LVS
lvs-devel
Google
 
Web LinuxVirtualServer.org

Re: [PATCH 3/3] ipvs: fix get_curr_sync_buff

To: Tinggong Wang <wangtinggong@xxxxxxxxx>
Subject: Re: [PATCH 3/3] ipvs: fix get_curr_sync_buff
Cc: Wensong Zhang <wensong@xxxxxxxxxxxx>, Simon Horman <horms@xxxxxxxxxxxx>, lvs-devel@xxxxxxxxxxxxxxx
From: Julian Anastasov <ja@xxxxxx>
Date: Tue, 14 Dec 2010 01:32:36 +0200 (EET)

        Hello,

On Sun, 12 Dec 2010, Tinggong Wang wrote:

use time_after get the current buffer created more than the specified time.

time_before in get_curr_sync_buff(2 * HZ) will get the buffer created newly.
if curr_sb has been created more than 2*HZ then it will still sit in master.
so use time_after instead.

Signed-off-by: Tinggong Wang <wangtinggong@xxxxxxxxx>
---
net/netfilter/ipvs/ip_vs_sync.c |    3 +--
1 files changed, 1 insertions(+), 2 deletions(-)

diff --git a/net/netfilter/ipvs/ip_vs_sync.c b/net/netfilter/ipvs/ip_vs_sync.c
index 2b6b0cb..555b0dd 100644
--- a/net/netfilter/ipvs/ip_vs_sync.c
+++ b/net/netfilter/ipvs/ip_vs_sync.c
@@ -221,8 +221,7 @@ get_curr_sync_buff(unsigned long time)
        struct ip_vs_sync_buff *sb;

        spin_lock_bh(&curr_sb_lock);
-       if (curr_sb && (time == 0 ||
-                       time_before(jiffies - curr_sb->firstuse, time))) {
+       if (curr_sb && time_after(jiffies - curr_sb->firstuse, time)) {

        This breaks the time=0 case when jiffies matches firstuse.

        May be the fix should be as follows?:

        if (curr_sb && time_after_eq(jiffies - curr_sb->firstuse, time)) {

        i.e. passed >= limit (2 or 0).

                sb = curr_sb;
                curr_sb = NULL;
        } else
--
1.7.2.3

Regards

--
Julian Anastasov <ja@xxxxxx>
--
To unsubscribe from this list: send the line "unsubscribe lvs-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html

<Prev in Thread] Current Thread [Next in Thread>