[SRU][J/N:linux-bluefield][PATCH v1 0/2] Updates to mlxbf-pmc

[David Thompson]

> -----Original Message-----
> From: Kuba Pawlak <kuba.pawlak at canonical.com>
> Sent: Wednesday, July 16, 2025 4:02 PM
> To: David Thompson <davthompson at nvidia.com>; Shravan Ramani
> <shravankr at nvidia.com>; kernel-team at lists.ubuntu.com
> Subject: Re: [SRU][J/N:linux-bluefield][PATCH v1 0/2] Updates to mlxbf-pmc
> 
> On 16.07.2025 19:19, David Thompson wrote:
> >> -----Original Message-----
> >> From: Shravan Ramani <shravankr at nvidia.com>
> >> Sent: Wednesday, July 9, 2025 5:20 AM
> >> To: kernel-team at lists.ubuntu.com
> >> Cc: Shravan Ramani <shravankr at nvidia.com>; David Thompson
> >> <davthompson at nvidia.com>
> >> Subject: [SRU][J/N:linux-bluefield][PATCH v1 0/2] Updates to
> >> mlxbf-pmc
> >>
> >> This submission addresses the following 2 issues:
> >> 1. Fix duplicate event ID in event_list
> >> BugLink: https://bugs.launchpad.net/bugs/2115400
> >>
> >> SRU Justification:
> >>
> >> [Impact]
> >> Same ID (103) was assigned to both GDC_BANK0_G_RSE_PIPE_CACHE_DATA0
> >> and GDC_BANK0_G_RSE_PIPE_CACHE_DATA1. This could lead to incorrect
> >> event mapping.
> >>
> >> [Fix]
> >> The fix is to remove the duplicate ID and assign the correct ID
> >> instead to GDC_BANK0_G_RSE_PIPE_CACHE_DATA1.
> >>
> >> [Test Case]
> >> Without the fix, programming event ID of 104 would return an error
> >> and there would be no way to assign the counter to track
> >> GDC_BANK0_G_RSE_PIPE_CACHE_DATA1.
> >> With the fix, event ID 104 should be programmable.
> >>
> >> [Regression Potential]
> >> Minimal
> >>
> >> 2. Validate user input for event/enable settings
> >> BugLink: https://bugs.launchpad.net/bugs/2113758
> >>
> >> SRU Justification:
> >>
> >> [Impact]
> >> The enable and event sysfs files are used to program and enable the
> >> performance counters. However, they also seem to take invalid input,
> >> which should be handled by the driver with an error.
> >>
> >> [Fix]
> >> Check "enable" input to allow only 0 or 1, and not other values. For "event"
> >> info, check if the user input is a valid event number for the
> >> corresponding block and if not, print an error message.
> >>
> >> [Test Case]
> >> Try invalid values to the "enable" sysfs in any block and it should
> >> not be accepted by the driver.
> >> For event sysfs, enter any event number that is not listed in the "event_list"
> >> and validate that the driver does not accept this input.
> >>
> >> [Regression Potential]
> >> Minimal
> >>
> >> Alok Tiwari (1):
> >>    platform/mellanox: mlxbf-pmc: Fix duplicate event ID for
> >> CACHE_DATA1
> >>
> >> Shravan Kumar Ramani (1):
> >>    UBUNTU: SAUCE: mlxbf-pmc: Validate user input for event/enable
> >>
> >>   drivers/platform/mellanox/mlxbf-pmc.c | 26 +++++++++++++++++---------
> >>   1 file changed, 17 insertions(+), 9 deletions(-)
> >>
> >> --
> >> 2.43.2
> > Kuba,
> >
> > I see this patch series was applied to Noble, but not Jammy.
> > 	Shravan Kumar Ramani      88228a17a14f UBUNTU: SAUCE: mlxbf-pmc:
> Validate user input for event/enable
> > 	Alok Tiwari               6df2ff7f1b0e platform/mellanox: mlxbf-pmc: Fix
> duplicate event ID for CACHE_DATA1
> >
> > The patch was sent with "J/N" tags and we're hoping to get it applied to Jammy
> as well.
> >
> > Can you check on your side if this can be applied to Jammy and part of next
> Jammy build?
> > I see that this series does cleanly apply to Jammy.
> >
> > - Dave
> 
> hi,
> 
> sorry about that, commits added
> 
> - Kuba

Thank you so much Kuba.

- Dave