From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-0.8 required=5.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, MAILING_LIST_MULTI,SPF_PASS,URIBL_BLOCKED autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 69E37FC6182 for ; Thu, 13 Sep 2018 21:42:18 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 0208D20C0A for ; Thu, 13 Sep 2018 21:42:17 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 0208D20C0A Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=intel.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-kernel-owner@vger.kernel.org Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1728226AbeINCxe (ORCPT ); Thu, 13 Sep 2018 22:53:34 -0400 Received: from mga01.intel.com ([192.55.52.88]:13954 "EHLO mga01.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1727734AbeINCxe (ORCPT ); Thu, 13 Sep 2018 22:53:34 -0400 X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from fmsmga006.fm.intel.com ([10.253.24.20]) by fmsmga101.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 13 Sep 2018 14:37:08 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.53,370,1531810800"; d="scan'208";a="263300609" Received: from rpedgeco-desk5.jf.intel.com ([10.54.75.168]) by fmsmga006.fm.intel.com with ESMTP; 13 Sep 2018 14:37:07 -0700 From: Rick Edgecombe To: tglx@linutronix.de, mingo@redhat.com, hpa@zytor.com, x86@kernel.org, linux-kernel@vger.kernel.org, linux-mm@kvack.org, kernel-hardening@lists.openwall.com, daniel@iogearbox.net, jannh@google.com, keescook@chromium.org, alexei.starovoitov@gmail.com Cc: kristen@linux.intel.com, dave.hansen@intel.com, arjan@linux.intel.com, Rick Edgecombe Subject: [PATCH v6 1/4] vmalloc: Add __vmalloc_node_try_addr function Date: Thu, 13 Sep 2018 14:31:35 -0700 Message-Id: <1536874298-23492-2-git-send-email-rick.p.edgecombe@intel.com> X-Mailer: git-send-email 2.7.4 In-Reply-To: <1536874298-23492-1-git-send-email-rick.p.edgecombe@intel.com> References: <1536874298-23492-1-git-send-email-rick.p.edgecombe@intel.com> Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Create __vmalloc_node_try_addr function that tries to allocate at a specific address and supports caller specified behavior for whether any lazy purging happens if there is a collision. This new function draws from the __vmalloc_node_range implementation. Attempts to merge the two into a single allocator resulted in logic that was difficult to follow, so they are left separate. Signed-off-by: Rick Edgecombe --- include/linux/vmalloc.h | 3 + mm/vmalloc.c | 177 +++++++++++++++++++++++++++++++++++++++++++++++- 2 files changed, 179 insertions(+), 1 deletion(-) diff --git a/include/linux/vmalloc.h b/include/linux/vmalloc.h index 398e9c9..c7712c8 100644 --- a/include/linux/vmalloc.h +++ b/include/linux/vmalloc.h @@ -82,6 +82,9 @@ extern void *__vmalloc_node_range(unsigned long size, unsigned long align, unsigned long start, unsigned long end, gfp_t gfp_mask, pgprot_t prot, unsigned long vm_flags, int node, const void *caller); +extern void *__vmalloc_node_try_addr(unsigned long addr, unsigned long size, + gfp_t gfp_mask, pgprot_t prot, unsigned long vm_flags, + int node, int try_purge, const void *caller); #ifndef CONFIG_MMU extern void *__vmalloc_node_flags(unsigned long size, int node, gfp_t flags); static inline void *__vmalloc_node_flags_caller(unsigned long size, int node, diff --git a/mm/vmalloc.c b/mm/vmalloc.c index a728fc4..1954458 100644 --- a/mm/vmalloc.c +++ b/mm/vmalloc.c @@ -1709,6 +1709,181 @@ static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask, return NULL; } +static bool pvm_find_next_prev(unsigned long end, + struct vmap_area **pnext, + struct vmap_area **pprev); + +/* Try to allocate a region of KVA of the specified address and size. */ +static struct vmap_area *try_alloc_vmap_area(unsigned long addr, + unsigned long size, int node, gfp_t gfp_mask, + int try_purge) +{ + struct vmap_area *va; + struct vmap_area *cur_va = NULL; + struct vmap_area *first_before = NULL; + int need_purge = 0; + int blocked = 0; + int purged = 0; + unsigned long addr_end; + + WARN_ON(!size); + WARN_ON(offset_in_page(size)); + + addr_end = addr + size; + if (addr > addr_end) + return ERR_PTR(-EOVERFLOW); + + might_sleep(); + + va = kmalloc_node(sizeof(struct vmap_area), + gfp_mask & GFP_RECLAIM_MASK, node); + if (unlikely(!va)) + return ERR_PTR(-ENOMEM); + + /* + * Only scan the relevant parts containing pointers to other objects + * to avoid false negatives. + */ + kmemleak_scan_area(&va->rb_node, SIZE_MAX, gfp_mask & GFP_RECLAIM_MASK); + +retry: + spin_lock(&vmap_area_lock); + + pvm_find_next_prev(addr, &cur_va, &first_before); + + if (!cur_va) + goto found; + + /* + * If there is no VA that starts before the target address, start the + * check from the closest VA in order to cover the case where the + * allocation overlaps at the end. + */ + if (first_before && addr < first_before->va_end) + cur_va = first_before; + + /* Linearly search through to make sure there is a hole */ + while (cur_va->va_start < addr_end) { + if (cur_va->va_end > addr) { + if (cur_va->flags & VM_LAZY_FREE) { + need_purge = 1; + } else { + blocked = 1; + break; + } + } + + if (list_is_last(&cur_va->list, &vmap_area_list)) + break; + + cur_va = list_next_entry(cur_va, list); + } + + /* + * If a non-lazy free va blocks the allocation, or + * we are not supposed to purge, but we need to, the + * allocation fails. + */ + if (blocked || (need_purge && !try_purge)) + goto fail; + + if (try_purge && need_purge) { + /* if purged once before, give up */ + if (purged) + goto fail; + + /* + * If the va blocking the allocation is set to + * be purged then purge all vmap_areas that are + * set to purged since this will flush the TLBs + * anyway. + */ + spin_unlock(&vmap_area_lock); + purge_vmap_area_lazy(); + need_purge = 0; + purged = 1; + goto retry; + } + +found: + va->va_start = addr; + va->va_end = addr_end; + va->flags = 0; + __insert_vmap_area(va); + spin_unlock(&vmap_area_lock); + + return va; +fail: + spin_unlock(&vmap_area_lock); + kfree(va); + if (need_purge && !blocked) + return ERR_PTR(-EUCLEAN); + return ERR_PTR(-EBUSY); +} + +/** + * __vmalloc_try_addr - try to alloc at a specific address + * @addr: address to try + * @size: size to try + * @gfp_mask: flags for the page level allocator + * @prot: protection mask for the allocated pages + * @vm_flags: additional vm area flags (e.g. %VM_NO_GUARD) + * @node: node to use for allocation or NUMA_NO_NODE + * @try_purge: try to purge if needed to fulfill and allocation + * @caller: caller's return address + * + * Try to allocate at the specific address. If it succeeds the address is + * returned. If it fails an EBUSY ERR_PTR is returned. If try_purge is + * zero, it will return an EUCLEAN ERR_PTR if it could have allocated if it + * was allowed to purge. It may trigger TLB flushes if a purge is needed, + * and try_purge is set. + */ +void *__vmalloc_node_try_addr(unsigned long addr, unsigned long size, + gfp_t gfp_mask, pgprot_t prot, unsigned long vm_flags, + int node, int try_purge, const void *caller) +{ + struct vmap_area *va; + struct vm_struct *area; + void *alloc_addr; + unsigned long real_size = size; + + size = PAGE_ALIGN(size); + if (!size || (size >> PAGE_SHIFT) > totalram_pages) + return NULL; + + WARN_ON(in_interrupt()); + + if (!(vm_flags & VM_NO_GUARD)) + size += PAGE_SIZE; + + va = try_alloc_vmap_area(addr, size, node, gfp_mask, try_purge); + if (IS_ERR(va)) + goto fail; + + area = kzalloc_node(sizeof(*area), gfp_mask & GFP_RECLAIM_MASK, node); + if (unlikely(!area)) { + warn_alloc(gfp_mask, NULL, "kmalloc: allocation failure"); + return ERR_PTR(-ENOMEM); + } + + setup_vmalloc_vm(area, va, vm_flags, caller); + + alloc_addr = __vmalloc_area_node(area, gfp_mask, prot, node); + if (!alloc_addr) { + warn_alloc(gfp_mask, NULL, + "vmalloc: allocation failure: %lu bytes", real_size); + return ERR_PTR(-ENOMEM); + } + + clear_vm_uninitialized_flag(area); + + kmemleak_vmalloc(area, real_size, gfp_mask); + + return alloc_addr; +fail: + return va; +} + /** * __vmalloc_node_range - allocate virtually contiguous memory * @size: allocation size @@ -2355,7 +2530,6 @@ void free_vm_area(struct vm_struct *area) } EXPORT_SYMBOL_GPL(free_vm_area); -#ifdef CONFIG_SMP static struct vmap_area *node_to_va(struct rb_node *n) { return rb_entry_safe(n, struct vmap_area, rb_node); @@ -2403,6 +2577,7 @@ static bool pvm_find_next_prev(unsigned long end, return true; } +#ifdef CONFIG_SMP /** * pvm_determine_end - find the highest aligned address between two vmap_areas * @pnext: in/out arg for the next vmap_area -- 2.7.4