Re: [PATCH net-next v16 09/12] virtio_net: Implement IPv4 ethtool flow rules

[Dan Jurgens <danielj@nvidia.com>]

On 1/22/26 11:21 AM, Michael S. Tsirkin wrote:
> On Wed, Jan 21, 2026 at 04:06:49PM -0600, Daniel Jurgens wrote:
>> Add support for IP_USER type rules from ethtool.
>>
>> Example:
>> $ ethtool -U ens9 flow-type ip4 src-ip 192.168.51.101 action -1
>> Added rule with ID 1
>>
>> The example rule will drop packets with the source IP specified.
>>
>> Signed-off-by: Daniel Jurgens <danielj@nvidia.com>
>> Reviewed-by: Parav Pandit <parav@nvidia.com>
>> Reviewed-by: Shahar Shitrit <shshitrit@nvidia.com>
>> Reviewed-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
>> ---
>> v4:
>>     - Fixed bug in protocol check of parse_ip4
>>     - (u8 *) to (void *) casting.
>>     - Alignment issues.
>>
>> v12
>>     - refactor calculate_flow_sizes to remove goto. MST
>>     - refactor build_and_insert to remove goto validate. MST
>>     - Move parse_ip4 l3_mask check to TCP/UDP patch. MST
>>     - Check saddr/daddr mask before copying in parse_ip4. MST
>>     - Remove tos check in setup_ip_key_mask.
>>     - check l4_4_bytes mask is 0 in setup_ip_key_mask. MST
>>     - changed return of setup_ip_key_mask to -EINVAL.
>>     - BUG_ON if key overflows u8 size in calculate_flow_sizes. MST
>>
>> v13:
>>     - Set tos field if applicable in parse_ip4. MST
>>     - Check tos in validate_ip4_mask. MST
>>     - check l3_mask before setting addr and mask in parse_ip4. MST
>>     - use has_ipv4 vs numhdrs for branching in build_and_insert. MST
>> ---
>> ---
>>  drivers/net/virtio_net.c | 129 +++++++++++++++++++++++++++++++++++++--
>>  1 file changed, 123 insertions(+), 6 deletions(-)
>>
>> diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
>> index ba231f10b803..7ce120baeb41 100644
>> --- a/drivers/net/virtio_net.c
>> +++ b/drivers/net/virtio_net.c
>> @@ -5818,6 +5818,39 @@ static bool validate_eth_mask(const struct virtnet_ff *ff,
>>  	return true;
>>  }
>>  
>> +static bool validate_ip4_mask(const struct virtnet_ff *ff,
>> +			      const struct virtio_net_ff_selector *sel,
>> +			      const struct virtio_net_ff_selector *sel_cap)
>> +{
>> +	bool partial_mask = !!(sel_cap->flags & VIRTIO_NET_FF_MASK_F_PARTIAL_MASK);
>> +	struct iphdr *cap, *mask;
>> +
>> +	cap = (struct iphdr *)&sel_cap->mask;
>> +	mask = (struct iphdr *)&sel->mask;
> 
> 
> this cast is only portable if sel and sel_cap are properly aligned.
> 
> 
> 
> Because if not then at least saddr/daddr accesses below are
> not portable and need
> Documentation/core-api/unaligned-memory-access.rst
> 

I'll fix this here, and in the next 2 patches.