On Mon, Oct 03, 2022 at 03:59:08PM +0200, Greg Kroah-Hartman wrote:
On Mon, Oct 03, 2022 at 02:48:19PM +0300, Peter Kosyh wrote:
Return value of a function 'xdp_convert_buff_to_frame' is dereferenced without checking for null, but it is usually checked for this function.
This fixed in upstream commit <e8223eeff02> while refactoring.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Signed-off-by: Peter Kosyh pkosyh@yandex.ru
drivers/net/ethernet/amazon/ena/ena_netdev.c | 2 ++ 1 file changed, 2 insertions(+)
diff --git a/drivers/net/ethernet/amazon/ena/ena_netdev.c b/drivers/net/ethernet/amazon/ena/ena_netdev.c index 52414ac2c901..9e6b2bd73dac 100644 --- a/drivers/net/ethernet/amazon/ena/ena_netdev.c +++ b/drivers/net/ethernet/amazon/ena/ena_netdev.c @@ -237,6 +237,8 @@ static int ena_xdp_tx_map_buff(struct ena_ring *xdp_ring, u32 size; tx_info->xdpf = xdp_convert_buff_to_frame(xdp);
- if (unlikely(!tx_info->xdpf))
goto error_report_dma_error;2.37.0
<formletter>
This is not the correct way to submit patches for inclusion in the stable kernel tree. Please read: https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html for how to do this properly.
</formletter>
To be specific, why is this non-upstream commit really needed? Can xdp_convert_buff_to_frame() ever fail under normal operation? Why was this one commit picked? And always properly reference commits in changelog text as the documentation asks you to.
Also, you didn't cc: everyone involved in the original commit, nor showed how you tested this in anyway (hint, read Documentation/process/researcher-guidelines.rst)
thanks,
greg k-h