linux-kernel - [PATCH] mm: fix racing of vb->va when kasan enabled

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-ID: <1653447164-15017-1-git-send-email-zhaoyang.huang@unisoc.com>
Date:   Wed, 25 May 2022 10:52:44 +0800
From:   "zhaoyang.huang" <zhaoyang.huang@...soc.com>
To:     Andrew Morton <akpm@...ux-foundation.org>,
        Zhaoyang Huang <huangzhaoyang@...il.com>, <linux-mm@...ck.org>,
        <linux-kernel@...r.kernel.org>, <ke.wang@...soc.com>
Subject: [PATCH] mm: fix racing of vb->va when kasan enabled

From: Zhaoyang Huang <zhaoyang.huang@...soc.com>

Accessing to vb->va could be deemed as use after free when KASAN is
enabled like bellowing. Fix it by expanding the mutex's range.

[   20.232335] ==================================================================
[   20.232365] BUG: KASAN: use-after-free in _vm_unmap_aliases+0x164/0x364
[   20.232376] Read of size 8 at addr ffffff80d84af780 by task modprobe/300
[   20.232380]
[   20.232395] CPU: 5 PID: 300 Comm: modprobe Tainted: G S       C O      5.4.161-android12-9-03238-gd43329d103de-ab20547 #1
[   20.232401] Hardware name: Spreadtrum UMS512-1H10 SoC (DT)
[   20.232407] Call trace:
[   20.232419]  dump_backtrace+0x0/0x2b4
[   20.232428]  show_stack+0x24/0x30
[   20.232443]  dump_stack+0x15c/0x1f4
[   20.232455]  print_address_description+0x88/0x568
[   20.232465]  __kasan_report+0x1b8/0x1dc
[   20.232474]  kasan_report+0x10/0x18
[   20.232486]  __asan_report_load8_noabort+0x1c/0x24
[   20.232495]  _vm_unmap_aliases+0x164/0x364
[   20.232505]  vm_unmap_aliases+0x20/0x28
[   20.232516]  change_memory_common+0x2c4/0x3ec
[   20.232524]  set_memory_ro+0x30/0x3c
[   20.232539]  module_enable_ro+0x144/0x3f0
[   20.232547]  load_module+0x54c0/0x8248
[   20.232555]  __se_sys_finit_module+0x174/0x1b0
[   20.232564]  __arm64_sys_finit_module+0x78/0x88
[   20.232573]  el0_svc_common+0x19c/0x354
[   20.232581]  el0_svc_handler+0x48/0x54
[   20.232591]  el0_svc+0x8/0xc
[   20.232595]
[   20.232602] Allocated by task 297:
[   20.232615]  __kasan_kmalloc+0x130/0x1f8
[   20.232625]  kasan_slab_alloc+0x14/0x1c
[   20.232638]  kmem_cache_alloc+0x1dc/0x394
[   20.232648]  alloc_vmap_area+0xb4/0x1630
[   20.232657]  vm_map_ram+0x3ac/0x768
[   20.232671]  z_erofs_decompress_generic+0x2f0/0x844
[   20.232681]  z_erofs_decompress+0xa8/0x594
[   20.232692]  z_erofs_decompress_pcluster+0xeb4/0x1458
[   20.232702]  z_erofs_vle_unzip_wq+0xe4/0x140
[   20.232715]  process_one_work+0x5c0/0x10ac
[   20.232724]  worker_thread+0x888/0x1128
[   20.232733]  kthread+0x290/0x304
[   20.232744]  ret_from_fork+0x10/0x18
[   20.232747]
[   20.232752] Freed by task 51:
[   20.232762]  __kasan_slab_free+0x1a0/0x270
[   20.232772]  kasan_slab_free+0x10/0x1c
[   20.232781]  slab_free_freelist_hook+0xd0/0x1ac
[   20.232792]  kmem_cache_free+0x110/0x368
[   20.232803]  __purge_vmap_area_lazy+0x524/0x13e4
[   20.232813]  _vm_unmap_aliases+0x290/0x364
[   20.232822]  __vunmap+0x45c/0x5c4
[   20.232831]  vfree+0x74/0x16c
[   20.232841]  module_memfree+0x44/0x7c
[   20.232850]  do_free_init+0x5c/0xac
[   20.232860]  process_one_work+0x5c0/0x10ac
[   20.232869]  worker_thread+0xb3c/0x1128
[   20.232877]  kthread+0x290/0x304
[   20.232887]  ret_from_fork+0x10/0x18

Signed-off-by: Zhaoyang Huang <zhaoyang.huang@...soc.com>
---
 mm/vmalloc.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/mm/vmalloc.c b/mm/vmalloc.c
index d2a00ad..028d65a 100644
--- a/mm/vmalloc.c
+++ b/mm/vmalloc.c
@@ -2081,7 +2081,7 @@ static void _vm_unmap_aliases(unsigned long start, unsigned long end, int flush)
 		return;
 
 	might_sleep();
-
+	mutex_lock(&vmap_purge_lock);
 	for_each_possible_cpu(cpu) {
 		struct vmap_block_queue *vbq = &per_cpu(vmap_block_queue, cpu);
 		struct vmap_block *vb;
@@ -2106,7 +2106,6 @@ static void _vm_unmap_aliases(unsigned long start, unsigned long end, int flush)
 		rcu_read_unlock();
 	}
 
-	mutex_lock(&vmap_purge_lock);
 	purge_fragmented_blocks_allcpus();
 	if (!__purge_vmap_area_lazy(start, end) && flush)
 		flush_tlb_kernel_range(start, end);
-- 
1.9.1