Date: Sun, 28 May 2017 10:10:47 +0200 From: =?UTF-8?Q?BERTRAND_Jo=c3=abl?= <joel.bertrand@systella.fr> To: freebsd-hackers@freebsd.org, pkgsrc-users@NetBSD.org Subject: Deadlock in pkg_jobs_update_universe_item_priority was Re: Issue with pkg upgrade on diskless workstation Message-ID: <6c18a445-93c6-8a65-cc0c-eedbef3eabc1@systella.fr> In-Reply-To: <d1818fc7-17ad-f6ca-97bb-585bc5a0f334@systella.fr> References: <aa873beb-7e5f-6591-33f8-107089b606df@systella.fr> <20170514082046.GA15092@lonesome.com> <fd582c6c-e8cf-6573-9db3-16756cf5410b@systella.fr> <fe9c666f-08ef-322d-ac3f-f13943bd1e06@systella.fr> <bf22b451-c1d4-053c-8965-2e12c4b5eef9@FreeBSD.org> <d1818fc7-17ad-f6ca-97bb-585bc5a0f334@systella.fr>
next in thread | previous in thread | raw e-mail | index | archive | help
Hello,
Deadlock I have seen is not related to NFS locks. I have added some
debug traces in libpkg and I have seen that pkg enters in deadlock in
pkg_jobs_update_universe_item_priority() function.
More precisely, a loop doesn't end :
line 712 of pkg_jobs_universe.c
while (rdeps_func(it->pkg, &d) == EPKG_OK) {
HASH_FIND_STR(universe->items, d->uid, found);
if (found == NULL)
{
continue;
}
LL_FOREACH(found, cur) {
...
}
}
Best regards,
JKB
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?6c18a445-93c6-8a65-cc0c-eedbef3eabc1>