]> git.dujemihanovic.xyz Git - linux.git/commitdiff
swiotlb: initialise restricted pool list_head when SWIOTLB_DYNAMIC=y
authorWill Deacon <will@kernel.org>
Thu, 2 May 2024 09:37:23 +0000 (10:37 +0100)
committerChristoph Hellwig <hch@lst.de>
Thu, 2 May 2024 12:57:04 +0000 (14:57 +0200)
Using restricted DMA pools (CONFIG_DMA_RESTRICTED_POOL=y) in conjunction
with dynamic SWIOTLB (CONFIG_SWIOTLB_DYNAMIC=y) leads to the following
crash when initialising the restricted pools at boot-time:

  | Unable to handle kernel NULL pointer dereference at virtual address 0000000000000008
  | Internal error: Oops: 0000000096000005 [#1] PREEMPT SMP
  | pc : rmem_swiotlb_device_init+0xfc/0x1ec
  | lr : rmem_swiotlb_device_init+0xf0/0x1ec
  | Call trace:
  |  rmem_swiotlb_device_init+0xfc/0x1ec
  |  of_reserved_mem_device_init_by_idx+0x18c/0x238
  |  of_dma_configure_id+0x31c/0x33c
  |  platform_dma_configure+0x34/0x80

faddr2line reveals that the crash is in the list validation code:

  include/linux/list.h:83
  include/linux/rculist.h:79
  include/linux/rculist.h:106
  kernel/dma/swiotlb.c:306
  kernel/dma/swiotlb.c:1695

because add_mem_pool() is trying to list_add_rcu() to a NULL
'mem->pools'.

Fix the crash by initialising the 'mem->pools' list_head in
rmem_swiotlb_device_init() before calling add_mem_pool().

Reported-by: Nikita Ioffe <ioffe@google.com>
Tested-by: Nikita Ioffe <ioffe@google.com>
Fixes: 1aaa736815eb ("swiotlb: allocate a new memory pool when existing pools are full")
Signed-off-by: Will Deacon <will@kernel.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
kernel/dma/swiotlb.c

index a5e0dfc44d24e22641e72bb0362511a33b23a1fd..0de66f0ff43aba1f64505d70b86d9faac8bc43ff 100644 (file)
@@ -1798,6 +1798,7 @@ static int rmem_swiotlb_device_init(struct reserved_mem *rmem,
                mem->for_alloc = true;
 #ifdef CONFIG_SWIOTLB_DYNAMIC
                spin_lock_init(&mem->lock);
+               INIT_LIST_HEAD_RCU(&mem->pools);
 #endif
                add_mem_pool(mem, pool);