02 Nov, 2010
13 commits
-
Use standard network statistics variables and routines.
Transmit counters are per queue, and skb mapping is already in
skb and does not need to be recomputed. Move SearchVcId to only
place it is used.Signed-off-by: Stephen Hemminger
-
Since classification and queue control are separate, allow
classification even if device is down now; this avoids races
on startup/shutdown.Signed-off-by: Stephen Hemminger
-
The driver allocates 18 queues (0..16) are for traffic, and 17
is a bit bucket.Signed-off-by: Stephen Hemminger
-
Use namespace tool from kernel scripts to identify dead code and
functions that should be static.Signed-off-by: Stephen Hemminger
-
Signed-off-by: Stephen Hemminger
-
Eliminate global variable in transmit path
The Leader can be on the stack, and get rid of unnecessary timeval.Signed-off-by: Stephen Hemminger
-
Even though wimax isn't really a broadcast medium, pretend it is.
Signed-off-by: Stephen Hemminger
-
Add more debug messages and make them similar to other drivers
Signed-off-by: Stephen Hemminger
-
get_seconds is lower overhead and fine if all driver wants to do
is keep track of seconds.Signed-off-by: Stephen Hemminger
-
Get rid of boot messages and put in correct place.
Signed-off-by: Stephen Hemminger
-
Provide standard interface to control verbosity of debug messages
Signed-off-by: Stephen Hemminger
-
Statistic point is now u32 (like it has to be).
Signed-off-by: Stephen Hemminger
-
Remove dead fields, change fields that only have true/false to boolean;
and rearrange to save space.Signed-off-by: Stephen Hemminger
01 Nov, 2010
23 commits
-
First step to supporting multiple devices, create device class
when module is initialized.Signed-off-by: Stephen Hemminger
-
Change indentation etc, to conform to acceptable kernel style
Signed-off-by: Stephen Hemminger
-
Signed-off-by: Stephen Hemminger
-
Not worth bothering to change printf format of messages which
are basically noise.Signed-off-by: Stephen Hemminger
-
Signed-off-by: Stephen Hemminger
-
Sparse caught several places where ioctl interface was incorrectly using user
memory. Fix all the ioctl cases for casting and __user annotation.Signed-off-by: Stephen Hemminger
-
Signed-off-by: Stephen Hemminger
-
This preserves the semantics of the original driver (unclassified packets
are dropped), but does it in a clean way; and fixes crash when packet
is sent to offline device.Signed-off-by: Stephen Hemminger
-
Remove commented out with '#if 0'
Signed-off-by: Stephen Hemminger
-
Current kernels have multi-queue support which can be used by
this device. This has the advantage that a single type of traffic
will not block other types.Signed-off-by: Stephen Hemminger
-
Many routines were tagged with inline_ but GCC does a better
job of deciding this.Signed-off-by: Stephen Hemminger
-
The function unregister_netdevice only called unregister_netdev.
Signed-off-by: Stephen Hemminger
-
Linux support NOARP flag, so the whole Arp spoofing routines
are not needed.Signed-off-by: Stephen Hemminger
-
Signed-off-by: Stephen Hemminger
-
Allocate Adapter structure as part of network device.
Signed-off-by: Stephen Hemminber
-
Start with carrier off.
Don't track up/down status in driver private flag.Signed-off-by: Stephen Hemminger
-
Reduce transmit queue length to avoid excess buffering
Signed-off-by: Stephen Hemminger
-
Signed-off-by: Stephen Hemminger
-
This adds basic ethtool support to get driver info and settings
Signed-off-by: Stephen Hemminger
-
Kernel complains loudly if thread does long uninterruptible sleep.
Also, dont wake up every 10ms even if no data present (wastes power).Signed-off-by: Stephen Hemminger
-
This is the convention used by Intel Wimax
Signed-off-by: Stephen Hemminger
-
Use native kernel functions for kmalloc/kfree directly
Signed-off-by: Stephen Hemminger
-
No longer need special hex dump routine
Signed-off-by: Stephen Hemminger
30 Oct, 2010
4 commits
-
Change how network device is setup:
* set pointer to device object so sysfs has eth0/device symlink
* set network device type
* eliminate all the compatiablity with older kernels.Signed-off-by: Stephen Hemminger
-
The only part of this structure still used was the network
device stats, and in recent kernel these are available in
network device itself.Signed-off-by: Stephen Hemminger
-
Get rid of empty header file and unused declarations
Signed-off-by: Stephen Hemminger
-
Just use presence of pointer
Signed-off-by: Stephen Hemminger