Re: [Patch nf] nf_conntrack_extend: silent a memory leak warning

2018-04-16 Thread Pablo Neira Ayuso
On Fri, Mar 30, 2018 at 01:22:06PM -0700, Cong Wang wrote:
> The following memory leak is false postive:
> 
> unreferenced object 0x8f37f156fb38 (size 128):
>   comm "softirq", pid 0, jiffies 4294899665 (age 11.292s)
>   hex dump (first 32 bytes):
> 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b  
> 00 00 00 00 30 00 20 00 48 6b 6b 6b 6b 6b 6b 6b  0. .Hkkk
>   backtrace:
> [<4fda266a>] __kmalloc_track_caller+0x10d/0x141
> [<7b0a7e3c>] __krealloc+0x45/0x62
> [] nf_ct_ext_add+0xdc/0x133
> [<99b47fd8>] init_conntrack+0x1b1/0x392
> [<86dc36ec>] nf_conntrack_in+0x1ee/0x34b
> [<940592de>] nf_hook_slow+0x36/0x95
> [] nf_hook.constprop.43+0x1c3/0x1dd
> [] __ip_local_out+0xae/0xb4
> [<3e4192a6>] ip_local_out+0x17/0x33
> [] igmp_ifc_timer_expire+0x23e/0x26f
> [<6a8f3032>] call_timer_fn+0x14c/0x2a5
> [<650c1725>] __run_timers.part.34+0x150/0x182
> [<90e6946e>] run_timer_softirq+0x2a/0x4c
> [<4d1e7293>] __do_softirq+0x1d1/0x3c2
> [<4643557d>] irq_exit+0x53/0xa2
> [<29ddee8f>] smp_apic_timer_interrupt+0x22a/0x235

Applied, thanks.


[Patch nf] nf_conntrack_extend: silent a memory leak warning

2018-03-30 Thread Cong Wang
The following memory leak is false postive:

unreferenced object 0x8f37f156fb38 (size 128):
  comm "softirq", pid 0, jiffies 4294899665 (age 11.292s)
  hex dump (first 32 bytes):
6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b  
00 00 00 00 30 00 20 00 48 6b 6b 6b 6b 6b 6b 6b  0. .Hkkk
  backtrace:
[<4fda266a>] __kmalloc_track_caller+0x10d/0x141
[<7b0a7e3c>] __krealloc+0x45/0x62
[] nf_ct_ext_add+0xdc/0x133
[<99b47fd8>] init_conntrack+0x1b1/0x392
[<86dc36ec>] nf_conntrack_in+0x1ee/0x34b
[<940592de>] nf_hook_slow+0x36/0x95
[] nf_hook.constprop.43+0x1c3/0x1dd
[] __ip_local_out+0xae/0xb4
[<3e4192a6>] ip_local_out+0x17/0x33
[] igmp_ifc_timer_expire+0x23e/0x26f
[<6a8f3032>] call_timer_fn+0x14c/0x2a5
[<650c1725>] __run_timers.part.34+0x150/0x182
[<90e6946e>] run_timer_softirq+0x2a/0x4c
[<4d1e7293>] __do_softirq+0x1d1/0x3c2
[<4643557d>] irq_exit+0x53/0xa2
[<29ddee8f>] smp_apic_timer_interrupt+0x22a/0x235

because __krealloc() is not supposed to release the old
memory and it is released later via kfree_rcu(). Since this is
the only external user of __krealloc(), just mark it as not leak
here.

Cc: Pablo Neira Ayuso 
Cc: Jozsef Kadlecsik 
Cc: Florian Westphal 
Signed-off-by: Cong Wang 
---
 net/netfilter/nf_conntrack_extend.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/net/netfilter/nf_conntrack_extend.c 
b/net/netfilter/nf_conntrack_extend.c
index 9fe0ddc333fb..bd71a828ebde 100644
--- a/net/netfilter/nf_conntrack_extend.c
+++ b/net/netfilter/nf_conntrack_extend.c
@@ -71,6 +71,7 @@ void *nf_ct_ext_add(struct nf_conn *ct, enum nf_ct_ext_id id, 
gfp_t gfp)
rcu_read_unlock();
 
alloc = max(newlen, NF_CT_EXT_PREALLOC);
+   kmemleak_not_leak(old);
new = __krealloc(old, alloc, gfp);
if (!new)
return NULL;
-- 
2.13.0