Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 14 Dec 2021 15:03:08 +0000
From:      bugzilla-noreply@freebsd.org
To:        bugs@FreeBSD.org
Subject:   [Bug 260406] pfctl: Cannot allocate memory (after a time)
Message-ID:  <bug-260406-227-yYKrXBWVoG@https.bugs.freebsd.org/bugzilla/>
In-Reply-To: <bug-260406-227@https.bugs.freebsd.org/bugzilla/>
References:  <bug-260406-227@https.bugs.freebsd.org/bugzilla/>

next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D260406

--- Comment #9 from Kristof Provost <kp@freebsd.org> ---
>  2  36537            pfr_ina_define:return c ffffa000c1e48be0
>  2  36537            pfr_ina_define:return 0 0

What? That's ... that can't be right. The first number should be the offset=
 in
the function where we returned from. Both 0xc and 0x0 make no sense there. =
The
second value is the return value. That should be zero or an error number. N=
ot a
pointer like we're clearly getting here.

Sigh, it looks like Dtrace's fbt is buggy on aarch64, and indeed if I'm rea=
ding
sys/cddl/dev/fbt/aarch64/fbt_isa.c correctly we're passing x[0] / x[1] in t=
he
return probe, rather than the return offset and the return value. So we sho=
uld
look for the return value in arg0, and indeed we see a return value of 0x0c=
 (or
12, ENOMEM).

So we now know at least that the big allocation succeeds, but that
pfr_ina_define() or one of the functions it calls fail to allocate memory.
Let's dig a bit more.

Try:
dtrace -n 'fbt:kernel:pfr_create_ktable:return { printf("%x", arg0); }' -n
'fbt:kernel:pf_find_or_create_kruleset:return { printf("%x", arg0); }' -n
'fbt:kernel:pfr_create_kentry:return { printf("%x", arg0); }'

--=20
You are receiving this mail because:
You are the assignee for the bug.=



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-260406-227-yYKrXBWVoG>