* [PATCH] tests/compress_common: replace coremask with corelist in compress_common @ 2025-07-11 20:30 Manit Mahajan 2025-07-11 21:30 ` Patrick Robb 2025-07-16 19:24 ` [PATCH v2] compress: resolve test failure from incomplete output capture Manit Mahajan 0 siblings, 2 replies; 8+ messages in thread From: Manit Mahajan @ 2025-07-11 20:30 UTC (permalink / raw) To: dts; +Cc: dmarchan, probb, Manit Mahajan The compress_common.py script previously used the --coremask parameter, which has been removed from DPDK. As a result, zlib_pmd test cases fail when invoking this script. This patch updates the script to use the --corelist parameter instead, which is the more common way to specify core assignment in DPDK. Signed-off-by: Manit Mahajan <mmahajan@iol.unh.edu> --- tests/compress_common.py | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/tests/compress_common.py b/tests/compress_common.py index aa591270..22ce4b74 100644 --- a/tests/compress_common.py +++ b/tests/compress_common.py @@ -94,12 +94,14 @@ def get_input_file(test_case): def run_unit(test_case, eal={}): cores = test_case.dut.get_core_list("1S/3C/1T") - core_mask = utils.create_mask(cores) + core_list = [] + for core in cores: + core_list.append(str(core)) + core_str = ",".join(core_list) mem_channels = test_case.dut.get_memory_channels() default = default_eals.copy() - default["l"] = None - default["c"] = core_mask + default["l"] = core_str default["n"] = mem_channels eal_str = get_opt_str(test_case, default, eal) -- 2.41.0 ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] tests/compress_common: replace coremask with corelist in compress_common 2025-07-11 20:30 [PATCH] tests/compress_common: replace coremask with corelist in compress_common Manit Mahajan @ 2025-07-11 21:30 ` Patrick Robb 2025-07-15 7:42 ` David Marchand 2025-07-16 19:24 ` [PATCH v2] compress: resolve test failure from incomplete output capture Manit Mahajan 1 sibling, 1 reply; 8+ messages in thread From: Patrick Robb @ 2025-07-11 21:30 UTC (permalink / raw) To: Manit Mahajan; +Cc: dts, dmarchan [-- Attachment #1: Type: text/plain, Size: 1709 bytes --] This seems fine in principle. Did you ever figure out why the compressdev_autotest was failing with coremask used in the first place? I haven't looked at the coremask deprecation patch that you had found, but David said on slack that it didn't remove the option to use the coremask EAL arg, it just discourages it with a warning. On Fri, Jul 11, 2025 at 4:32 PM Manit Mahajan <mmahajan@iol.unh.edu> wrote: > The compress_common.py script previously used the --coremask parameter, > which has been removed from DPDK. As a result, zlib_pmd test cases > fail when invoking this script. This patch updates the script to > use the --corelist parameter instead, which is the more common > way to specify core assignment in DPDK. > > Signed-off-by: Manit Mahajan <mmahajan@iol.unh.edu> > --- > tests/compress_common.py | 8 +++++--- > 1 file changed, 5 insertions(+), 3 deletions(-) > > diff --git a/tests/compress_common.py b/tests/compress_common.py > index aa591270..22ce4b74 100644 > --- a/tests/compress_common.py > +++ b/tests/compress_common.py > @@ -94,12 +94,14 @@ def get_input_file(test_case): > > def run_unit(test_case, eal={}): > cores = test_case.dut.get_core_list("1S/3C/1T") > - core_mask = utils.create_mask(cores) > + core_list = [] > + for core in cores: > + core_list.append(str(core)) > + core_str = ",".join(core_list) > mem_channels = test_case.dut.get_memory_channels() > > default = default_eals.copy() > - default["l"] = None > - default["c"] = core_mask > + default["l"] = core_str > default["n"] = mem_channels > > eal_str = get_opt_str(test_case, default, eal) > -- > 2.41.0 > > [-- Attachment #2: Type: text/html, Size: 2273 bytes --] ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] tests/compress_common: replace coremask with corelist in compress_common 2025-07-11 21:30 ` Patrick Robb @ 2025-07-15 7:42 ` David Marchand 2025-07-16 15:53 ` Manit Mahajan 0 siblings, 1 reply; 8+ messages in thread From: David Marchand @ 2025-07-15 7:42 UTC (permalink / raw) To: Patrick Robb, Manit Mahajan; +Cc: dts, Thomas Monjalon Hello, On Fri, Jul 11, 2025 at 11:36 PM Patrick Robb <probb@iol.unh.edu> wrote: > > This seems fine in principle. > > Did you ever figure out why the compressdev_autotest was failing with coremask used in the first place? I haven't looked at the coremask deprecation patch that you had found, but David said on slack that it didn't remove the option to use the coremask EAL arg, it just discourages it with a warning. > > On Fri, Jul 11, 2025 at 4:32 PM Manit Mahajan <mmahajan@iol.unh.edu> wrote: >> >> The compress_common.py script previously used the --coremask parameter, >> which has been removed from DPDK. As a result, zlib_pmd test cases >> fail when invoking this script. This patch updates the script to There is no --coremask EAL option. And the option -c was not removed. Please refine the rca. >> use the --corelist parameter instead, which is the more common >> way to specify core assignment in DPDK. There is no --corelist EAL option. Only -c, -l and --lcores exist atm. -- David Marchand ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] tests/compress_common: replace coremask with corelist in compress_common 2025-07-15 7:42 ` David Marchand @ 2025-07-16 15:53 ` Manit Mahajan 0 siblings, 0 replies; 8+ messages in thread From: Manit Mahajan @ 2025-07-16 15:53 UTC (permalink / raw) To: David Marchand; +Cc: Patrick Robb, dts, Thomas Monjalon [-- Attachment #1: Type: text/plain, Size: 1701 bytes --] Hi, Thanks for the clarification. I had misunderstood the patch that added the coremask warning and incorrectly assumed that -c was deprecated. Upon closer review, I see that the patch only adds warning logs when using coremask but does not deprecate the -c option. The actual problem is that the warning message includes a '>', which causes send_expect('>') to return prematurely and miss important test output. I will address this in v2 by changing the expected prompt to 'RTE>'. On Tue, Jul 15, 2025 at 3:42 AM David Marchand <david.marchand@redhat.com> wrote: > Hello, > > On Fri, Jul 11, 2025 at 11:36 PM Patrick Robb <probb@iol.unh.edu> wrote: > > > > This seems fine in principle. > > > > Did you ever figure out why the compressdev_autotest was failing with > coremask used in the first place? I haven't looked at the coremask > deprecation patch that you had found, but David said on slack that it > didn't remove the option to use the coremask EAL arg, it just discourages > it with a warning. > > > > On Fri, Jul 11, 2025 at 4:32 PM Manit Mahajan <mmahajan@iol.unh.edu> > wrote: > >> > >> The compress_common.py script previously used the --coremask parameter, > >> which has been removed from DPDK. As a result, zlib_pmd test cases > >> fail when invoking this script. This patch updates the script to > > There is no --coremask EAL option. > And the option -c was not removed. > > Please refine the rca. > > >> use the --corelist parameter instead, which is the more common > >> way to specify core assignment in DPDK. > > There is no --corelist EAL option. > > Only -c, -l and --lcores exist atm. > > > -- > David Marchand > > [-- Attachment #2: Type: text/html, Size: 2286 bytes --] ^ permalink raw reply [flat|nested] 8+ messages in thread
* [PATCH v2] compress: resolve test failure from incomplete output capture 2025-07-11 20:30 [PATCH] tests/compress_common: replace coremask with corelist in compress_common Manit Mahajan 2025-07-11 21:30 ` Patrick Robb @ 2025-07-16 19:24 ` Manit Mahajan 2025-07-16 20:30 ` Patrick Robb 1 sibling, 1 reply; 8+ messages in thread From: Manit Mahajan @ 2025-07-16 19:24 UTC (permalink / raw) To: probb; +Cc: dts, david.marchand, Manit Mahajan The compress unit test uses send_expect() with '>' to determine when the command completed. A recent patch introduced a deprecation warning for coremask, which includes a '>' in the output. This causes send_expect() to exit early, resulting in missed test output, leading to false failures. This is fixed by changing the expected termination prompt from '>' to 'RTE>'. This change address issues from the v1 patch, which incorrectly modified the EAL option from coremask to corelist. Signed-off-by: Manit Mahajan <mmahajan@iol.unh.edu> --- tests/compress_common.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tests/compress_common.py b/tests/compress_common.py index aa591270..b4fd1921 100644 --- a/tests/compress_common.py +++ b/tests/compress_common.py @@ -106,8 +106,8 @@ def run_unit(test_case, eal={}): cmdline = "./{app_path} {eal}".format( app_path=test_case.dut.apps_name["test"], eal=eal_str ) - test_case.dut.send_expect(cmdline, ">", 30) - out = test_case.dut.send_expect("compressdev_autotest", ">", 30) + test_case.dut.send_expect(cmdline, "RTE>", 30) + out = test_case.dut.send_expect("compressdev_autotest", "RTE>", 30) test_case.dut.send_expect("quit", "# ", 30) print(out) -- 2.41.0 ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v2] compress: resolve test failure from incomplete output capture 2025-07-16 19:24 ` [PATCH v2] compress: resolve test failure from incomplete output capture Manit Mahajan @ 2025-07-16 20:30 ` Patrick Robb 2025-07-17 8:40 ` David Marchand 0 siblings, 1 reply; 8+ messages in thread From: Patrick Robb @ 2025-07-16 20:30 UTC (permalink / raw) To: Manit Mahajan; +Cc: dts, david.marchand [-- Attachment #1: Type: text/plain, Size: 1627 bytes --] Thanks. So, if we change the expect string to "RTE>", the warning will not trigger the expect and unit tests will complete before the execution proceeds? Sounds good. On Wed, Jul 16, 2025 at 3:24 PM Manit Mahajan <mmahajan@iol.unh.edu> wrote: > The compress unit test uses send_expect() with '>' to determine > when the command completed. A recent patch introduced a > deprecation warning for coremask, which includes a '>' in the > output. This causes send_expect() to exit early, > resulting in missed test output, leading to false failures. > > This is fixed by changing the expected termination prompt from > '>' to 'RTE>'. This change address issues from the v1 patch, > which incorrectly modified the EAL option from coremask to > corelist. > > Signed-off-by: Manit Mahajan <mmahajan@iol.unh.edu> > --- > tests/compress_common.py | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/tests/compress_common.py b/tests/compress_common.py > index aa591270..b4fd1921 100644 > --- a/tests/compress_common.py > +++ b/tests/compress_common.py > @@ -106,8 +106,8 @@ def run_unit(test_case, eal={}): > cmdline = "./{app_path} {eal}".format( > app_path=test_case.dut.apps_name["test"], eal=eal_str > ) > - test_case.dut.send_expect(cmdline, ">", 30) > - out = test_case.dut.send_expect("compressdev_autotest", ">", 30) > + test_case.dut.send_expect(cmdline, "RTE>", 30) > + out = test_case.dut.send_expect("compressdev_autotest", "RTE>", 30) > test_case.dut.send_expect("quit", "# ", 30) > print(out) > > -- > 2.41.0 > > [-- Attachment #2: Type: text/html, Size: 2256 bytes --] ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v2] compress: resolve test failure from incomplete output capture 2025-07-16 20:30 ` Patrick Robb @ 2025-07-17 8:40 ` David Marchand 2025-07-17 18:33 ` Patrick Robb 0 siblings, 1 reply; 8+ messages in thread From: David Marchand @ 2025-07-17 8:40 UTC (permalink / raw) To: Patrick Robb, Manit Mahajan; +Cc: dts, Thomas Monjalon Hello, On Wed, Jul 16, 2025 at 10:36 PM Patrick Robb <probb@iol.unh.edu> wrote: > > On Wed, Jul 16, 2025 at 3:24 PM Manit Mahajan <mmahajan@iol.unh.edu> wrote: >> >> The compress unit test uses send_expect() with '>' to determine >> when the command completed. A recent patch introduced a >> deprecation warning for coremask, which includes a '>' in the >> output. This causes send_expect() to exit early, >> resulting in missed test output, leading to false failures. >> >> This is fixed by changing the expected termination prompt from >> '>' to 'RTE>'. This change address issues from the v1 patch, >> which incorrectly modified the EAL option from coremask to >> corelist. >> >> Signed-off-by: Manit Mahajan <mmahajan@iol.unh.edu> > > Thanks. So, if we change the expect string to "RTE>", the warning will not trigger the expect and unit tests will complete before the execution proceeds? Sounds good. On the principle, that sounds good to me too. Just a note that other parts of "old" dts may be affected: $ git grep 'send_expect.*">"' tests/TestSuite_ieee1588.py: self.dut.send_expect("start", ">", 10) tests/TestSuite_multiprocess.py: self.session_secondary.send_expect("send hello_primary", ">") tests/TestSuite_multiprocess.py: self.session_secondary.send_expect("send hello_secondary", ">") tests/TestSuite_multiprocess.py: self.dut.send_expect("send %s" % line, ">") tests/TestSuite_multiprocess.py: self.session_secondary.send_expect("send hello_primary", ">") tests/TestSuite_multiprocess.py: self.session_secondary.send_expect("send hello_secondary", ">", 100) tests/TestSuite_multiprocess_iavf.py: self.session_secondary.send_expect("send hello_primary", ">") tests/TestSuite_multiprocess_iavf.py: self.session_secondary.send_expect("send hello_secondary", ">") tests/TestSuite_multiprocess_iavf.py: self.dut.send_expect("send %s" % line, ">") tests/TestSuite_multiprocess_iavf.py: self.session_secondary.send_expect("send hello_primary", ">") tests/TestSuite_multiprocess_iavf.py: self.session_secondary.send_expect("send hello_secondary", ">", 100) tests/TestSuite_shutdown_api.py: self.dut.send_expect("set link-up port %d" % i, ">") tests/TestSuite_speed_capabilities.py: self.dut.send_expect("set link-up port %d" % i, ">") tests/compress_common.py: test_case.dut.send_expect(cmdline, ">", 30) tests/compress_common.py: out = test_case.dut.send_expect("compressdev_autotest", ">", 30) How do we proceed now? Do you want me to push this fix in the dts legacy repo? -- David Marchand ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v2] compress: resolve test failure from incomplete output capture 2025-07-17 8:40 ` David Marchand @ 2025-07-17 18:33 ` Patrick Robb 0 siblings, 0 replies; 8+ messages in thread From: Patrick Robb @ 2025-07-17 18:33 UTC (permalink / raw) To: David Marchand; +Cc: Manit Mahajan, dts, Thomas Monjalon [-- Attachment #1: Type: text/plain, Size: 3432 bytes --] Hi David, Because Legacy DTS is deprecated, I think you should just push Manit's patch as is. In the case of the Zlib test, this is a legacy DTS test which we are still running at UNH, and we had to patch it locally to keep using the test, so of course we felt we should push the patch to patchwork for visibility at a minimum, and to be merged if maintainers want. Again I don't think it is the best use of Manit's time to have him doing further fixes in legacy DTS when he could be making other contributions to DPDK testing efforts elsewhere. But, if you don't agree, we will adjust. On Thu, Jul 17, 2025 at 4:40 AM David Marchand <david.marchand@redhat.com> wrote: > Hello, > > On Wed, Jul 16, 2025 at 10:36 PM Patrick Robb <probb@iol.unh.edu> wrote: > > > > On Wed, Jul 16, 2025 at 3:24 PM Manit Mahajan <mmahajan@iol.unh.edu> > wrote: > >> > >> The compress unit test uses send_expect() with '>' to determine > >> when the command completed. A recent patch introduced a > >> deprecation warning for coremask, which includes a '>' in the > >> output. This causes send_expect() to exit early, > >> resulting in missed test output, leading to false failures. > >> > >> This is fixed by changing the expected termination prompt from > >> '>' to 'RTE>'. This change address issues from the v1 patch, > >> which incorrectly modified the EAL option from coremask to > >> corelist. > >> > >> Signed-off-by: Manit Mahajan <mmahajan@iol.unh.edu> > > > > Thanks. So, if we change the expect string to "RTE>", the warning will > not trigger the expect and unit tests will complete before the execution > proceeds? Sounds good. > > On the principle, that sounds good to me too. > > Just a note that other parts of "old" dts may be affected: > $ git grep 'send_expect.*">"' > tests/TestSuite_ieee1588.py: self.dut.send_expect("start", ">", 10) > tests/TestSuite_multiprocess.py: > self.session_secondary.send_expect("send hello_primary", ">") > tests/TestSuite_multiprocess.py: > self.session_secondary.send_expect("send hello_secondary", ">") > tests/TestSuite_multiprocess.py: self.dut.send_expect("send > %s" % line, ">") > tests/TestSuite_multiprocess.py: > self.session_secondary.send_expect("send hello_primary", ">") > tests/TestSuite_multiprocess.py: > self.session_secondary.send_expect("send hello_secondary", ">", 100) > tests/TestSuite_multiprocess_iavf.py: > self.session_secondary.send_expect("send hello_primary", ">") > tests/TestSuite_multiprocess_iavf.py: > self.session_secondary.send_expect("send hello_secondary", ">") > tests/TestSuite_multiprocess_iavf.py: > self.dut.send_expect("send %s" % line, ">") > tests/TestSuite_multiprocess_iavf.py: > self.session_secondary.send_expect("send hello_primary", ">") > tests/TestSuite_multiprocess_iavf.py: > self.session_secondary.send_expect("send hello_secondary", ">", 100) > tests/TestSuite_shutdown_api.py: self.dut.send_expect("set > link-up port %d" % i, ">") > tests/TestSuite_speed_capabilities.py: > self.dut.send_expect("set link-up port %d" % i, ">") > tests/compress_common.py: test_case.dut.send_expect(cmdline, ">", 30) > tests/compress_common.py: out = > test_case.dut.send_expect("compressdev_autotest", ">", 30) > > > How do we proceed now? > Do you want me to push this fix in the dts legacy repo? > > > -- > David Marchand > > [-- Attachment #2: Type: text/html, Size: 4649 bytes --] ^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2025-07-17 18:39 UTC | newest] Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2025-07-11 20:30 [PATCH] tests/compress_common: replace coremask with corelist in compress_common Manit Mahajan 2025-07-11 21:30 ` Patrick Robb 2025-07-15 7:42 ` David Marchand 2025-07-16 15:53 ` Manit Mahajan 2025-07-16 19:24 ` [PATCH v2] compress: resolve test failure from incomplete output capture Manit Mahajan 2025-07-16 20:30 ` Patrick Robb 2025-07-17 8:40 ` David Marchand 2025-07-17 18:33 ` Patrick Robb
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).