Re: [PATCH] sfc: potential dereference of null pointer
flat view
From: Jakub Kicinski <kuba@kernel.org>
Date: 2021-12-17 04:19:49
Also in:
bpf, lkml
On Thu, 16 Dec 2021 16:49:02 +0800 Jiasheng Jiang wrote:
[PATCH] sfc: potential dereference of null pointer
Please give the commit a more unique name.
The return value of kcalloc() needs to be checked. To avoid dereference of null pointer in case of the failure of alloc, such as efx_fini_rx_recycle_ring(). Therefore, it should be better to change the definition of page_ptr_mask to signed int and then assign the page_ptr_mask to -1 when page_ring is NULL, in order to avoid the use in the loop in efx_fini_rx_recycle_ring().
What about uses of this ring, you mention avoiding null-deref in the freeing function but presumably something is actually using this ring at runtime. Seems like we should return -ENOMEM.
Fixes: 3d95b884392f ("sfc: move more rx code")Moving buggy code does not introduce the problem, please find a Fixes tag referring to where the bad code was _introduced_.
quoted hunk ↗ jump to hunk
Signed-off-by: Jiasheng Jiang <redacted> --- drivers/net/ethernet/sfc/net_driver.h | 2 +- drivers/net/ethernet/sfc/rx_common.c | 5 ++++- 2 files changed, 5 insertions(+), 2 deletions(-)diff --git a/drivers/net/ethernet/sfc/net_driver.h b/drivers/net/ethernet/sfc/net_driver.h index 9b4b25704271..beba3e0a6027 100644 --- a/drivers/net/ethernet/sfc/net_driver.h +++ b/drivers/net/ethernet/sfc/net_driver.h@@ -407,7 +407,7 @@ struct efx_rx_queue { unsigned int page_recycle_count; unsigned int page_recycle_failed; unsigned int page_recycle_full; - unsigned int page_ptr_mask; + int page_ptr_mask; unsigned int max_fill; unsigned int fast_fill_trigger; unsigned int min_fill;diff --git a/drivers/net/ethernet/sfc/rx_common.c b/drivers/net/ethernet/sfc/rx_common.c index 68fc7d317693..d9d0a5805f1c 100644 --- a/drivers/net/ethernet/sfc/rx_common.c +++ b/drivers/net/ethernet/sfc/rx_common.c@@ -150,7 +150,10 @@ static void efx_init_rx_recycle_ring(struct efx_rx_queue *rx_queue) efx->rx_bufs_per_page); rx_queue->page_ring = kcalloc(page_ring_size, sizeof(*rx_queue->page_ring), GFP_KERNEL); - rx_queue->page_ptr_mask = page_ring_size - 1; + if (!rx_queue->page_ring) + rx_queue->page_ptr_mask = -1; + else + rx_queue->page_ptr_mask = page_ring_size - 1; } static void efx_fini_rx_recycle_ring(struct efx_rx_queue *rx_queue)