Re: Still seeing Failed assertion: "p[i] == 0" on armv7 buildworld

On Dec 8, 2025, at 22:00, Mark Millard <[email protected]> wrote:

> There was a commit on "Mon, 08 Dec 2025":
> 
>    • git: cff67bc43df1 - main - vm_fault: only rely on PG_ZERO when the page was newly allocated Konstantin Belousov
> 
> This looks to be more changes related to avoiding the "p[i] == 0"
> problem taht jemalloc was detecting and reporting:
> 
> QUOTE
>    vm_fault: only rely on PG_ZERO when the page was newly allocated
> 
> If the fs->m page was found invalid on the object queue, PG_ZERO flag is
> stale. Track the source of the page in the new fault state variable
> m_needs_zero, and ignore PG_ZERO if the page did not came from the
> allocator.
> END QUOTE
> 
> It is indicated for a MFC in 1 week.

If I gather right, the following commits to main
are the completion of the intended changes for
the issues exposes by jemalloc 5.3.0's "p[i] == 0"
debug build based asserts :

Wed, 17 Dec 2025
    • git: bdfd695037a9 - main - swap_pager_seek_data(): improve assertions Konstantin Belousov 
    • git: 30ae88e3b217 - main - swap_pager_seek_data(): move the clipping at the object size to consumers Konstantin Belousov 
    • git: f94bfc469df9 - main - vm_object_coalesce(): check that coalescing does not revive stale pages Konstantin Belousov

They are scheduled for MFC after 1 week.


===
Mark Millard
marklmi at yahoo.com