From mboxrd@z Thu Jan 1 00:00:00 1970 From: Yu Zhao Subject: [PATCH v2 2/3] arm64: mm: don't call page table ctors for init_mm Date: Mon, 18 Feb 2019 16:13:18 -0700 Message-ID: <20190218231319.178224-2-yuzhao@google.com> References: <20190214211642.2200-1-yuzhao@google.com> <20190218231319.178224-1-yuzhao@google.com> Mime-Version: 1.0 Content-Transfer-Encoding: 8bit Return-path: In-Reply-To: <20190218231319.178224-1-yuzhao@google.com> Sender: linux-kernel-owner@vger.kernel.org To: Catalin Marinas , Will Deacon Cc: "Aneesh Kumar K . V" , Andrew Morton , Nick Piggin , Peter Zijlstra , Joel Fernandes , "Kirill A . Shutemov" , Mark Rutland , Ard Biesheuvel , Chintan Pandya , Jun Yao , Laura Abbott , linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org, linux-mm@kvack.org, Yu Zhao List-Id: linux-arch.vger.kernel.org init_mm doesn't require page table lock to be initialized at any level. Add a separate page table allocator for it, and the new one skips page table ctors. The ctors allocate memory when ALLOC_SPLIT_PTLOCKS is set. Not calling them avoids memory leak in case we call pte_free_kernel() on init_mm. Signed-off-by: Yu Zhao --- arch/arm64/mm/mmu.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-) diff --git a/arch/arm64/mm/mmu.c b/arch/arm64/mm/mmu.c index fa7351877af3..e8bf8a6300e8 100644 --- a/arch/arm64/mm/mmu.c +++ b/arch/arm64/mm/mmu.c @@ -370,6 +370,16 @@ static void __create_pgd_mapping(pgd_t *pgdir, phys_addr_t phys, } while (pgdp++, addr = next, addr != end); } +static phys_addr_t pgd_kernel_pgtable_alloc(int shift) +{ + void *ptr = (void *)__get_free_page(PGALLOC_GFP); + BUG_ON(!ptr); + + /* Ensure the zeroed page is visible to the page table walker */ + dsb(ishst); + return __pa(ptr); +} + static phys_addr_t pgd_pgtable_alloc(int shift) { void *ptr = (void *)__get_free_page(PGALLOC_GFP); @@ -591,7 +601,7 @@ static int __init map_entry_trampoline(void) /* Map only the text into the trampoline page table */ memset(tramp_pg_dir, 0, PGD_SIZE); __create_pgd_mapping(tramp_pg_dir, pa_start, TRAMP_VALIAS, PAGE_SIZE, - prot, pgd_pgtable_alloc, 0); + prot, pgd_kernel_pgtable_alloc, 0); /* Map both the text and data into the kernel page table */ __set_fixmap(FIX_ENTRY_TRAMP_TEXT, pa_start, prot); @@ -1067,7 +1077,8 @@ int arch_add_memory(int nid, u64 start, u64 size, struct vmem_altmap *altmap, flags = NO_BLOCK_MAPPINGS | NO_CONT_MAPPINGS; __create_pgd_mapping(swapper_pg_dir, start, __phys_to_virt(start), - size, PAGE_KERNEL, pgd_pgtable_alloc, flags); + size, PAGE_KERNEL, pgd_kernel_pgtable_alloc, + flags); return __add_pages(nid, start >> PAGE_SHIFT, size >> PAGE_SHIFT, altmap, want_memblock); -- 2.21.0.rc0.258.g878e2cd30e-goog From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-it1-f194.google.com ([209.85.166.194]:40334 "EHLO mail-it1-f194.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1730219AbfBRXN1 (ORCPT ); Mon, 18 Feb 2019 18:13:27 -0500 Received: by mail-it1-f194.google.com with SMTP id i2so1890326ite.5 for ; Mon, 18 Feb 2019 15:13:26 -0800 (PST) From: Yu Zhao Subject: [PATCH v2 2/3] arm64: mm: don't call page table ctors for init_mm Date: Mon, 18 Feb 2019 16:13:18 -0700 Message-ID: <20190218231319.178224-2-yuzhao@google.com> In-Reply-To: <20190218231319.178224-1-yuzhao@google.com> References: <20190214211642.2200-1-yuzhao@google.com> <20190218231319.178224-1-yuzhao@google.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Sender: linux-arch-owner@vger.kernel.org List-ID: To: Catalin Marinas , Will Deacon Cc: "Aneesh Kumar K . V" , Andrew Morton , Nick Piggin , Peter Zijlstra , Joel Fernandes , "Kirill A . Shutemov" , Mark Rutland , Ard Biesheuvel , Chintan Pandya , Jun Yao , Laura Abbott , linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org, linux-mm@kvack.org, Yu Zhao Message-ID: <20190218231318.c803pOfRP__bhXpzB7C0gQC3K5fv9qItK1ijXOTceA4@z> init_mm doesn't require page table lock to be initialized at any level. Add a separate page table allocator for it, and the new one skips page table ctors. The ctors allocate memory when ALLOC_SPLIT_PTLOCKS is set. Not calling them avoids memory leak in case we call pte_free_kernel() on init_mm. Signed-off-by: Yu Zhao --- arch/arm64/mm/mmu.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-) diff --git a/arch/arm64/mm/mmu.c b/arch/arm64/mm/mmu.c index fa7351877af3..e8bf8a6300e8 100644 --- a/arch/arm64/mm/mmu.c +++ b/arch/arm64/mm/mmu.c @@ -370,6 +370,16 @@ static void __create_pgd_mapping(pgd_t *pgdir, phys_addr_t phys, } while (pgdp++, addr = next, addr != end); } +static phys_addr_t pgd_kernel_pgtable_alloc(int shift) +{ + void *ptr = (void *)__get_free_page(PGALLOC_GFP); + BUG_ON(!ptr); + + /* Ensure the zeroed page is visible to the page table walker */ + dsb(ishst); + return __pa(ptr); +} + static phys_addr_t pgd_pgtable_alloc(int shift) { void *ptr = (void *)__get_free_page(PGALLOC_GFP); @@ -591,7 +601,7 @@ static int __init map_entry_trampoline(void) /* Map only the text into the trampoline page table */ memset(tramp_pg_dir, 0, PGD_SIZE); __create_pgd_mapping(tramp_pg_dir, pa_start, TRAMP_VALIAS, PAGE_SIZE, - prot, pgd_pgtable_alloc, 0); + prot, pgd_kernel_pgtable_alloc, 0); /* Map both the text and data into the kernel page table */ __set_fixmap(FIX_ENTRY_TRAMP_TEXT, pa_start, prot); @@ -1067,7 +1077,8 @@ int arch_add_memory(int nid, u64 start, u64 size, struct vmem_altmap *altmap, flags = NO_BLOCK_MAPPINGS | NO_CONT_MAPPINGS; __create_pgd_mapping(swapper_pg_dir, start, __phys_to_virt(start), - size, PAGE_KERNEL, pgd_pgtable_alloc, flags); + size, PAGE_KERNEL, pgd_kernel_pgtable_alloc, + flags); return __add_pages(nid, start >> PAGE_SHIFT, size >> PAGE_SHIFT, altmap, want_memblock); -- 2.21.0.rc0.258.g878e2cd30e-goog