Re: mmap( MAP_ANON) is broken on current. (was Still seeing Failed assertion: "p[i] == 0" on armv7 buildworld)

On Sat, Nov 22, 2025 at 09:37:15AM -0800, Mark Millard wrote:
> Michal Meloun <mmel_at_freebsd.org> wrote on
> Date: Sat, 22 Nov 2025 16:37:19 UTC :
> 
> > On 22.11.2025 16:40, Konstantin Belousov wrote:
> > > On Sat, Nov 22, 2025 at 03:31:24PM +0100, Michal Meloun wrote:
> > >> This patch KASSERTs almost immediately when the system enters multi-user
> > >> mode while processing mmap() syscall:
> > >>
> > >> panic: vm_object_coalesce: obj 0xc73ddb28 next_pindex 0x13 next_size 0x5
> > >> obj_size 0x176
> > > 
> > > Yes, the assert was mis-placed. Please try this variant.
> > > 
> > > commit 2b1a1bcd2926bd89b8422c665b0aa411e29c883b
> > > Author: Konstantin Belousov <kib@FreeBSD.org>
> > > Date: Sat Nov 22 16:02:50 2025 +0200
> > > 
> > > vm_object_coalesce(): fix logic to detect coalesce possibility, simplify
> > > 
> > > diff --git a/sys/vm/vm_object.c b/sys/vm/vm_object.c
> > > index 5b4517d2bf0c..9bb4e54edd96 100644
> > > --- a/sys/vm/vm_object.c
> > > +++ b/sys/vm/vm_object.c
> > > @@ -2189,13 +2189,19 @@ vm_object_coalesce(vm_object_t prev_object, vm_ooffset_t prev_offset,
> > > next_size >>= PAGE_SHIFT;
> > > next_pindex = OFF_TO_IDX(prev_offset) + prev_size;
> > > 
> > > - if (prev_object->ref_count > 1 &&
> > > - prev_object->size != next_pindex &&
> > > + if (prev_object->ref_count > 1 ||
> > > + prev_object->size != next_pindex ||
> > > (prev_object->flags & OBJ_ONEMAPPING) == 0) {
> > > VM_OBJECT_WUNLOCK(prev_object);
> > > return (FALSE);
> > > }
> > > 
> > > + KASSERT(next_pindex + next_size > prev_object->size,
> > > + ("vm_object_coalesce: "
> > > + "obj %p next_pindex %#jx next_size %#jx obj_size %#jx",
> > > + prev_object, (uintmax_t)next_pindex, (uintmax_t)next_size,
> > > + (uintmax_t)prev_object->size));
> > > +
> > > /*
> > > * Account for the charge.
> > > */
> > > @@ -2222,26 +2228,13 @@ vm_object_coalesce(vm_object_t prev_object, vm_ooffset_t prev_offset,
> > > * Remove any pages that may still be in the object from a previous
> > > * deallocation.
> > > */
> > > - if (next_pindex < prev_object->size) {
> > > - vm_object_page_remove(prev_object, next_pindex, next_pindex +
> > > - next_size, 0);
> > > -#if 0
> > > - if (prev_object->cred != NULL) {
> > > - KASSERT(prev_object->charge >=
> > > - ptoa(prev_object->size - next_pindex),
> > > - ("object %p overcharged 1 %jx %jx", prev_object,
> > > - (uintmax_t)next_pindex, (uintmax_t)next_size));
> > > - prev_object->charge -= ptoa(prev_object->size -
> > > - next_pindex);
> > > - }
> > > -#endif
> > > - }
> > > + vm_object_page_remove(prev_object, next_pindex, next_pindex +
> > > + next_size, 0);
> > > 
> > > /*
> > > * Extend the object if necessary.
> > > */
> > > - if (next_pindex + next_size > prev_object->size)
> > > - prev_object->size = next_pindex + next_size;
> > > + prev_object->size = next_pindex + next_size;
> > > 
> > > VM_OBJECT_WUNLOCK(prev_object);
> > > return (TRUE);
> > 
> > Unfortunately, that didn't help. I will try the vm_map.c patch again 
> > for confirmation.
> 
> On amd64 I could not complete a boot: the KASSERT failed for equality
> instead of > :   "next_pindex Oxf next_size 0x4 obj_size 0x19"
> 
> QUOTE (from a prior message to the list):
> No serial console so a summary from a picture
> (expect typos):
> 
> . . .
> ue0: link state changed to UP
> panic: vm_object_coalesce: obj Oxfffff800090a27c0 next_pindex Oxf next_size 0x4 obj_size 0x19

Remove this assertion from your source.  The right fix is to move it later,
but it does not matter for the testing.