From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andy Lutomirski Subject: [PATCH v4 29/29] fork: Cache two thread stacks per cpu if CONFIG_VMAP_STACK is set Date: Sun, 26 Jun 2016 14:55:51 -0700 Message-ID: <32d90073b4c98c2fab5af97d9eecf7eea2f8c251.1466974736.git.luto@kernel.org> References: Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: Received: from mail.kernel.org ([198.145.29.136]:60768 "EHLO mail.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751705AbcFZV4j (ORCPT ); Sun, 26 Jun 2016 17:56:39 -0400 In-Reply-To: In-Reply-To: References: Sender: linux-arch-owner@vger.kernel.org List-ID: To: x86@kernel.org Cc: linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org, Borislav Petkov , Nadav Amit , Kees Cook , Brian Gerst , "kernel-hardening@lists.openwall.com" , Linus Torvalds , Josh Poimboeuf , Jann Horn , Heiko Carstens , Andy Lutomirski vmalloc is a bit slow, and pounding vmalloc/vfree will eventually force a global TLB flush. To reduce pressure on them, if CONFIG_VMAP_STACK, cache two thread stacks per cpu. This will let us quickly allocate a hopefully cache-hot, TLB-hot stack under heavy forking workloads (shell script style). On my silly pthread_create benchmark, it saves about 2 =C2=B5s per pthread_create+join with CONFIG_VMAP_STACK=3Dy. Signed-off-by: Andy Lutomirski --- kernel/fork.c | 54 ++++++++++++++++++++++++++++++++++++++++++++++++++-= --- 1 file changed, 50 insertions(+), 4 deletions(-) diff --git a/kernel/fork.c b/kernel/fork.c index 8dd1329e1bf8..4b8ea904e47b 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -159,10 +159,37 @@ void __weak arch_release_thread_stack(unsigned lo= ng *stack) * kmemcache based allocator. */ # if THREAD_SIZE >=3D PAGE_SIZE || defined(CONFIG_VMAP_STACK) + +#ifdef CONFIG_VMAP_STACK +/* + * vmalloc is a bit slow, and calling vfree enough times will force a = TLB + * flush. Try to minimize the number of calls by caching stacks. + */ +#define NR_CACHED_STACKS 2 +static DEFINE_PER_CPU(struct vm_struct *, cached_stacks[NR_CACHED_STAC= KS]); +#endif + static unsigned long *alloc_thread_stack_node(struct task_struct *tsk,= int node) { #ifdef CONFIG_VMAP_STACK - void *stack =3D __vmalloc_node_range( + void *stack; + int i; + + local_irq_disable(); + for (i =3D 0; i < NR_CACHED_STACKS; i++) { + struct vm_struct *s =3D this_cpu_read(cached_stacks[i]); + + if (!s) + continue; + this_cpu_write(cached_stacks[i], NULL); + + tsk->stack_vm_area =3D s; + local_irq_enable(); + return s->addr; + } + local_irq_enable(); + + stack =3D __vmalloc_node_range( THREAD_SIZE, THREAD_SIZE, VMALLOC_START, VMALLOC_END, THREADINFO_GFP | __GFP_HIGHMEM, PAGE_KERNEL, 0, node, __builtin_return_address(0)); @@ -185,10 +212,29 @@ static unsigned long *alloc_thread_stack_node(str= uct task_struct *tsk, int node) =20 static inline void free_thread_stack(struct task_struct *tsk) { - if (task_stack_vm_area(tsk)) +#ifdef CONFIG_VMAP_STACK + if (task_stack_vm_area(tsk)) { + unsigned long flags; + int i; + + local_irq_save(flags); + for (i =3D 0; i < NR_CACHED_STACKS; i++) { + if (this_cpu_read(cached_stacks[i])) + continue; + + this_cpu_write(cached_stacks[i], tsk->stack_vm_area); + goto done; + } + vfree(tsk->stack); - else - free_kmem_pages((unsigned long)tsk->stack, THREAD_SIZE_ORDER); + +done: + local_irq_restore(flags); + return; + } +#endif + + free_kmem_pages((unsigned long)tsk->stack, THREAD_SIZE_ORDER); } # else static struct kmem_cache *thread_stack_cache; --=20 2.7.4