Re: [RFC PATCH 1/1] idr: do not create idr if new id would be outside given range

[Matthew Wilcox]

On Thu, Nov 27, 2025 at 03:03:20PM +0100, Christian König wrote:
> On 11/27/25 14:54, Matthew Wilcox wrote:
> > On Thu, Nov 27, 2025 at 10:27:32AM +0100, Jan Sokolowski wrote:
> >> A scenario was found where trying to add id in range 0,1
> >> would return an id of 2, which is outside the range and thus
> >> now what the user would expect.
> > 
> > Can you do a bit better with this bug report?  Under what circumstances
> > does this happen?  Preferably answer in the form of a test case for the
> > IDR test suite.  Here's my attempt to recreate your situation based on
> > what I read in that thread.  It doesn't show a problem, so clearly I got
> > something wrong.
> 
> According to Jan the observation he has is that this code:
> 
> idr_init_base(&idr, 1);
> id = idr_alloc(&idr, dummy_ptr, 0, 1, GFP_KERNEL);
> 
> Gives him id=2 in return.

Hm.  That's not what it does for me.  It gives me id == 1, which isn't
correct!  I'll look into that, but it'd be helpful to know what
combination of inputs gives us 2.

To be completely clear, here's what I'm looking at:

+void idr_alloc2_test(void)
+{
+       int id;
+       struct idr idr = IDR_INIT_BASE(idr, 1);
+
+       id = idr_alloc(&idr, idr_alloc2_test, 0, 1, GFP_KERNEL);
+       printf("id = %d\n", id);

and I think that should return -ENOSPC instead of 1, since we told it to
allocate exclusive of 1.