Given many versions of ethtool's reluctance to do anything other than
byte accesses to the EEPROM interface, it is easier to update the driver
to support byte accesses so that all the ethtool versions that have been
observed in Debian can write the EEPROM.
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Reviewed-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
FTGMAC100 Ethernet Media Access Controller supports 10/100/1000 Mbps
and MII/GMII. This driver has been working on some ARM/NDS32 SoC's
including Faraday A369 and Andes AG102.
Signed-off-by: Po-Yu Chuang <ratbert@faraday-tech.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
arch/m68k/emu/nfeth.c: In function ‘nfeth_init’:
arch/m68k/emu/nfeth.c:243: error: implicit declaration of function ‘request_irq’
arch/m68k/emu/nfeth.c:243: error: ‘IRQF_SHARED’ undeclared (first use in this function)
arch/m68k/emu/nfeth.c:243: error: (Each undeclared identifier is reported only once
arch/m68k/emu/nfeth.c:243: error: for each function it appears in.)
arch/m68k/emu/nfeth.c: In function ‘nfeth_cleanup’:
arch/m68k/emu/nfeth.c:266: error: implicit declaration of function ‘free_irq’
drivers/net/apne.c: In function ‘apne_probe’:
drivers/net/apne.c:189: error: implicit declaration of function ‘free_irq’
drivers/net/apne.c: In function ‘apne_probe1’:
drivers/net/apne.c:317: error: implicit declaration of function ‘request_irq’
drivers/net/apne.c:317: error: ‘IRQF_SHARED’ undeclared (first use in this function)
drivers/net/apne.c:317: error: (Each undeclared identifier is reported only once
drivers/net/apne.c:317: error: for each function it appears in.)
Introduced by commit a6b7a40786 ("net: remove interrupt.h inclusion from
netdevice.h").
Include <linux/interrupt.h> in the individual drivers to fix the build.
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The message size allocated for rtnl ifinfo dumps was limited to
a single page. This is not enough for additional interface info
available with devices that support SR-IOV and caused a bug in
which VF info would not be displayed if more than approximately
40 VFs were created per interface.
Implement a new function pointer for the rtnl_register service that will
calculate the amount of data required for the ifinfo dump and allocate
enough data to satisfy the request.
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch changes the way versioning is done for igb in the kernel by
removing the number after the "k." It has been determined that just the
"k" is sufficient to identify a kernel version and the following number
was used in an inconsistent manner.
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com>
Tested-by: Jeff Pieper <jeffrey.e.pieper@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Update the version number to match version conventions. Bump the major
version to indicate that new hardware support (i350) has been added.
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Commit 5d03078a68 added a redundant 'select
CRC32'; remove it.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The Host Wakeup Active bit in the PHY Port General Configuration register
(page 769 register 17) must be cleared after every PHY reset to prevent an
unexpected wake signal from the PHY. Originally, this was accomplished by
simply reading the PHY Wakeup Control register on page 800 which clears the
Host Wakeup Active bit as a side-effect. Unfortunately, a hardware bug on
the 82577 and 82578 PHY can cause unexpected behavior when registers on
page 800 are accessed while in gigabit mode.
This patch changes the remaining instances when the Host Wakeup Active bit
needs to be cleared while possibly in gigabit mode by accessing the Port
General Configuration register directly instead of accessing any register
on page 800.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Jeff Pieper <jeffrey.e.pieper@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Doing a PHY page select can take a long time, relatively speaking. This
can cause a significant delay when updating a number of PHY registers on
the same page by unnecessarily setting the page for each PHY access. For
example when going to Sx, all the PHY wakeup registers (WUC, RAR[], MTA[],
SHRAR[], IP4AT[], IP6AT[], etc.) on 82577/8/9 need to be updated which
takes a long time which can cause issues when suspending.
This patch introduces new PHY ops function pointers to allow callers to
set the page directly and do any number of PHY accesses on that page.
This feature is currently only implemented for 82577, 82578 and 82579
PHYs for both the normally addressed registers as well as the special-
case addressing of the PHY wakeup registers on page 800. For the latter
registers, the existing function for accessing the wakeup registers has
been divided up into three- 1) enable access to the wakeup register page,
2) perform the register access and 3) disable access to the wakeup register
page. The two functions that enable/disable access to the wakeup register
page are necessarily available to the caller so that the caller can restore
the value of the Port Control (a.k.a. Wakeup Enable) register after the
wakeup register accesses are done.
All instances of writing to multiple PHY registers on the same page are
updated to use this new method and to acquire any PHY locking mechanism
before setting the page and performing the register accesses, and release
the locking mechanism afterward.
Some affiliated magic number cleanup is done as well.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Jeff Pieper <jeffrey.e.pieper@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Start the Tx queue when the interface is brought up in e1000e_up() but do
not schedule the queue until link is up as detected in the watchdog task
which sets netif_carrier_on.
Also flush the descriptors and clean the Tx and Rx rings before resetting
the hardware when bringing the interface down otherwise there is a small
window where the watchdog task can be triggered with netif_carrier_off
and the Tx ring not yet empty which causes an additional and unnecessary
reset.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Since EXTCNF_CTRL.SWFLAG (used in the ownership arbitration of shared
resources, e.g. the PHY shared between the s/w, f/w, and h/w clients)
can be cleared by any of those clients, log a debug message when
software attempts to clear it and it is already cleared unexpectedly.
And since the swflag is cleared by a hardware reset, the driver does
not need to do that, but the mutex acquired when the bit is set must
still be cleared.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
When repeatedly cycling Sx->S0 states with the network cable unplugged,
the 82579 PHY may not initialize as expected and may require a full power
cycle to recover functionality to the device. Workaround this by testing
access of the PHY registers after resuming; if that returns unexpected
results toggle the LANPHYPC signal to power cycle the PHY.
This is implemented in the new function e1000_resume_workarounds_pchlan()
which calls another new function, e1000_toggle_lanphypc_value_ich8lan(),
which has been created to reduce code duplication (same functionality
required by a previous workaround). Also, e1000e_disable_gig_wol_ich8lan
is now e1000_suspend_workarounds_ich8lan to better reflect what it does.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The ESB2 LAN includes a debug feature that enables far-end loopback (FELB)
of the SerDes/Kumeran interface. This feature is activated when receiving
a sequence of symbols that includes a reserved codeword. On a perfect
link, FELB would never be activated. In the presence of bit errors, there
is a very small, but non-zero, probability of FELB being activated.
If the FELB is activated, the SerDes link becomes non-functional and must
be reset. It could also corrupt the switching tables in the switch since
the ESB2 is transmitting packets with a different source MAC address.
This patch disables the FELB feature.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Now all received packets are handled by bond_handle_frame,
and arp_mon_pt isn't used any more.
Reviewed-by: WANG Cong <xiyou.wangcong@gmail.com>
Signed-off-by: Weiping Pan <panweiping3@gmail.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Now we use agg_select_timer and ad_work.
Reviewed-by: WANG Cong <xiyou.wangcong@gmail.com>
Signed-off-by: Weiping Pan <panweiping3@gmail.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
bond_params->ad_select and ad_bond_info->agg_select_mode have the same
meaning, they are duplicate and need extra synchronization.
__get_agg_selection_mode() get ad_select from bond_params directly.
Signed-off-by: Weiping Pan <panweiping3@gmail.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
These is also a bug, that if you modify lacp_rate via sysfs,
and add new slaves in bonding, new slaves won't use the latest lacp_rate,
since ad_bond_info->lacp_fast is initialized only once,
in bond_3ad_initialize().
Since both struct bond_params and ad_bond_info have lacp_fast,
they are duplicate and need extra synchronization.
bond_3ad_bind_slave() can use bond_params->lacp_fast to initialize port.
So we can just remove lacp_fast from struct ad_bond_info.
Signed-off-by: Weiping Pan <panweiping3@gmail.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There is bug that when you modify lacp_rate via sysfs,
802.3ad won't use the new value of lacp_rate to transmit packets.
This is because port->actor_oper_port_state isn't changed.
Signed-off-by: Weiping Pan <panweiping3@gmail.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Christian Benvenuti <benve@cisco.com>
Signed-off-by: Danny Guo <dannguo@cisco.com>
Signed-off-by: Vasanthy Kolluri <vkolluri@cisco.com>
Signed-off-by: Roopa Prabhu <roprabhu@cisco.com>
Signed-off-by: David Wang <dwang2@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Instead of deriving the index of a transmit/receive interrupt resource
from the transmit/receive queue index, always save and retrieve it
using an additional variable.
Signed-off-by: Christian Benvenuti <benve@cisco.com>
Signed-off-by: Danny Guo <dannguo@cisco.com>
Signed-off-by: Vasanthy Kolluri <vkolluri@cisco.com>
Signed-off-by: Roopa Prabhu <roprabhu@cisco.com>
Signed-off-by: David Wang <dwang2@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Christian Benvenuti <benve@cisco.com>
Signed-off-by: Danny Guo <dannguo@cisco.com>
Signed-off-by: Vasanthy Kolluri <vkolluri@cisco.com>
Signed-off-by: Roopa Prabhu <roprabhu@cisco.com>
Signed-off-by: David Wang <dwang2@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
enic driver currently passes 802.1p bits to the upper layers for packets
tagged with non-zero vlan ids only. This patch extends such behaviour to
zero vlan tagged packets also.
The patch is dependant on the following kernel patches:
1) vlan_dev: VLAN 0 should be treated as "no vlan tag" (802.1p packet)
- net-next-2.6 git commit: ad1afb0039
- Available 2.6.36 and later
2) vlan: Centralize handling of hardware acceleration.
- net-next-2.6 git commit: 3701e51382
- Available 2.6.37 and later
Signed-off-by: Christian Benvenuti <benve@cisco.com>
Signed-off-by: Danny Guo <dannguo@cisco.com>
Signed-off-by: Vasanthy Kolluri <vkolluri@cisco.com>
Signed-off-by: Roopa Prabhu <roprabhu@cisco.com>
Signed-off-by: David Wang <dwang2@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The definition NO_FLAGS was introduced to make the code more
readable and shall be used to initialize flag fields.
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
Signed-off-by: Sven Eckelmann <sven@narfation.org>
CodingStyle "Chapter 12: Macros, Enums and RTL" recommends to use enums
for several related constants. Internal states can be used without
defining the actual value, but all values which are visible to the
outside must be defined as before. Normal values are assigned as usual
and flags are defined by shifts of a bit.
Signed-off-by: Sven Eckelmann <sven@narfation.org>
CodingStyle "Chapter 12: Macros, Enums and RTL" highly recommends to use
functions instead of macros were possible. This ensures type safety and
prevents shadowing of other variables.
Signed-off-by: Sven Eckelmann <sven@narfation.org>
strict_strtoul as used in parse_gw_bandwidth is defined for unsigned
long and strict_strtol should be used instead for long.
Signed-off-by: Sven Eckelmann <sven@narfation.org>
gw_node_delete is defined with "void" as return type, but still tries to
return a value. The called function gw_node_delete is also return as
void and thus doesn't provide a value for us.
Signed-off-by: Sven Eckelmann <sven@narfation.org>
When a link is saturated (re)broadcasts of OGMs are delayed. Under heavy
load this delay may exceed the orig interval which leads to OGMs being
dropped (the code would only accept an OGM rebroadcast if it arrived
before the next OGM was broadcasted). With this patch batman-adv will
also accept delayed OGMs in order to avoid a bogus influence on the
routing metric.
Signed-off-by: Daniele Furlan <daniele.furlan@gmail.com>
Signed-off-by: Sven Eckelmann <sven@narfation.org>
Log:
drivers/net/xilinx_emaclite.c: In function 'xemaclite_open':
drivers/net/xilinx_emaclite.c:961: error: implicit declaration of function 'request_irq'
drivers/net/xilinx_emaclite.c: In function 'xemaclite_close':
drivers/net/xilinx_emaclite.c:995: error: implicit declaration of function 'free_irq'
make[2]: *** [drivers/net/xilinx_emaclite.o] Error 1
make[1]: *** [drivers/net] Error 2
make: *** [drivers] Error 2
Signed-off-by: Michal Simek <monstr@monstr.eu>
Signed-off-by: David S. Miller <davem@davemloft.net>
Perf shows a relatively high rate (about 8%) race in
spin_lock_irqsave() when doing netperf between external host and
guest. It's mainly becuase the lock contention between the
tun_do_read() and tun_xmit_skb(), so this patch do not put self into
waitqueue to reduce this kind of race. After this patch, it drops to
4%.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Amos Kong <akong@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Current standard practice is to not mark most functions as inline
and let compiler decide instead.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The tun driver allocates skb's to hold data from user and then passes
the data into the network stack as received data. Most network devices
allocate the receive skb with routines like dev_alloc_skb() that reserves
additional space for use by network protocol stack but tun does not.
Because of the lack of padding, when the packet is passed through bridge
netfilter a new skb has to be allocated.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Based on earlier patch from Neil Horman <nhorman@tuxdriver.com>
If iSCSI is not supported on a bnx2 device, bnx2_cnic_probe() will
return NULL and the cnic device will not be visible to bnx2i. This
will prevent bnx2i from registering and then unregistering during
cnic_start() and cause the warning message:
bnx2 0003:01:00.1: eth1: Failed waiting for ULP up call to complete
Signed-off-by: Michael Chan <mchan@broadcom.com>
Cc: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
During NETDEV_UP, we use symbol_get() to get the net driver's cnic
probe function. This sometimes doesn't work if NETDEV_UP happens
right after NETDEV_REGISTER and the net driver is still running module
init code. As a result, the cnic device may not be discovered. We
fix this by probing on all NETDEV events if the device's netif_running
state is up.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This reduces the likelihood of port re-use when re-loading the driver.
Signed-off-by: Eddie Wai <eddie.wai@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
During iSCSI connection terminations, if the target is also terminating
at about the same time, the firmware may not complete the driver's
request to close or reset the connection. This is fixed by handling
other events (instead of the expected completion event) as an indication
that the driver's request has been rejected.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We need to keep looping until cnic_get_kcqes() returns 0. cnic_get_kcqes()
returns a maximum of 64 entries. If there are more entries in the queue
and we don't loop back, the remaining entries may not be serviced for a
long time.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Profiles show false sharing in addr_compare() because refcnt/dtime
changes dirty the first inet_peer cache line, where are lying the keys
used at lookup time. If many cpus are calling inet_getpeer() and
inet_putpeer(), or need frag ids, addr_compare() is in 2nd position in
"perf top".
Before patch, my udpflood bench (16 threads) on my 2x4x2 machine :
5784.00 9.7% csum_partial_copy_generic [kernel]
3356.00 5.6% addr_compare [kernel]
2638.00 4.4% fib_table_lookup [kernel]
2625.00 4.4% ip_fragment [kernel]
1934.00 3.2% neigh_lookup [kernel]
1617.00 2.7% udp_sendmsg [kernel]
1608.00 2.7% __ip_route_output_key [kernel]
1480.00 2.5% __ip_append_data [kernel]
1396.00 2.3% kfree [kernel]
1195.00 2.0% kmem_cache_free [kernel]
1157.00 1.9% inet_getpeer [kernel]
1121.00 1.9% neigh_resolve_output [kernel]
1012.00 1.7% dev_queue_xmit [kernel]
# time ./udpflood.sh
real 0m44.511s
user 0m20.020s
sys 11m22.780s
# time ./udpflood.sh
real 0m44.099s
user 0m20.140s
sys 11m15.870s
After patch, no more addr_compare() in profiles :
4171.00 10.7% csum_partial_copy_generic [kernel]
1787.00 4.6% fib_table_lookup [kernel]
1756.00 4.5% ip_fragment [kernel]
1234.00 3.2% udp_sendmsg [kernel]
1191.00 3.0% neigh_lookup [kernel]
1118.00 2.9% __ip_append_data [kernel]
1022.00 2.6% kfree [kernel]
993.00 2.5% __ip_route_output_key [kernel]
841.00 2.2% neigh_resolve_output [kernel]
816.00 2.1% kmem_cache_free [kernel]
658.00 1.7% ia32_sysenter_target [kernel]
632.00 1.6% kmem_cache_alloc_node [kernel]
# time ./udpflood.sh
real 0m41.587s
user 0m19.190s
sys 10m36.370s
# time ./udpflood.sh
real 0m41.486s
user 0m19.290s
sys 10m33.650s
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Driver was already keeping 64 bit counters, just not using the new interface.
Ps: IMHO drivers should not be duplicating network device
stats into ethtool stats. It is useless duplication.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The device driver already uses 64 bit statistics, it just
doesn't use the 64 bit interface.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Change to 64 bit statistics interface, driver was already maintaining 64 bit
value.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Acked-by: Amit Kumar Salecha <amit.salecha@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Not much change, device was already keeping per cpu statistics.
Use recent 64 statistics interface.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Convert vmxnet3 driver to 64 bit statistics interface.
This driver was already counting packet per queue in a 64 bit value so not
a huge change. Eliminate unused old net_device_stats structure.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: Scott J. Goldman <scottjg@vmware.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Andi Kleen and Tim Chen reported huge contention on inetpeer
unused_peers.lock, on memcached workload on a 40 core machine, with
disabled route cache.
It appears we constantly flip peers refcnt between 0 and 1 values, and
we must insert/remove peers from unused_peers.list, holding a contended
spinlock.
Remove this list completely and perform a garbage collection on-the-fly,
at lookup time, using the expired nodes we met during the tree
traversal.
This removes a lot of code, makes locking more standard, and obsoletes
two sysctls (inet_peer_gc_mintime and inet_peer_gc_maxtime). This also
removes two pointers in inet_peer structure.
There is still a false sharing effect because refcnt is in first cache
line of object [were the links and keys used by lookups are located], we
might move it at the end of inet_peer structure to let this first cache
line mostly read by cpus.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
CC: Andi Kleen <andi@firstfloor.org>
CC: Tim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch lowers the default initRTO from 3secs to 1sec per
RFC2988bis. It falls back to 3secs if the SYN or SYN-ACK packet
has been retransmitted, AND the TCP timestamp option is not on.
It also adds support to take RTT sample during 3WHS on the passive
open side, just like its active open counterpart, and uses it, if
valid, to seed the initRTO for the data transmission phase.
The patch also resets ssthresh to its initial default at the
beginning of the data transmission phase, and reduces cwnd to 1 if
there has been MORE THAN ONE retransmission during 3WHS per RFC5681.
Signed-off-by: H.K. Jerry Chu <hkchu@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ben Dooks
Michio Honda
Po-Yu Chuang
Geert Uytterhoeven
Greg Rose
Carolyn Wyborny
Williams, Mitch A
Bruce Allan
Peter Pan(潘卫平)
Vasanthy Kolluri
David S. Miller
Marek Lindner
Sven Eckelmann
Daniele Furlan
Michal Simek
Amos Kong
stephen hemminger
Michael Chan
Eddie Wai
Eric Dumazet
Jerry Chu