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@xxxxxxxxx> > > --- > > 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; > > size = tx_info->xdpf->len; > > ena_buf = tx_info->bufs; > > > > -- > > 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
