In article <200109250803.f8P83sF02609@euphoria.confusion.net>,
Michael K. Sanders <msanders@confusion.net> wrote:
>I have a couple SMC Power155 PCI ATM cards, which are apparently just
>OEM versions of the Efficient card.  I played with them a few years ago
>and I believe they probed just fine, though I never actually used them.
>
>Now, however, something isn't working.
>
>en0 at pci0 dev 11 function 0: Efficent Networks 155P-MF1 ATM (ASIC) (rev. 0x00)
>en0: interrupting at irq 10
>en0: unexpected timeout in tx DMA test
...
>en0: unexpected timeout in tx DMA test
>en0: WARNING: WMAYBE DMA test failed 56 time(s)
>en0: ATM midway v0, board IDs 6.0, Utopia (pipelined), 512KB on-board RAM
>en0: maximum DMA burst length = 64 bytes (must align)
>en0: 7 32KB receive buffers, 8 32KB transmit buffers allocated
>en0: End Station Identifier (mac address) 00:00:c0:44:61:e0

>Any ideas?



when the driver attaches to an "en" it does a whole bunch of
test DMA operations to see what your system supports.   some
motherboards fail in such a way that it causes an 
"unexpected timeout in tx DMA test" error on some tests.

once the driver finishes the tests it imposes a set of DMA
restrictions so that only known-good DMA operations are used.
in your case it concluded:
	en0: maximum DMA burst length = 64 bytes (must align)

64 is the max burst size, but your system only allows it if
the starting address is on a 64-byte aligned boundary (that's
the "must align" restriction).    

the driver may have to do extra DMA operations to meet this 
restriction, but after autoconfig it should (in theory) work
normally.


chuck