|author||Thomas Schwinge <firstname.lastname@example.org>||2013-09-25 21:45:38 +0200|
|committer||Thomas Schwinge <email@example.com>||2013-09-25 21:45:38 +0200|
Diffstat (limited to 'open_issues/exec_memory_leaks.mdwn')
1 files changed, 51 insertions, 3 deletions
diff --git a/open_issues/exec_memory_leaks.mdwn b/open_issues/exec_memory_leaks.mdwn
index d504c4f..67281bd 100644
@@ -1,4 +1,4 @@
-[[!meta copyright="Copyright © 2012 Free Software Foundation, Inc."]]
+[[!meta copyright="Copyright © 2012, 2013 Free Software Foundation, Inc."]]
[[!meta license="""[[!toggle id="license" text="GFDL 1.2+"]][[!toggleable
id="license" text="Permission is granted to copy, distribute and/or modify this
@@ -12,8 +12,56 @@ License|/fdl]]."]]"""]]
There are is some memory leak in [[`exec`|hurd/translator/exec]].
+# IRC, freenode, #hurd, 2012-08-11
+ <braunr> the exec servers seems to leak a lot
+ <braunr> server*
+ <braunr> exec now uses 109M on darnassus
+ <braunr> it really leaks a lot
+ <pinotree> only 109mb? few months ago, exec on exodar was taking more than
+ 200mb after few days of uptime with builds done
+ <braunr> i wonder how much it takes on the buildds
+## IRC, freenode, #hurd, 2012-08-17
+ <braunr> the exec leak is tricky
+ <braunr> bddebian: btw, look at the TODO file in the hurd source code
+ <braunr> bddebian: there is a not from thomas bushnell about that
+ <braunr> "*** Handle dead name notifications on execserver ports. !
+ <braunr> not sure it's still a todo item, but it might be worth checking
+ <bddebian> braunr: diskfs_execboot_class = ports_create_class (0, 0);
+ This is what would need to change right? It should call some cleanup
+ routine in the first argument?
+ <bddebian> Would be ideal if it could just use deadboot() from exec.
+ <braunr> bddebian: possible
+ <braunr> bddebian: hum execboot, i'm not so sure
+ <bddebian> Execboot is the exec task, no?
+ <braunr> i don't know what execboot is
+ <bddebian> It's from libdiskfs
+ <braunr> but "diskfs_execboot_class" looks like a class of ports used at
+ startup only
+ <braunr> ah
+ <braunr> then it's something run in the diskfs users ?
+ <bddebian> yes
+ <braunr> the leak is in exec
+ <braunr> if clients misbehave, it shouldn't affect that server
+ <bddebian> That's a different issue, this was about the TODO thing
+ <braunr> ah
+ <braunr> i don't know
+ <bddebian> Me either :)
+ <bddebian> For the leak I'm still focusing on do-bunzip2 but I am baffled
+ at my results..
+ <braunr> ?
+ <bddebian> Where my counters are zero if I always increment on different
+ vars but wild freaking numbers if I increment on malloc and decrement on
After twelve hours worth of `fork/exec` ([[GCC]]'s `check-c` part of the
testsuite), we got:
@@ -29,7 +77,7 @@ quite noticeable. In comparison:
276 0 3 1 1 344 442M 28.2M 0.6 48:09.36 91min /hurd/ext2fs /dev/hd2s5
After running the libtool testsuite for some time: