| Message ID | 20240520190131.94904-1-ericsunshine@charter.net (mailing list archive) |
|---|---|
| Headers | show |
| Series | improve chainlint.pl CPU count computation | expand |
On Mon, 2024-05-20 at 15:01 -0400, Eric Sunshine wrote: > From: Eric Sunshine <sunshine@sunshineco.com> > > This series replaces a patch[1] sent by John Paul Adrian Glaubitz to fix > chainlint.pl CPU count computation on Linux SPARC. > > Unlike its predecessor, this series also fixes an underlying problem in > which ncores() could return 0 which would result in chainlint.pl not > processing any of its input test scripts. Patch [3/3] also fixes CPU > count detection on Alpha[2]. > > Patch [2/3] of this series is more or less Adrian's original patch[1] so > it retains his authorship, though I simplified the regular-expression > and tweaked the commit message. > > [1]: https://lore.kernel.org/git/20240520111109.99882-1-glaubitz@physik.fu-berlin.de/ > [2]: https://lore.kernel.org/git/503a99f3511559722a3eeef15d31027dfe617fa1.camel@physik.fu-berlin.de/ > > Eric Sunshine (2): > chainlint.pl: make CPU count computation more robust > chainlint.pl: latch CPU count directly reported by /proc/cpuinfo > > John Paul Adrian Glaubitz (1): > chainlint.pl: fix incorrect CPU count on Linux SPARC > > t/chainlint.pl | 20 +++++++++++++++++--- > 1 file changed, 17 insertions(+), 3 deletions(-) > Works as expected on my Linux SPARC machine running Debian unstable. Tested-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> Adrian
On Mon, May 20, 2024 at 3:17 PM John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> wrote: > On Mon, 2024-05-20 at 15:01 -0400, Eric Sunshine wrote: > > From: Eric Sunshine <sunshine@sunshineco.com> > > This series replaces a patch[1] sent by John Paul Adrian Glaubitz to fix > > chainlint.pl CPU count computation on Linux SPARC. > > > > Unlike its predecessor, this series also fixes an underlying problem in > > which ncores() could return 0 which would result in chainlint.pl not > > processing any of its input test scripts. Patch [3/3] also fixes CPU > > count detection on Alpha[2]. > > Works as expected on my Linux SPARC machine running Debian unstable. > > Tested-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> Thanks for testing. Were you able to check whether it fixes CPU count detection on Alpha, as well?
On Mon, 2024-05-20 at 15:19 -0400, Eric Sunshine wrote: > On Mon, May 20, 2024 at 3:17 PM John Paul Adrian Glaubitz > <glaubitz@physik.fu-berlin.de> wrote: > > On Mon, 2024-05-20 at 15:01 -0400, Eric Sunshine wrote: > > > From: Eric Sunshine <sunshine@sunshineco.com> > > > This series replaces a patch[1] sent by John Paul Adrian Glaubitz to fix > > > chainlint.pl CPU count computation on Linux SPARC. > > > > > > Unlike its predecessor, this series also fixes an underlying problem in > > > which ncores() could return 0 which would result in chainlint.pl not > > > processing any of its input test scripts. Patch [3/3] also fixes CPU > > > count detection on Alpha[2]. > > > > Works as expected on my Linux SPARC machine running Debian unstable. > > > > Tested-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> > > Thanks for testing. Were you able to check whether it fixes CPU count > detection on Alpha, as well? I can test on Alpha, but that will take a little longer as I don't have my setup ready. Will try to report back by tomorrow. Let me CC Michael Cree and Matt Turner who both own fast Alpha machines and might report back faster. @Michael, Matt: Could you test this patch series against the current git development tree? It should fix the testsuite on Alpha. Adrian
Hi Eric, On Mon, 2024-05-20 at 21:23 +0200, John Paul Adrian Glaubitz wrote: > On Mon, 2024-05-20 at 15:19 -0400, Eric Sunshine wrote: > > Thanks for testing. Were you able to check whether it fixes CPU count > > detection on Alpha, as well? > > I can test on Alpha, but that will take a little longer as I don't have > my setup ready. Will try to report back by tomorrow. I have tested it now on single-core Alpha and it works as expected, so I think it's safe to land the patches. I currently cannot test on SMP as I need to build a custom kernel for that first which disables a problematic kernel option. Adrian
On Tue, May 21, 2024 at 10:28 AM John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> wrote: > On Mon, 2024-05-20 at 21:23 +0200, John Paul Adrian Glaubitz wrote: > > On Mon, 2024-05-20 at 15:19 -0400, Eric Sunshine wrote: > > > Thanks for testing. Were you able to check whether it fixes CPU count > > > detection on Alpha, as well? > > > > I can test on Alpha, but that will take a little longer as I don't have > > my setup ready. Will try to report back by tomorrow. > > I have tested it now on single-core Alpha and it works as expected, so I > think it's safe to land the patches. Thank you for testing.
From: Eric Sunshine <sunshine@sunshineco.com> This series replaces a patch[1] sent by John Paul Adrian Glaubitz to fix chainlint.pl CPU count computation on Linux SPARC. Unlike its predecessor, this series also fixes an underlying problem in which ncores() could return 0 which would result in chainlint.pl not processing any of its input test scripts. Patch [3/3] also fixes CPU count detection on Alpha[2]. Patch [2/3] of this series is more or less Adrian's original patch[1] so it retains his authorship, though I simplified the regular-expression and tweaked the commit message. [1]: https://lore.kernel.org/git/20240520111109.99882-1-glaubitz@physik.fu-berlin.de/ [2]: https://lore.kernel.org/git/503a99f3511559722a3eeef15d31027dfe617fa1.camel@physik.fu-berlin.de/ Eric Sunshine (2): chainlint.pl: make CPU count computation more robust chainlint.pl: latch CPU count directly reported by /proc/cpuinfo John Paul Adrian Glaubitz (1): chainlint.pl: fix incorrect CPU count on Linux SPARC t/chainlint.pl | 20 +++++++++++++++++--- 1 file changed, 17 insertions(+), 3 deletions(-)