On 03/08/2018 03:30 AM, Uri Lublin wrote:
On 03/01/2018 11:08 AM, zhenwei.pi wrote:
In some bad cases, for example, host OS crashes without
sending any FIN to usbredirserver, and usbredirserver
will keep idle connection for a long time.
We can also set the kernel arguments, it means that other
Signed-off-by: zhenwei.pi <zhenwei...@youruncloud.com>
---
usbredirserver/usbredirserver.c | 31 ++-
1 file changed, 30 insertions(+), 1 deletion(-)
diff --git a/usbredirserver/usbredirserver.c b/usbredirserver/usbredirserver.c
index 5575181..3a7620a 100644
---
Detecting disconnection costs time is :
30s + 10s * 3 = 60s
If setting TCP keepalive fails with errno ENOTSUP, ignore
the specific error.
Signed-off-by: zhenwei.pi <zhenwei...@youruncloud.com>
---
usbredirserver/usbredirserver.c | 41 -
1 file change
connection costs time is :
idle time + 10s * 3
If setting TCP keepalive fails with errno ENOTSUP, ignore
the specific error.
Signed-off-by: zhenwei.pi <zhenwei...@youruncloud.com>
Tested-by: Uri Lublin <u...@redhat.com>
Acked-by: Frediano Ziglio <fzig...@redhat
Signed-off-by: zhenwei.pi <zhenwei...@youruncloud.com>
I think I got completely wrong the problem.
So this patch is trying to use tcp keepalives to detect disconnection
making possible new connections. The problem is that the server
stops accepting connections when it handle one client and ret