Re: bcm283x_dwcotg0: usbpf: Out of memory

Mark Millard <[email protected]> Fri, 17 Apr 2026 10:45:27 -0700
Newsgroups gmane.os.freebsd.devel.arm
Message-ID <[email protected]>
On 4/17/26 01:37, void wrote:
> On Thu, Apr 16, 2026 at 11:27:32AM -0700, bob prohaska wrote:
>>
>> It turns out a 1.8GB swap partition does not provoke the "excess
>> swap" warning and seems sufficient for a -j3 buildworld to proceed.
>> -j4 works usually, but if Clang is fully rebuilt in -j4 the machine
>> _can_ run out of swap and kill processes.
>>
>> Using swap on microSD or USB flash led (I think) to subtle binary
>> corruption that you helped me sort out. Since then I've used only
>> mechanical hard disks for swap and have not seen any obvious repeats.
> 
> Even with -j3 and smaller swap slices (1.5GBx3) I'd get the odd crash,

I'll note that the warnings about potential mis-tuning are based on the
total swap space enabled, not based on separate swap-space sizes that
contribute.

> always in a different place. Since then, I've modified the build like
> this:
> 
> 1. in /stc/src.conf there's WITHOUT_CLEAN=
> 2. in /etc/src-env.conf there's WITH_META_MODE= and filemon is loaded.
> 3. ccache is in use

Normally WITH_META_MODE use automatically sets WITHOUT_CLEAN (NO_CLEAN)
when appropriate (or acts as if it had). Explicitly setting
WITHOUT_CLEAN (NO_CLEAN) might set it when such is inappropriate. (A
possibility might be when newvers.sh has changed.)

FYI: To my knowledge, Bob P. does not use ccache for his build
activities. (Nor do I.) That might make comparison/contrast of results
messy.

> 
> I'm using -j2 right now because this is building a -current
> buildworld where userland is still 15-stable (but kernel is -current)
> so maybe there's an extra stress point there.
> 
> Also this is the first (ongoing) buildworld with meta mode.


-- 
===
Mark Millard
marklmi at yahoo.com