Hello,
Brent W. Baccala, on ven. 05 janv. 2018 19:47:37 -0500, wrote:
> Well, in that case, perhaps it should work by checking to see if the port is
> in
> the hash, rather than by looking at how many weak references it's got.
> Something more like this:
>
> if ((refcounts_hard_references(&i->pi.refcounts) == 0)
> && (* i->id_hashloc == i))
> {
> /* Nobody got a send right in between, we can remove from the hash. */
> hurd_ihash_locp_remove (&idhash, i->id_hashloc);
> ports_port_deref_weak (&i->pi);
> }
You can't afford dereferencing id_hashloc like that: it could have
gotten away in between.
Could you try the attached patch instead?
> Also, what documentation? The hurd texinfo file? That's the only place that
> I
> know of where this is documented.
Yes: “If nonzero, @var{dropweak_routine} will be called to request
weak references to be dropped.” It doesn't say "one weak reference", but
"weak references".
Samuel
diff --git a/libfshelp/get-identity.c b/libfshelp/get-identity.c
index f88e0f82..ac1b4f0b 100644
--- a/libfshelp/get-identity.c
+++ b/libfshelp/get-identity.c
@@ -70,10 +70,12 @@ id_clean (void *cookie)
{
struct idspec *i = cookie;
pthread_mutex_lock (&idlock);
- if (refcounts_hard_references(&i->pi.refcounts) == 0)
+ if (refcounts_hard_references(&i->pi.refcounts) == 0
+ && i->id_hashloc != NULL)
{
- /* Nobody got a send right in between, we can remove from the hash. */
+ /* Nobody got a send right in between, we can remove i from the hash. */
hurd_ihash_locp_remove (&idhash, i->id_hashloc);
+ i->id_hashloc = NULL;
ports_port_deref_weak (&i->pi);
}
pthread_mutex_unlock (&idlock);
