On Fri, Nov 16, 2012 at 08:25:06AM +0000, Konstantin Belousov wrote: > Author: kib > Date: Fri Nov 16 08:25:06 2012 > New Revision: 243142 > URL: http://svnweb.freebsd.org/changeset/base/243142 > > Log: > In pget(9), if PGET_NOTWEXIT flag is not specified, also search the > zombie list for the pid. This allows several kern.proc sysctls to > report useful information for zombies. > > Hold the allproc_lock around all searches instead of relocking it. > Remove private pfind_locked() from the new nfs client code. > > Requested and reviewed by: pjd > Tested by: pho > MFC after: 3 weeks [...] > @@ -364,12 +374,16 @@ pget(pid_t pid, int flags, struct proc * > struct proc *p; > int error; > > + sx_slock(&allproc_lock); > if (pid <= PID_MAX) > - p = pfind(pid); > + p = pfind_locked(pid); > else if ((flags & PGET_NOTID) == 0) > - p = pfind_tid(pid); > + p = pfind_tid_locked(pid); > else > p = NULL; > + if (p == NULL && (flags & PGET_NOTWEXIT) == 0) > + p = zpfind_locked(pid); > + sx_sunlock(&allproc_lock); > if (p == NULL) > return (ESRCH);
I think we should move zpfind_locked() under 'pid <= PID_MAX':
sx_slock(&allproc_lock);
if (pid <= PID_MAX) {
p = pfind_locked(pid);
if (p == NULL && (flags & PGET_NOTWEXIT) == 0)
p = zpfind_locked(pid);
} else if ((flags & PGET_NOTID) == 0) {
p = pfind_tid_locked(pid);
} else {
p = NULL;
}
sx_sunlock(&allproc_lock);
if (p == NULL)
return (ESRCH);
--
Pawel Jakub Dawidek http://www.wheelsystems.com
FreeBSD committer http://www.FreeBSD.org
Am I Evil? Yes, I Am! http://tupytaj.pl
pgpws7X3zEMAu.pgp
Description: PGP signature
