* [9front] lock: Fix some memory leaks
@ 2023-01-21 0:17 Josiah Frentsos
2023-01-21 7:55 ` Steve Simon
2023-01-22 20:46 ` cinap_lenrek
0 siblings, 2 replies; 6+ messages in thread
From: Josiah Frentsos @ 2023-01-21 0:17 UTC (permalink / raw)
To: 9front
diff bb36ba0617b5aa8263ea9b5ece8c1a5249fedc86 uncommitted
--- a/sys/src/cmd/lock.c
+++ b/sys/src/cmd/lock.c
@@ -35,6 +35,7 @@
}
if (w->pid == pid)
return w;
+ free(w);
}
}
@@ -141,7 +142,8 @@
error("wait");
postnote(PNPROC, lckpid, "die");
- waitfor(lckpid);
+ free(waitfor(lckpid));
+
if(w->msg[0]){
p = utfrune(w->msg, ':');
if(p && p[1])
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [9front] lock: Fix some memory leaks
2023-01-21 0:17 [9front] lock: Fix some memory leaks Josiah Frentsos
@ 2023-01-21 7:55 ` Steve Simon
2023-01-22 0:36 ` Dan Cross
2023-01-22 20:46 ` cinap_lenrek
1 sibling, 1 reply; 6+ messages in thread
From: Steve Simon @ 2023-01-21 7:55 UTC (permalink / raw)
To: 9front
sorry, but this patch looks very wrong.
free after return does nothing.
calling waitfor twice allocates data twice.
i would also consider what happens after the waitfor() call - i have not looked at the source but i have a suspicion that lock exits, so there is no point in freeing memory anyway.
-Steve
> On 21 Jan 2023, at 12:18 am, Josiah Frentsos <jfrent@tilde.team> wrote:
>
> diff bb36ba0617b5aa8263ea9b5ece8c1a5249fedc86 uncommitted
> --- a/sys/src/cmd/lock.c
> +++ b/sys/src/cmd/lock.c
> @@ -35,6 +35,7 @@
> }
> if (w->pid == pid)
> return w;
> + free(w);
> }
> }
>
> @@ -141,7 +142,8 @@
> error("wait");
>
> postnote(PNPROC, lckpid, "die");
> - waitfor(lckpid);
> + free(waitfor(lckpid));
> +
> if(w->msg[0]){
> p = utfrune(w->msg, ':');
> if(p && p[1])
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [9front] lock: Fix some memory leaks
2023-01-21 7:55 ` Steve Simon
@ 2023-01-22 0:36 ` Dan Cross
2023-01-22 7:09 ` Benjamin Purcell
0 siblings, 1 reply; 6+ messages in thread
From: Dan Cross @ 2023-01-22 0:36 UTC (permalink / raw)
To: 9front
On Sat, Jan 21, 2023 at 3:52 AM Steve Simon <steve@quintile.net> wrote:
> sorry, but this patch looks very wrong.
>
> free after return does nothing.
> calling waitfor twice allocates data twice.
I've got no real context here, but a couple of brief observations: The
`return` mentioned above is conditional on some pid having an expected
value. The `free` added in this patch is for the non-return case where
the conditional does not evaluate to true.
I don't see how `waitfor` is called twice in this patch. Rather, it
appears that a call to `waitfor(...)` was replaced with
`free(waitfor(...))`
> i would also consider what happens after the waitfor() call - i have not looked at the source but i have a suspicion that lock exits, so there is no point in freeing memory anyway.
It does appear that `lock` exits here.
- Dan C.
> > On 21 Jan 2023, at 12:18 am, Josiah Frentsos <jfrent@tilde.team> wrote:
> >
> > diff bb36ba0617b5aa8263ea9b5ece8c1a5249fedc86 uncommitted
> > --- a/sys/src/cmd/lock.c
> > +++ b/sys/src/cmd/lock.c
> > @@ -35,6 +35,7 @@
> > }
> > if (w->pid == pid)
> > return w;
> > + free(w);
> > }
> > }
> >
> > @@ -141,7 +142,8 @@
> > error("wait");
> >
> > postnote(PNPROC, lckpid, "die");
> > - waitfor(lckpid);
> > + free(waitfor(lckpid));
> > +
> > if(w->msg[0]){
> > p = utfrune(w->msg, ':');
> > if(p && p[1])
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [9front] lock: Fix some memory leaks
2023-01-22 0:36 ` Dan Cross
@ 2023-01-22 7:09 ` Benjamin Purcell
2023-01-22 8:11 ` Steve Simon
0 siblings, 1 reply; 6+ messages in thread
From: Benjamin Purcell @ 2023-01-22 7:09 UTC (permalink / raw)
To: 9front
My guess would be that Steve's email client does something weird with leading whitespace.
On Sat, Jan 21, 2023, at 18:36, Dan Cross wrote:
> On Sat, Jan 21, 2023 at 3:52 AM Steve Simon <steve@quintile.net> wrote:
>> sorry, but this patch looks very wrong.
>>
>> free after return does nothing.
>> calling waitfor twice allocates data twice.
>
> I've got no real context here, but a couple of brief observations: The
> `return` mentioned above is conditional on some pid having an expected
> value. The `free` added in this patch is for the non-return case where
> the conditional does not evaluate to true.
>
> I don't see how `waitfor` is called twice in this patch. Rather, it
> appears that a call to `waitfor(...)` was replaced with
> `free(waitfor(...))`
>
>> i would also consider what happens after the waitfor() call - i have not looked at the source but i have a suspicion that lock exits, so there is no point in freeing memory anyway.
>
> It does appear that `lock` exits here.
>
> - Dan C.
>
>> > On 21 Jan 2023, at 12:18 am, Josiah Frentsos <jfrent@tilde.team> wrote:
>> >
>> > diff bb36ba0617b5aa8263ea9b5ece8c1a5249fedc86 uncommitted
>> > --- a/sys/src/cmd/lock.c
>> > +++ b/sys/src/cmd/lock.c
>> > @@ -35,6 +35,7 @@
>> > }
>> > if (w->pid == pid)
>> > return w;
>> > + free(w);
>> > }
>> > }
>> >
>> > @@ -141,7 +142,8 @@
>> > error("wait");
>> >
>> > postnote(PNPROC, lckpid, "die");
>> > - waitfor(lckpid);
>> > + free(waitfor(lckpid));
>> > +
>> > if(w->msg[0]){
>> > p = utfrune(w->msg, ':');
>> > if(p && p[1])
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [9front] lock: Fix some memory leaks
2023-01-21 0:17 [9front] lock: Fix some memory leaks Josiah Frentsos
2023-01-21 7:55 ` Steve Simon
@ 2023-01-22 20:46 ` cinap_lenrek
1 sibling, 0 replies; 6+ messages in thread
From: cinap_lenrek @ 2023-01-22 20:46 UTC (permalink / raw)
To: 9front
applied!
--
cinap
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2023-01-22 20:48 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-01-21 0:17 [9front] lock: Fix some memory leaks Josiah Frentsos
2023-01-21 7:55 ` Steve Simon
2023-01-22 0:36 ` Dan Cross
2023-01-22 7:09 ` Benjamin Purcell
2023-01-22 8:11 ` Steve Simon
2023-01-22 20:46 ` cinap_lenrek
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).