| Message ID | 20230322011628.23359-1-krisman@suse.de (mailing list archive) |
|---|---|
| Headers | show |
| Series | io-wq: cleanup io_wq and io_wqe | expand |
On 3/21/23 7:16?PM, Gabriel Krisman Bertazi wrote: > Hi Jens, > > This tides up the io-wq internal interface by dropping the io_wqe/io_wq > separation, which no longer makes sense since commit > 0654b05e7e65 ("io_uring: One wqe per wq"). We currently have a single > io_wqe instance per io_wq, which is embedded in the structure. This > patchset merges the two, dropping bit of code to go from one to the > other in the io-wq implementation. > > I don't expect it to have any positive impact on performance, of course, > since hopefully the compiler optimizes it, but still, it is nice clean > up. To be sure, I measured with some mmtests microbenchmarks and I haven't > seen differences with or without the patchset. > > Patch 2 is slightly big to review but the use of wq and wqe is > intrinsically connected; it was a bit hard to break it in more pieces. > > Tested by running liburing's testsuite and mmtests performance > microbenchmarks (which uses fio). > > Based on your for-next branch. Nice! This is a great continuation of getting rid of the per-node worker setup.
On Tue, 21 Mar 2023 22:16:26 -0300, Gabriel Krisman Bertazi wrote: > This tides up the io-wq internal interface by dropping the io_wqe/io_wq > separation, which no longer makes sense since commit > 0654b05e7e65 ("io_uring: One wqe per wq"). We currently have a single > io_wqe instance per io_wq, which is embedded in the structure. This > patchset merges the two, dropping bit of code to go from one to the > other in the io-wq implementation. > > [...] Applied, thanks! [1/2] io-wq: Move wq accounting to io_wq commit: 98bab464dbc90c0f0495bcf9bc6e2d5a30da192d [2/2] io-wq: Drop struct io_wqe commit: 98bab464dbc90c0f0495bcf9bc6e2d5a30da192d Best regards,
Hi Jens, This tides up the io-wq internal interface by dropping the io_wqe/io_wq separation, which no longer makes sense since commit 0654b05e7e65 ("io_uring: One wqe per wq"). We currently have a single io_wqe instance per io_wq, which is embedded in the structure. This patchset merges the two, dropping bit of code to go from one to the other in the io-wq implementation. I don't expect it to have any positive impact on performance, of course, since hopefully the compiler optimizes it, but still, it is nice clean up. To be sure, I measured with some mmtests microbenchmarks and I haven't seen differences with or without the patchset. Patch 2 is slightly big to review but the use of wq and wqe is intrinsically connected; it was a bit hard to break it in more pieces. Tested by running liburing's testsuite and mmtests performance microbenchmarks (which uses fio). Based on your for-next branch. Thanks, Gabriel Krisman Bertazi (2): io-wq: Move wq accounting to io_wq io-wq: Drop struct io_wqe io_uring/io-wq.c | 408 ++++++++++++++++++++++------------------------- 1 file changed, 189 insertions(+), 219 deletions(-)