> -----Original Message----- > From: Daniel Borkmann <daniel@xxxxxxxxxxxxx> > Sent: Wednesday, August 10, 2022 4:18 PM > To: Fijalkowski, Maciej <maciej.fijalkowski@xxxxxxxxx>; Koikkara Reeny, > Shibin <shibin.koikkara.reeny@xxxxxxxxx> > Cc: bpf@xxxxxxxxxxxxxxx; ast@xxxxxxxxxx; netdev@xxxxxxxxxxxxxxx; Karlsson, > Magnus <magnus.karlsson@xxxxxxxxx>; bjorn@xxxxxxxxxx; > kuba@xxxxxxxxxx; andrii@xxxxxxxxxx; Loftus, Ciara <ciara.loftus@xxxxxxxxx> > Subject: Re: [PATCH bpf-next v4] selftests: xsk: Update poll test cases > > On 8/10/22 3:08 PM, Maciej Fijalkowski wrote: > > On Wed, Aug 03, 2022 at 02:43:54PM +0000, Shibin Koikkara Reeny wrote: > >> Poll test case was not testing all the functionality of the poll > >> feature in the testsuite. This patch update the poll test case which > >> contain 2 testcases to > > > > updates, contains, test cases > > > >> test the RX and the TX poll functionality and additional > >> 2 more testcases to check the timeout features of the > > > > feature > > > >> poll event. > >> > >> Poll testsuite have 4 test cases: > > > > test suite, has > > > >> > >> 1. TEST_TYPE_RX_POLL: > >> Check if RX path POLLIN function work as expect. TX path > > > > works > > > >> can use any method to sent the traffic. > > > > send > > > >> > >> 2. TEST_TYPE_TX_POLL: > >> Check if TX path POLLOUT function work as expect. RX path can use any > >> method to receive the traffic. > >> > >> 3. TEST_TYPE_POLL_RXQ_EMPTY: > >> Call poll function with parameter POLLIN on empty rx queue will cause > >> timeout.If return timeout then test case is pass. > > > > space after dot > > > >> > >> 4. TEST_TYPE_POLL_TXQ_FULL: > >> When txq is filled and packets are not cleaned by the kernel then if > >> we invoke the poll function with POLLOUT then it should trigger > >> timeout. > >> > >> v1: > >> https://lore.kernel.org/bpf/20220718095712.588513-1-shibin.koikkara.r > >> eeny@xxxxxxxxx/ > >> v2: > >> https://lore.kernel.org/bpf/20220726101723.250746-1-shibin.koikkara.r > >> eeny@xxxxxxxxx/ > >> v3: > >> https://lore.kernel.org/bpf/20220729132337.211443-1-shibin.koikkara.r > >> eeny@xxxxxxxxx/ > >> > >> Changes in v2: > >> * Updated the commit message > >> * fixed the while loop flow in receive_pkts function. > >> Changes in v3: > >> * Introduced single thread validation function. > >> * Removed pkt_stream_invalid(). > >> * Updated TEST_TYPE_POLL_TXQ_FULL testcase to create invalid frame. > >> * Removed timer from send_pkts(). > >> * Removed boolean variable skip_rx and skip_tx. > >> Change in v4: > >> * Added is_umem_valid() > > > > for single patches, I believe that it's concerned a better practice to > > include the versioning below the '---' line? > > > >> > >> Signed-off-by: Shibin Koikkara Reeny > >> <shibin.koikkara.reeny@xxxxxxxxx> > >> --- > >> tools/testing/selftests/bpf/xskxceiver.c | 166 +++++++++++++++++----- > - > >> tools/testing/selftests/bpf/xskxceiver.h | 8 +- > >> 2 files changed, 134 insertions(+), 40 deletions(-) > > > > I don't think these grammar suggestions require a new revision, so: > > Reviewed-by: Maciej Fijalkowski <maciej.fijalkowski@xxxxxxxxx> > > I cleaned these up while applying. Shibin, please take care of this before > sending out next time, thanks guys! Thank you Daniel. Appreciate it. 😊
