Re: [PATCH bpf-next 2/6] selftests/xsk: Add tail adjustment functionality to XDP

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 02/25, Vyavahare, Tushar wrote:
> 
> 
> > -----Original Message-----
> > From: Stanislav Fomichev <stfomichev@xxxxxxxxx>
> > Sent: Thursday, February 20, 2025 11:17 PM
> > To: Vyavahare, Tushar <tushar.vyavahare@xxxxxxxxx>
> > Cc: bpf@xxxxxxxxxxxxxxx; netdev@xxxxxxxxxxxxxxx; bjorn@xxxxxxxxxx; Karlsson,
> > Magnus <magnus.karlsson@xxxxxxxxx>; Fijalkowski, Maciej
> > <maciej.fijalkowski@xxxxxxxxx>; jonathan.lemon@xxxxxxxxx;
> > davem@xxxxxxxxxxxxx; kuba@xxxxxxxxxx; pabeni@xxxxxxxxxx; ast@xxxxxxxxxx;
> > daniel@xxxxxxxxxxxxx
> > Subject: Re: [PATCH bpf-next 2/6] selftests/xsk: Add tail adjustment functionality
> > to XDP
> > 
> > On 02/20, Tushar Vyavahare wrote:
> > > Introduce a new function, xsk_xdp_adjust_tail, within the XDP program
> > > to adjust the tail of packets. This function utilizes
> > > bpf_xdp_adjust_tail to modify the packet size dynamically based on the 'count'
> > variable.
> > >
> > > If the adjustment fails, the packet is dropped using XDP_DROP to
> > > ensure processing of only correctly modified packets.
> > >
> > > Signed-off-by: Tushar Vyavahare <tushar.vyavahare@xxxxxxxxx>
> > 
> > Any reason not to combine patches 2..5 into a single one? I looked through each
> > one briefly and it's a bit hard to follow when trying to put everything together..
> 
> Maybe that was too many patches. How about this?
>  
> #1: selftests/xsk: Add packet stream replacement function
> #2: selftests/xsk: Add tail adjustment test functionality to AF_XDP.
> #3: selftests/xsk: Add support check for bpf_xdp_adjust_tail() helper in
>     xskxceiver
> #4: selftests/xsk: Implement and test packet resizing with
>     bpf_xdp_adjust_tail
 
Even that might be too much. #1 is clearly refactoring - keep it separate. The
rest seems like it belongs to the same testcase(s)? I'd put patches 2,3,4,5,6
in the same patch.

For example, you add bpf xsk_xdp_adjust_tail function, but there are
not callers in the same patch, so it's not clear what's the context.
Same for testapp_xdp_adjust_tail - there is no caller, so the reader
needs to go to the next patch to put it all together. Same for
testapp_adjust_tail...




[Index of Archives]     [Linux Samsung SoC]     [Linux Rockchip SoC]     [Linux Actions SoC]     [Linux for Synopsys ARC Processors]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]


  Powered by Linux