Problem unfixing a bug report

[Alan Stern]

I tried unfixing a bug report, but nothing happened:

https://lore.kernel.org/linux-usb/97b26c3e-d025-44c9...@rowland.harvard.edu/

What's wrong? Is there any other way to get syzbot to run a test for
me?

Alan Stern

[Taras Madan]

Hi Alan,
Thanks for reporting.

Let's track it in https://github.com/google/syzkaller/issues/7131.

BR,
Taras

> --
> You received this message because you are subscribed to the Google Groups "syzkaller" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to syzkaller+...@googlegroups.com.
> To view this discussion visit https://groups.google.com/d/msgid/syzkaller/794dedb3-a3e9-4c3c-ae03-5b52c65b8e04%40rowland.harvard.edu.

[Alan Stern]

On Mon, Apr 27, 2026 at 12:43:10PM +0200, Taras Madan wrote:
> Hi Alan,
> Thanks for reporting.
>
> Let's track it in https://github.com/google/syzkaller/issues/7131.

I see; syzbot doesn't allow changes to bug reports after they are
closed. But what exactly is the difference between "fixed" and
"closed"? And why does the dashboard page list the status as "fixed"
rather than "closed"?

More to the point: Is there any way to convince syzbot to run the
reproducer for that bug with a new test patch? I want to try fixing the
original problem in a different way, and it would be a big help to use
syzbot to test the new approach.

Alan Stern

[Taras Madan]

Hi Alan,

The bug may be closed for several reasons:

1. The bug value went down (consider an old bug in a rapidly evolving tree like linux-next).

2. It doesn't reproduce for a long time.

3. Other scenarios.

You can see some here. 

> More to the point: Is there any way to convince syzbot to run the
reproducer for that bug with a new test patch?

I don't think it is possible. CC @Aleksandr Nogikh in case I'm missing something.

You can download the reproducer, build the old kernel locally and experiment locally.

It is the best idea I have so far.

BR,
Taras

[Alan Stern]

On Mon, Apr 27, 2026 at 05:36:32PM +0200, Taras Madan wrote:
> Hi Alan,
>

> The bug may be closed for several reasons:
> 1. The bug value went down (consider an old bug in a rapidly evolving tree
> like linux-next).
> 2. It doesn't reproduce for a long time.
> 3. Other scenarios.
>

> You can see some here <https://syzbot.org/upstream/invalid>.

But what about the other way around? That is, a bug that is fixed but
not closed?

> > More to the point: Is there any way to convince syzbot to run the
> reproducer for that bug with a new test patch?
>

> I don't think it is possible. CC @Aleksandr Nogikh <nog...@google.com> in

> case I'm missing something.
> You can download the reproducer, build the old kernel locally and
> experiment locally.
> It is the best idea I have so far.

Would this make a good enhancement request?

Alan