Also move the EXPORT_SYMBOL(tcp_statistics) to net/ipv4/tcp.c, so that
finally the tcp.h header can be removed from net/netsyms.c, ctags doesn't
catches the variables defined with DEFINE_SNMP_STAT :-\

wants 'phases'

Cset exclude: geert@linux-m68k.org|ChangeSet|20030912135245|42657

into home.osdl.org:/home/torvalds/v2.5/linux

into redhat.com:/spare/repo/misc-2.5

into home.osdl.org:/home/torvalds/v2.5/linux

The patch below switches MII to be select'ed instead of including it in
the Makefile.

Note that this patch requires a recent Linus' tree with the select CRC32
patch included.

diffstat output:

 drivers/net/Kconfig          |   11 +++++++++++
 drivers/net/Makefile         |   28 +++++++++++-----------------
 drivers/net/pcmcia/Kconfig   |    1 +
 drivers/net/tulip/Kconfig    |    1 +
 drivers/usb/net/Kconfig      |    2 ++
 drivers/usb/net/Makefile.mii |    6 ------
 6 files changed, 26 insertions(+), 23 deletions(-)


Tangential to the patch I observed a small problem (not fixed in the patch):
MII depends on NET_ETHERNET, but USB_PEGASUS and USB_USBNET depend
only on NET.

cu
Adrian

Similar to the patch Brian Gerst sent for CRC32, this patch changes
ZLIB_{IN,DE}FLATE to be select'ed.

The dependencies are much better maintainable that way.

Existing code for drivers/net/wan/z8530 is riddled with self-deadlocks
and irq flag confusion.  For example:
	z8530_init -> do_z8530_init -> write_zsreg
self deadlocks on the channel lock.

Several places acquire both the channel and dma lock and then
reuse the same irq flags variable - ouch.

This code at least, correctly probes (for no device case) on SMP.
Other paths verified by inspection.

The sealevel driver called sppp_attach before checking that board existed
and never called detach in the error path.

My change is to call sppp_attach from the netdev->init hook which happens
later in the process, and call detach from the uninit hook.

Also, changed the structure element 'netdev' to 'pppdev' to avoid confusion.

Here is the fix against 2.6.0-test5 latest