (Postfix) with ESMTPSA id F09C2C19422; Mon, 19 Jan 2026 22:07:13 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1768860436; bh=6S73vOd5UI/VdrU0A44F+umXZrxl6lMbP+XMOUHhvwk=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=kLvk1sihBmhLNqGYOrWl1P7Kw1zbXh0a/WibGYGBJWG/wR4GOOs4Du3QtEXlr5cUW MWcLRK0eayhgnz/5gbIG4C+ZvxXLfUht3UJ71f/wLDIHEJR/wXiwxh5hKWUsbgY8OD 6TT9tuseGPWEur3xLRTn9vz+Cuq6WQljYgkyqjJUrN9bi89j8FEH3FBccDYsCIt6Ka dmbdHf/UL9h/6JbGJECapaG/DQ89XY2TKGYIQRPAUfiiFZKc5WzlZGxXDhLcnQpFv2 5uoA5ur8LNLZYJT/D9TVGiUCSew3hqT07U5YyvJ2jicSxHHZsPkpIBlyHW1EPlBzs6 SufBVr7O0WNFw== From: "David Hildenbrand (Red Hat)" To: linux-kernel@vger.kernel.org Cc: linux-mm@kvack.org, "David Hildenbrand (Red Hat)" , Andrew Morton , Lorenzo Stoakes , "Liam R. Howlett" , Vlastimil Babka , Mike Rapoport , Suren Baghdasaryan , Michal Hocko , Qi Zheng Subject: [PATCH v1 1/2] mm: move pte table reclaim code to memory.c Date: Mon, 19 Jan 2026 23:07:07 +0100 Message-ID: <20260119220708.3438514-2-david@kernel.org> X-Mailer: git-send-email 2.52.0 In-Reply-To: <20260119220708.3438514-1-david@kernel.org> References: <20260119220708.3438514-1-david@kernel.org> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" The pte-table reclaim code is only called from memory.c, while zapping pages, and it better also stays that way in the long run. If we ever have to call it from other files, we should expose proper high-level helpers for zapping if the existing helpers are not good enough. So, let's move the code over (it's not a lot) and slightly clean it up a bit by: - Renaming the functions. - Dropping the "Check if it is empty PTE page" comment, which is now self-explaining given the function name. - Making zap_pte_table_if_empty() return whether zapping worked so the caller can free it. - Adding a comment in pte_table_reclaim_possible(). - Inlining free_pte() in the last remaining user. - In zap_empty_pte_table(), switch from pmdp_get_lcokless() to pmd_clear(), we are holding the PMD PT lock. By moving the code over, compilers can also easily figure out when zap_empty_pte_table() does not initialize the pmdval variable, avoiding false-positive warnings about the variable possibly not being initialized. Signed-off-by: David Hildenbrand (Red Hat) Reviewed-by: Qi Zheng --- MAINTAINERS | 1 - mm/Makefile | 1 - mm/internal.h | 18 ------------- mm/memory.c | 68 +++++++++++++++++++++++++++++++++++++++++----- mm/pt_reclaim.c | 72 ------------------------------------------------- 5 files changed, 62 insertions(+), 98 deletions(-) delete mode 100644 mm/pt_reclaim.c diff --git a/MAINTAINERS b/MAINTAINERS index 11720728d92f2..28e8e28bca3e5 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -16692,7 +16692,6 @@ R: Shakeel Butt R: Lorenzo Stoakes L: linux-mm@kvack.org S: Maintained -F: mm/pt_reclaim.c F: mm/vmscan.c F: mm/workingset.c =20 diff --git a/mm/Makefile b/mm/Makefile index 0d85b10dbdde4..53ca5d4b1929b 100644 --- a/mm/Makefile +++ b/mm/Makefile @@ -146,5 +146,4 @@ obj-$(CONFIG_GENERIC_IOREMAP) +=3D ioremap.o obj-$(CONFIG_SHRINKER_DEBUG) +=3D shrinker_debug.o obj-$(CONFIG_EXECMEM) +=3D execmem.o obj-$(CONFIG_TMPFS_QUOTA) +=3D shmem_quota.o -obj-$(CONFIG_PT_RECLAIM) +=3D pt_reclaim.o obj-$(CONFIG_LAZY_MMU_MODE_KUNIT_TEST) +=3D tests/lazy_mmu_mode_kunit.o diff --git a/mm/internal.h b/mm/internal.h index 9508dbaf47cd4..ef71a1d9991f2 100644 --- a/mm/internal.h +++ b/mm/internal.h @@ -1745,24 +1745,6 @@ int walk_page_range_debug(struct mm_struct *mm, unsi= gned long start, unsigned long end, const struct mm_walk_ops *ops, pgd_t *pgd, void *private); =20 -/* pt_reclaim.c */ -bool try_get_and_clear_pmd(struct mm_struct *mm, pmd_t *pmd, pmd_t *pmdval= ); -void free_pte(struct mm_struct *mm, unsigned long addr, struct mmu_gather = *tlb, - pmd_t pmdval); -void try_to_free_pte(struct mm_struct *mm, pmd_t *pmd, unsigned long addr, - struct mmu_gather *tlb); - -#ifdef CONFIG_PT_RECLAIM -bool reclaim_pt_is_enabled(unsigned long start, unsigned long end, - struct zap_details *details); -#else -static inline bool reclaim_pt_is_enabled(unsigned long start, unsigned lon= g end, - struct zap_details *details) -{ - return false; -} -#endif /* CONFIG_PT_RECLAIM */ - void dup_mm_exe_file(struct mm_struct *mm, struct mm_struct *oldmm); int dup_mmap(struct mm_struct *mm, struct mm_struct *oldmm); =20 diff --git a/mm/memory.c b/mm/memory.c index f2e9e05388743..c3055b2577c27 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -1824,11 +1824,68 @@ static inline int do_zap_pte_range(struct mmu_gathe= r *tlb, return nr; } =20 +static bool pte_table_reclaim_possible(unsigned long start, unsigned long = end, + struct zap_details *details) +{ + if (!IS_ENABLED(CONFIG_PT_RECLAIM)) + return false; + /* Only zap if we are allowed to and cover the full page table. */ + return details && details->reclaim_pt && (end - start >=3D PMD_SIZE); +} + +static bool zap_empty_pte_table(struct mm_struct *mm, pmd_t *pmd, pmd_t *p= mdval) +{ + spinlock_t *pml =3D pmd_lockptr(mm, pmd); + + if (!spin_trylock(pml)) + return false; + + *pmdval =3D pmdp_get(pmd); + pmd_clear(pmd); + spin_unlock(pml); + return true; +} + +static bool zap_pte_table_if_empty(struct mm_struct *mm, pmd_t *pmd, + unsigned long addr, pmd_t *pmdval) +{ + spinlock_t *pml, *ptl =3D NULL; + pte_t *start_pte, *pte; + int i; + + pml =3D pmd_lock(mm, pmd); + start_pte =3D pte_offset_map_rw_nolock(mm, pmd, addr, pmdval, &ptl); + if (!start_pte) + goto out_ptl; + if (ptl !=3D pml) + spin_lock_nested(ptl, SINGLE_DEPTH_NESTING); + + for (i =3D 0, pte =3D start_pte; i < PTRS_PER_PTE; i++, pte++) { + if (!pte_none(ptep_get(pte))) + goto out_ptl; + } + pte_unmap(start_pte); + + pmd_clear(pmd); + + if (ptl !=3D pml) + spin_unlock(ptl); + spin_unlock(pml); + return true; +out_ptl: + if (start_pte) + pte_unmap_unlock(start_pte, ptl); + if (ptl !=3D pml) + spin_unlock(pml); + return false; +} + static unsigned long zap_pte_range(struct mmu_gather *tlb, struct vm_area_struct *vma, pmd_t *pmd, unsigned long addr, unsigned long end, struct zap_details *details) { + bool can_reclaim_pt =3D pte_table_reclaim_possible(addr, end, details); bool force_flush =3D false, force_break =3D false; struct mm_struct *mm =3D tlb->mm; int rss[NR_MM_COUNTERS]; @@ -1837,7 +1894,6 @@ static unsigned long zap_pte_range(struct mmu_gather = *tlb, pte_t *pte; pmd_t pmdval; unsigned long start =3D addr; - bool can_reclaim_pt =3D reclaim_pt_is_enabled(start, end, details); bool direct_reclaim =3D true; int nr; =20 @@ -1878,7 +1934,7 @@ static unsigned long zap_pte_range(struct mmu_gather = *tlb, * from being repopulated by another thread. */ if (can_reclaim_pt && direct_reclaim && addr =3D=3D end) - direct_reclaim =3D try_get_and_clear_pmd(mm, pmd, &pmdval); + direct_reclaim =3D zap_empty_pte_table(mm, pmd, &pmdval); =20 add_mm_rss_vec(mm, rss); lazy_mmu_mode_disable(); @@ -1907,10 +1963,10 @@ static unsigned long zap_pte_range(struct mmu_gathe= r *tlb, } =20 if (can_reclaim_pt) { - if (direct_reclaim) - free_pte(mm, start, tlb, pmdval); - else - try_to_free_pte(mm, pmd, start, tlb); + if (direct_reclaim || zap_pte_table_if_empty(mm, pmd, start, &pmdval)) { + pte_free_tlb(tlb, pmd_pgtable(pmdval), addr); + mm_dec_nr_ptes(mm); + } } =20 return addr; diff --git a/mm/pt_reclaim.c b/mm/pt_reclaim.c deleted file mode 100644 index 46771cfff8239..0000000000000 --- a/mm/pt_reclaim.c +++ /dev/null @@ -1,72 +0,0 @@ -// SPDX-License-Identifier: GPL-2.0 -#include -#include - -#include - -#include "internal.h" - -bool reclaim_pt_is_enabled(unsigned long start, unsigned long end, - struct zap_details *details) -{ - return details && details->reclaim_pt && (end - start >=3D PMD_SIZE); -} - -bool try_get_and_clear_pmd(struct mm_struct *mm, pmd_t *pmd, pmd_t *pmdval) -{ - spinlock_t *pml =3D pmd_lockptr(mm, pmd); - - if (!spin_trylock(pml)) - return false; - - *pmdval =3D pmdp_get_lockless(pmd); - pmd_clear(pmd); - spin_unlock(pml); - - return true; -} - -void free_pte(struct mm_struct *mm, unsigned long addr, struct mmu_gather = *tlb, - pmd_t pmdval) -{ - pte_free_tlb(tlb, pmd_pgtable(pmdval), addr); - mm_dec_nr_ptes(mm); -} - -void try_to_free_pte(struct mm_struct *mm, pmd_t *pmd, unsigned long addr, - struct mmu_gather *tlb) -{ - pmd_t pmdval; - spinlock_t *pml, *ptl =3D NULL; - pte_t *start_pte, *pte; - int i; - - pml =3D pmd_lock(mm, pmd); - start_pte =3D pte_offset_map_rw_nolock(mm, pmd, addr, &pmdval, &ptl); - if (!start_pte) - goto out_ptl; - if (ptl !=3D pml) - spin_lock_nested(ptl, SINGLE_DEPTH_NESTING); - - /* Check if it is empty PTE page */ - for (i =3D 0, pte =3D start_pte; i < PTRS_PER_PTE; i++, pte++) { - if (!pte_none(ptep_get(pte))) - goto out_ptl; - } - pte_unmap(start_pte); - - pmd_clear(pmd); - - if (ptl !=3D pml) - spin_unlock(ptl); - spin_unlock(pml); - - free_pte(mm, addr, tlb, pmdval); - - return; -out_ptl: - if (start_pte) - pte_unmap_unlock(start_pte, ptl); - if (ptl !=3D pml) - spin_unlock(pml); -} --=20 2.52.0 From nobody Sat Feb 7 20:39:39 2026 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-1.web.codeaurora.org [10.30.226.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id A9D9F2FCBE3 for ; Mon, 19 Jan 2026 22:07:20 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=10.30.226.201 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1768860440; cv=none; b=NQIxpxPishopLHwlRwlwh1FAA+u75D5kAsr3BAJJKCqXw4sPpFwiZoqZp4RjZxofFbEgI/9q5+drCmWgYsdYGiRnB6n7fmK4ARvKRTLOOod2GNWJJHZQDYNuYiGxmHhq8YCJg5lfodF6sYJKFYDGDe1UNjyz5dNPWRe1PPBIXoo= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1768860440; c=relaxed/simple; bh=Uz+rgdCACY4xIfDsBJdwJG1Wp19Ieqw85OkYIQtZgKM=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=H2hO5AZJKspbZj/uhnNBvbLHXGVJ3m65TpzDfRuXuXTEfFntJoAqI1icZjACk4QB7tGYTvJlWQdtg+FkDwHLm67pRBQScXVfYXCQmjq5Jrmv6cKPHp/+viU787O+82ygo3XFqi7DfGuthxKpywjk/nQkHxU7DAirQw3in/HXjsA= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b=kJSJA0Ye; arc=none smtp.client-ip=10.30.226.201 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b="kJSJA0Ye" Received: by smtp.kernel.org (Postfix) with ESMTPSA id 59401C116C6; Mon, 19 Jan 2026 22:07:17 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1768860440; bh=Uz+rgdCACY4xIfDsBJdwJG1Wp19Ieqw85OkYIQtZgKM=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=kJSJA0Yex7DNtH457O8X0WqNKsu9ixkcZRM/TTyY4EYQV8NQUNuD4KbD2p9teGs6q d6zn4/v7fCR3N30JmDUlvQAoPdRhzWAr3MUhVV29GSCsI3wSKqhe7Gt56UIAJUEnDB LBBXRpjcYEY2sIz8agcCU5MittvsbydGbHOuHt7kl3NzlEQwj47lukZLwK3/UZLLQv JcEG+KeuJCx2k1mMTADQ61A/ZJcBfKA/QdA8pa2+Z4P9tqBDab8J1AZRx8s/x8ARsw +rc+7Cqs5rCU6MjaI25AqwZmETUFiCZePv3rXlZNU+FrLPTIOi4Ese/gE+sdsMnOoK M05bKsRL3WFpw== From: "David Hildenbrand (Red Hat)" To: linux-kernel@vger.kernel.org Cc: linux-mm@kvack.org, "David Hildenbrand (Red Hat)" , Andrew Morton , Lorenzo Stoakes , "Liam R. Howlett" , Vlastimil Babka , Mike Rapoport , Suren Baghdasaryan , Michal Hocko , Qi Zheng Subject: [PATCH v1 2/2] mm/memory: handle non-split locks correctly in zap_empty_pte_table() Date: Mon, 19 Jan 2026 23:07:08 +0100 Message-ID: <20260119220708.3438514-3-david@kernel.org> X-Mailer: git-send-email 2.52.0 In-Reply-To: <20260119220708.3438514-1-david@kernel.org> References: <20260119220708.3438514-1-david@kernel.org> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" While we handle pte_lockptr() =3D=3D pmd_lockptr() correctly in zap_pte_table_if_empty(), we don't handle it in zap_empty_pte_table(), making the spin_trylock() always fail and forcing us onto the slow path. So let's handle the scenario where pte_lockptr() =3D=3D pmd_lockptr() better, which can only happen if CONFIG_SPLIT_PTE_PTLOCKS is not set. This is only relevant once we unlock CONFIG_PT_RECLAIM on architectures that are not x86-64. Signed-off-by: David Hildenbrand (Red Hat) Reviewed-by: Qi Zheng --- mm/memory.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) diff --git a/mm/memory.c b/mm/memory.c index c3055b2577c27..3852075ea62d4 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -1833,16 +1833,18 @@ static bool pte_table_reclaim_possible(unsigned lon= g start, unsigned long end, return details && details->reclaim_pt && (end - start >=3D PMD_SIZE); } =20 -static bool zap_empty_pte_table(struct mm_struct *mm, pmd_t *pmd, pmd_t *p= mdval) +static bool zap_empty_pte_table(struct mm_struct *mm, pmd_t *pmd, + spinlock_t *ptl, pmd_t *pmdval) { spinlock_t *pml =3D pmd_lockptr(mm, pmd); =20 - if (!spin_trylock(pml)) + if (ptl !=3D pml && !spin_trylock(pml)) return false; =20 *pmdval =3D pmdp_get(pmd); pmd_clear(pmd); - spin_unlock(pml); + if (ptl !=3D pml) + spin_unlock(pml); return true; } =20 @@ -1934,7 +1936,7 @@ static unsigned long zap_pte_range(struct mmu_gather = *tlb, * from being repopulated by another thread. */ if (can_reclaim_pt && direct_reclaim && addr =3D=3D end) - direct_reclaim =3D zap_empty_pte_table(mm, pmd, &pmdval); + direct_reclaim =3D zap_empty_pte_table(mm, pmd, ptl, &pmdval); =20 add_mm_rss_vec(mm, rss); lazy_mmu_mode_disable(); --=20 2.52.0