fedora 14 kernel performance with ip forwarding workload
by Jesse Brandeburg
The other day I was running the stock fedora kernel on my ip
forwarding setup, to see what the performance was, and the performance
wasn't very good.
system is S5520HC dual socket 2.93GHz Xeon 5570 (Nehalem) with 3 quad
port 82580 adapters (12 ports). Traffic is bidirectional 64 byte
packets being forwarded and received on each port, basically port to
port routing. I am only using 12 flows currently.
The driver is igb, and I am using an affinity script that lines up
each pair of ports that are forwarding traffic into optimal
configurations for cache locality. I am also disabling
remote_node_defrag_ratio to stop cross node traffic.
With the fedora default kernel from F14 it appears that
CONFIG_NETFILTER=y means that I cannot unload all of netfilter even if
I stop iptables service.
perf showed netfilter being prominent, and removing it gives me much
higher throughput. Is there a reason CONFIG_NETFILTER=y ? Isn't it a
good thing to be able to disable netfilter if you want to?
Jesse
7 years, 11 months
[PATCH] Drop the old Adaptec AIC7xxx driver
by Paul Bolle
Ever since v2.4.3 this driver is advertised as not being under active
development. It's main advantage to most users will be the warnings it
generates at build time:
drivers/scsi/aic7xxx_old.c: In function ‘aic7xxx_register’:
drivers/scsi/aic7xxx_old.c:7901:5: warning: case value ‘257’ not in enumerated type ‘ahc_chip’ [-Wswitch]
drivers/scsi/aic7xxx_old.c:7898:5: warning: case value ‘513’ not in enumerated type ‘ahc_chip’ [-Wswitch]
drivers/scsi/aic7xxx_old.c: In function ‘aic7xxx_load_seeprom’:
drivers/scsi/aic7xxx_old.c:8517:5: warning: case value ‘257’ not in enumerated type ‘ahc_chip’ [-Wswitch]
drivers/scsi/aic7xxx_old.c:8510:5: warning: case value ‘513’ not in enumerated type ‘ahc_chip’ [-Wswitch]
Let's drop this old driver.
Signed-off-by: Paul Bolle <pebolle(a)tiscali.nl>
---
0) I'm trying to get my kernel builds (which use the Fedora kernel
config) warning free. Upstream declined the brilliant fix to silence
these warnings (see https://lkml.org/lkml/2012/9/21/104 ).
1) Plan B is to simply drop this old driver in the Fedora kernel
altogether. That should silence these warnings too, shouldn't it? But is
dropping this old driver acceptable?
2) This patch is entirely untested.
config-generic | 1 -
1 file changed, 1 deletion(-)
diff --git a/config-generic b/config-generic
index 4a8015f..5cc8f8d 100644
--- a/config-generic
+++ b/config-generic
@@ -341,7 +341,6 @@ CONFIG_SCSI_3W_9XXX=m
CONFIG_SCSI_ACARD=m
CONFIG_SCSI_AACRAID=m
CONFIG_SCSI_AIC7XXX=m
-CONFIG_SCSI_AIC7XXX_OLD=m
CONFIG_AIC7XXX_CMDS_PER_DEVICE=4
CONFIG_AIC7XXX_RESET_DELAY_MS=15000
# CONFIG_AIC7XXX_BUILD_FIRMWARE is not set
--
1.7.11.7
10 years, 3 months
Fedora Kernel Meeting 2012-12-14 cancelled
by Josh Boyer
Due to a rather limited amount of stuff to discuss, we are going to
cancel today's Fedora Kernel IRC meeting. If there is anything anyone
wishes to discuss, please feel free to stop by #fedora-kernel or send
us an email on the kernel list!
josh
10 years, 5 months
[PATCH F18] team: backport fixes from linus's tree
by Jiri Pirko
Backport fixes from linus's tree.
upstream commits backported:
commit 403f43c937d24832b18524f65415c0bbba6b5064
team: bcast: convert return value of team_dev_queue_xmit() to bool correctly
commit 3ed7147189d2fbe8ac6da95db2fd9d6d52f53ce9
team: fix hw_features setup
Signed-off-by: Jiri Pirko <jpirko(a)redhat.com>
---
drivers/net/team/team.c | 4 +++-
drivers/net/team/team_mode_broadcast.c | 6 +++---
2 files changed, 6 insertions(+), 4 deletions(-)
diff --git a/drivers/net/team/team.c b/drivers/net/team/team.c
index 5d8e1cb..f504773 100644
--- a/drivers/net/team/team.c
+++ b/drivers/net/team/team.c
@@ -1792,10 +1792,12 @@ static void team_setup(struct net_device *dev)
dev->features |= NETIF_F_LLTX;
dev->features |= NETIF_F_GRO;
- dev->hw_features = NETIF_F_HW_VLAN_TX |
+ dev->hw_features = TEAM_VLAN_FEATURES |
+ NETIF_F_HW_VLAN_TX |
NETIF_F_HW_VLAN_RX |
NETIF_F_HW_VLAN_FILTER;
+ dev->hw_features &= ~(NETIF_F_ALL_CSUM & ~NETIF_F_HW_CSUM);
dev->features |= dev->hw_features;
}
diff --git a/drivers/net/team/team_mode_broadcast.c b/drivers/net/team/team_mode_broadcast.c
index 9db0171..c5db428 100644
--- a/drivers/net/team/team_mode_broadcast.c
+++ b/drivers/net/team/team_mode_broadcast.c
@@ -29,8 +29,8 @@ static bool bc_transmit(struct team *team, struct sk_buff *skb)
if (last) {
skb2 = skb_clone(skb, GFP_ATOMIC);
if (skb2) {
- ret = team_dev_queue_xmit(team, last,
- skb2);
+ ret = !team_dev_queue_xmit(team, last,
+ skb2);
if (!sum_ret)
sum_ret = ret;
}
@@ -39,7 +39,7 @@ static bool bc_transmit(struct team *team, struct sk_buff *skb)
}
}
if (last) {
- ret = team_dev_queue_xmit(team, last, skb);
+ ret = !team_dev_queue_xmit(team, last, skb);
if (!sum_ret)
sum_ret = ret;
}
--
1.7.11.7
10 years, 5 months
