It seems that skb_padto security fixes in 2.4 and 2.5 trying
to fix "CAN-2003-0001:Multiple ethernet NID device drivers
do not pad frames with null bytes", do not put the skb_padto
blocks in proper places in the  3c527, eth16i, fmv18x, seeq8005,
yellowfin device drivers.   

In case a driver calls skb_padto(), it is possible
that the space available in the original skb buffer tailroom is less
than the space to pad. In this case, in short, the skb_padto()
will create a new skb buffer, copy data from the original
skb buffer to a new skb buffer, free the original buffer,
and finally return the new buffer.

If this happens to the aforementioned device drivers, they come to
point to wrong data. And, for 3c527 and yellowfin, the drivers can
unexpectedly double free the original skb buffers since they still
point to the original skb buffers. The attached patch against
2.4.23pre1 fixes these issues.

The only thing that differentiated most of the entries in the
board_info[] table and the board_t type was the vendor branding
string for the board.  This table is a pain to maintain, so we
prefer to simply use "RTL8129" or "RTL8139".

a.k.a. Sega Broadband Adapter.

Updated drivers;  3c574_cs, axnet_cs, pcnet_cs, xirc2ps_cs

Thanks to Komuro for pointing this out.

* only set PCIDAC (64-bit PCI) bit in hardware if
  sizeof(dma_addr_t) > 32.  Need a better test for whether
  64-bit mode is _really_ needed.
* cache chip command register in private struct.  this allows
  the setting of rx-vlan, rx-csum, and other features to be
  persistent across the entire lifetime of the net device.
* remove dead private struct members frag_skb, dropping_frag,
  and pci_using_dac.

* make sure rx_frags is still accounted
* query RxMissed register, and clear, upon each get-stats func call

* remove netif_queue_stopped test, netif_wake_queue already does that
* move vlan stuff to top of file
* remove __dev markers
* update todo list at top of file
* remove pci_set_dma_mask argument casts; ULL suffixes preferred.

NAPI is turned on unconditionally for the RX path.  The hardware
supports interrupt mitigation, so that should be investigated too.

RX fragment handling removed.  We simply ensure that we alloc
buffers large enough to hold incoming packets.  Any stray RX
frags that occur (shouldn't be any) will be dropped.

(previously it was ifdef'd)

Also, bump version to 1.0.

Missing pieces for DMA-API on the Rx side:
- SiS190_init_ring: the global area for the received data is mapped.
  This area is persistent during the whole driver's life.
  It only needs to be unmapped in SiS190_close() as no other exit/error
  path exists.
- SiS190_rx_interrupt: no map/unmap for received data buffer. A single
  sync operation is done. Btw, there is no need to store the same value
  in RxDescArray[cur_rx].buf_addr over and over again.
- Remove driver dependancy on CONFIG_BROKEN.

* call pci_set_dma_mask
* remove erroneous call to unregister_netdev in _init_board()

I realized that I've had this patch in my tree for a while, and forgot
to send it to you and lkml.  The patch below fixes bug number 923:
	http://bugme.osdl.org/show_bug.cgi?id=923
(basically keeps us from calling pci_find_device from interrupt
context.)

It's been tested by a few people with this device, and they say it works
just fine for them.  Please forward it on up the food chain.

instead of homebrew MII bitbanging.

Driver does not need to enforce 256 byte alignment for data returned
from pci_alloc_consistent().
- {rx/tx}_dma_aligned and {rx/td}_dma_raw are both replaced by {rx/tx}_dma;
- {rx/tx}_desc_raw is replaced by direct use of {Rx/Tx}DescArray;
- SiS190_open()
  + fixup for a lack of kmalloc() failure handling;
  + (return status) there is no need for both retval/rc: merge them;
  + anonymous printk() fixup: the name of the guilty device is printed;
- define {RX/TX}_DESC_TOTAL_SIZE because I am too lazy to read twice the
  same lengthy arithmetic expression.

Also, export PCI bus id via ETHTOOL_GDRVINFO.

Drivers updated: fmvj18x_cs, ibmtr_cs, nmclan_cs, pcnet_cs,
xirc2ps_cs.

Until RX path is cleaned up to use PCI DMA API and
not virt_to_bus.

Also, minor changes:
* mark ->hard_start_xmit ETH_ZLEN test as unlikely()
* use cpu_to_le32() and le32_to_cpu() in TX path
* fix two leak in error path, in ->hard_start_xmit
* don't test netif_queue_stopped() in TX completion path,
  netif_wake_queue() already does that.

Looks like this driver wasn't tested on SMP :)