Age | Commit message (Collapse) | Author | Files | Lines |
|
get_tick() is not a good function name, namespace wise, so use
kbench_get_tick() instead.
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Signed-off-by: Vincent Bernat <vincent@bernat.im>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
The former one is not available since Linux 4.6 and boot_cpu_has() is
available since a long time.
Signed-off-by: Vincent Bernat <vincent@bernat.im>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
This patch:
1. Add IPv6 support
2. Print timing for every 65536 fib insert operations to observe
the gc effect (mostly for IPv6 fib).
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Flow keys can be specified on the kernel command line.
A warmup of "warmup_count" lookups are performed, then
a single cycle counted lookup is performed with the
cycle count reported in the kernel logs.
X86, Powerpc, and Sparc64 are currently supported.
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
This avoids us seeing overhead from netlink reply generation.
As a backup we keep the socket filter there in case the kernel
does not support this feature.
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Since mnl_socket_get_fd() exists, we can use that.
Thanks to Pablo.
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Install a socket filter to reduce the pure netlink overhead.
Unfortunately the libmnl library does not provide a way to
set socket options that are of level other than SOL_NETLINK.
So we hack it by knowing some things about libmnl internals.
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
|