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

On Nov 16, 2025, at 18:03, Konstantin Belousov <[email protected]> wrote:

> On Sun, Nov 16, 2025 at 10:11:15PM +0100, Michal Meloun wrote:
>> 
>> 
>> On 16.11.2025 18:51, Warner Losh wrote:
>>> Maybe try main with the following patch. Adrian noticed the TLS
>>> mismatch. I don't think it will matter, but TLS thread model stuff
>>> always gives me a big headache. If the following fails to apply, just
>>> copy the JEMALLOC_TLS_MODEL line from i386 to arm. The default changed
>>> elsewhere, but this wasn't updated here.
>>> 
>>> Warner
>> 
>> Unfortunately, that doesn't help. I'm out of ideas on how to debug this, all
>> of my attempts have failed.
>> 
>> The problem only occurs when Clang compiles a larger project and is
>> intermediate. Attempt to compile the clang generated reproducer is always
>> successful.
>> It's clear that the parallelism introduced by make plays a significant role.
>> But the system never reached an OOM condition before failure.
> 
> Is the problem reproducable on armv7 userspace running on arm64 kernel


Yes, I've replicated the problem on a Windows Dev Kit 2023 in an
armv7 chroot with the system boot's having hw.physmem=2G via
/boot/loader.conf and the SWAP space sets to 3.5 GiBytes.
"env META_MODE= make -j8 buildworld" was executed in the chroot
until it failed with the assert Bob P. has been getting.

The aarch64 system was an installation of an official pkgbase
distribution, not a personal build. Same for the world in the
chroot. All main [so: 16 at this point].

Absent the memory pressure from the small hw.physmem, I did not
get the failure.


I also tried experiments with aarch64 doing an armv7 buildworld
with some small sizes for hw.physmem . I got no failures. So far,
only armv7 has been shown to have the issue.
 

> , or only
> on native armv7 host?

The armv7 kernel does not need to be involved at all: just
the armv7 world.


===
Mark Millard
marklmi at yahoo.com