| Message ID | 166393728005.2213882.4162674859542409548.stgit@firesoul (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | b860a1b964be7917ed3dcaa674ec0a2ba0a62aa0 |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [net-next] xdp: Adjust xdp_frame layout to avoid using bitfields | expand |
Hello: This patch was applied to netdev/net-next.git (master) by Jakub Kicinski <kuba@kernel.org>: On Fri, 23 Sep 2022 14:48:00 +0200 you wrote: > Practical experience (and advice from Alexei) tell us that bitfields in > structs lead to un-optimized assemply code. I've verified this change > does lead to better x86_64 assemply, both via objdump and playing with > code snippets in godbolt.org. > > Using scripts/bloat-o-meter shows the code size is reduced with 24 > bytes for xdp_convert_buff_to_frame() that gets inlined e.g. in > i40e_xmit_xdp_tx_ring() which were used for microbenchmarking. > > [...] Here is the summary with links: - [net-next] xdp: Adjust xdp_frame layout to avoid using bitfields https://git.kernel.org/netdev/net-next/c/b860a1b964be You are awesome, thank you!
diff --git a/include/net/xdp.h b/include/net/xdp.h index 04c852c7a77f..55dbc68bfffc 100644 --- a/include/net/xdp.h +++ b/include/net/xdp.h @@ -164,13 +164,13 @@ struct xdp_frame { void *data; u16 len; u16 headroom; - u32 metasize:8; - u32 frame_sz:24; + u32 metasize; /* uses lower 8-bits */ /* Lifetime of xdp_rxq_info is limited to NAPI/enqueue time, * while mem info is valid on remote CPU. */ struct xdp_mem_info mem; struct net_device *dev_rx; /* used by cpumap */ + u32 frame_sz; u32 flags; /* supported values defined in xdp_buff_flags */ };
Practical experience (and advice from Alexei) tell us that bitfields in structs lead to un-optimized assemply code. I've verified this change does lead to better x86_64 assemply, both via objdump and playing with code snippets in godbolt.org. Using scripts/bloat-o-meter shows the code size is reduced with 24 bytes for xdp_convert_buff_to_frame() that gets inlined e.g. in i40e_xmit_xdp_tx_ring() which were used for microbenchmarking. Microbenchmarking results do show improvements, but very small and varying between 0.5 to 2 nanosec improvement per packet. The member @metasize is changed from u8 to u32. Future users of this area could split this into two u16 fields. I've also benchmarked with two u16 fields showing equal performance gains and code size reduction. The moved member @frame_sz doesn't change sizeof struct due to existing padding. Like xdp_buff member @frame_sz is placed next to @flags, which allows compiler to optimize assignment of these. Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com> --- include/net/xdp.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)