kern: cpuset: drop the lock to allocate domainsets

Restructure the loop a little bit to make it a little more clear how it
really operates: we never allocate any domains at the beginning of the first
iteration, and it will run until we've satisfied the amount we need or we
encounter an error.

The lock is now taken outside of the loop to make stuff inside the loop
easier to evaluate w.r.t. locking.

This fixes it to not try and allocate any domains for the freelist under the
spinlock, which would have happened before if we needed any new domains.

Reported by:	syzbot+6743fa07b9b7528dc561@syzkaller.appspotmail.com
Reviewed by:	markj
MFC after:	3 days
Differential Revision:	https://reviews.freebsd.org/D27371
This commit is contained in:
Kyle Evans 2020-11-28 01:21:11 +00:00
parent 15add60d37
commit e07e3fa3c9

View File

@ -794,14 +794,11 @@ cpuset_modify_domain(struct cpuset *set, struct domainset *domain)
return (EPERM);
domainset_freelist_init(&domains, 0);
domain = domainset_create(domain);
ndomains = needed = 0;
do {
if (ndomains < needed) {
domainset_freelist_add(&domains, needed - ndomains);
ndomains = needed;
}
ndomains = 0;
mtx_lock_spin(&cpuset_lock);
for (;;) {
root = cpuset_getroot(set);
mtx_lock_spin(&cpuset_lock);
dset = root->cs_domain;
/*
* Verify that we have access to this set of domains.
@ -826,7 +823,15 @@ cpuset_modify_domain(struct cpuset *set, struct domainset *domain)
&needed, 0);
if (error)
goto out;
} while (ndomains < needed);
if (ndomains >= needed)
break;
/* Dropping the lock; we'll need to re-evaluate again. */
mtx_unlock_spin(&cpuset_lock);
domainset_freelist_add(&domains, needed - ndomains);
ndomains = needed;
mtx_lock_spin(&cpuset_lock);
}
dset = set->cs_domain;
cpuset_update_domain(set, domain, dset, &domains);
out: