patchwork-bot incorrect reports

patchwork-bot incorrect reports

[Geert Uytterhoeven]

Hi Konstantin,

For quite a while, sometimes the patchwork-bot incorrectly reports that
a patch was applied.  Usually this happens when a series is applied
partially.  In that case, sometimes one or more of the unapplied
patches still appear in the patchwork-bot email report, while they are
still marked as "Action Required" in patchwork, so no real harm is done.

However, today I got a report[1] that is worse:
  1. It claims that Jakub Kicinski applied a patch to my repository,
  2. Its patchwork state[2] was changed to "Mainlined",
  3.  Its patchwork entry points to an unrelated merge commit in netdev[3].

Do you know what's going on?
Thanks!

[1] https://lore.kernel.org/all/174703868703.506001.16395245401308177497.git-patchwork-summary@kernel.org
[2] https://patchwork.kernel.org/project/linux-renesas-soc/patch/TYRPR01MB14284C8E8145635A1AD8314F4828AA@TYRPR01MB14284.jpnprd01.prod.outlook.com
[3] https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/commit/?id=ccb52a9c8dfaa4e2cbb2524b68a7ead72038b039
Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: patchwork-bot incorrect reports

[Jakub Kicinski]

On Mon, 12 May 2025 15:27:57 +0200 Geert Uytterhoeven wrote:
> Hi Konstantin,
> 
> For quite a while, sometimes the patchwork-bot incorrectly reports that
> a patch was applied.  Usually this happens when a series is applied
> partially.  In that case, sometimes one or more of the unapplied
> patches still appear in the patchwork-bot email report, while they are
> still marked as "Action Required" in patchwork, so no real harm is done.
> 
> However, today I got a report[1] that is worse:
>   1. It claims that Jakub Kicinski applied a patch to my repository,
>   2. Its patchwork state[2] was changed to "Mainlined",
>   3.  Its patchwork entry points to an unrelated merge commit in netdev[3].
> 
> Do you know what's going on?
> Thanks!
> 
> [1] https://lore.kernel.org/all/174703868703.506001.16395245401308177497.git-patchwork-summary@kernel.org
> [2] https://patchwork.kernel.org/project/linux-renesas-soc/patch/TYRPR01MB14284C8E8145635A1AD8314F4828AA@TYRPR01MB14284.jpnprd01.prod.outlook.com
> [3] https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/commit/?id=ccb52a9c8dfaa4e2cbb2524b68a7ead72038b039

Hm, weird. I wonder if I'm doing something special. BPF had a similar
problem and it's usually my merge commits.

Maybe it's because I add https://patch.msgid.link URLs to merge commits,
which are technically not patches? Should I switch to lore links on
merges but stick to patch.msgid.link on patches?

Re: patchwork-bot incorrect reports

[Konstantin Ryabitsev]

On Mon, May 12, 2025 at 10:31:55AM -0700, Jakub Kicinski wrote:
> > For quite a while, sometimes the patchwork-bot incorrectly reports that
> > a patch was applied.  Usually this happens when a series is applied
> > partially.  In that case, sometimes one or more of the unapplied
> > patches still appear in the patchwork-bot email report, while they are
> > still marked as "Action Required" in patchwork, so no real harm is done.
> > 
> > However, today I got a report[1] that is worse:
> >   1. It claims that Jakub Kicinski applied a patch to my repository,
> >   2. Its patchwork state[2] was changed to "Mainlined",
> >   3.  Its patchwork entry points to an unrelated merge commit in netdev[3].
> > 
> > Do you know what's going on?
> > Thanks!
> > 
> > [1] https://lore.kernel.org/all/174703868703.506001.16395245401308177497.git-patchwork-summary@kernel.org
> > [2] https://patchwork.kernel.org/project/linux-renesas-soc/patch/TYRPR01MB14284C8E8145635A1AD8314F4828AA@TYRPR01MB14284.jpnprd01.prod.outlook.com
> > [3] https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/commit/?id=ccb52a9c8dfaa4e2cbb2524b68a7ead72038b039
> 
> Hm, weird. I wonder if I'm doing something special. BPF had a similar
> problem and it's usually my merge commits.
> 
> Maybe it's because I add https://patch.msgid.link URLs to merge commits,
> which are technically not patches? Should I switch to lore links on
> merges but stick to patch.msgid.link on patches?

That's my conclusion as well -- the bot is naive in how it treats the links
and will match a commit to any message when it sees a Link: that leads to a
known patch. I'm working on adding some logic to have it be more discerning.

-K