directly, call the function pointer (*if_input)(ifp, m). The input routine
expects the packet header to be at the head of the packet, and will adjust
as necessary. Privatize the layer 2 input and output routines, allowing
*_ifattach() to set them up as appropriate.
TULIP_BUSMODE_BIGENDIAN does bswap packet buffers also, so we should use
TULIP_BUSMODE_DESC_BIGENDIAN on big-endian machines. (PR 7027)
XXX 21040 doesn't have this bit, but supporting only 21041+ is better than
nothing.
descriptor count gets miscounted on txprobes. The second (and more
important) is that transmit stalls should now be fixed. The problem
was not due to lack of ring resources but dmamaps. When Jason changed
the driver to use pre-allocated maps (instead of the dynamic ones I
used), and when there were no more maps, the driver just gave up instead
of calling tx_intr to free any transmitted but unreclaimed dma maps.
Since there was nothing being transmitted, no transmit interrupts
would fire to restore things (and OACTIVE prevented other transmits
from happenning). So it stayed starved until another interrupt cause
(like a received packet) "woke" it up.
-use of uninitialized variable in attach (mostly harmless)
-removal of dead code in attach()
-fix of interpretation of SIA blocks on >=21142 (serious, caused strange
things with 10Mbit)
it simple, and allocate one for each transmit and receive descriptor.
In addition to being simpler and faster, this fixes a serious memory leak
in the transmit path.
tested on any other platform other than i386. Use of bus_dma(9) can be
turned by defining TULIP_BUS_DMA_NOTX and TULIP_BUS_DMA_NORX. These allow
one to determine if the problem is in the transmit or receive path.
A problem reported by mycroft is also fixed.
pseudo-device rnd # /dev/random and in-kernel generator
in config files.
o Add declaration to all architectures.
o Clean up copyright message in rnd.c, rnd.h, and rndpool.c to include
that this code is derived in part from Ted Tyso's linux code.
memory-mapped or i/o-mapped access to the device registers, and always
choose memory-mapped if it is enabled. In particular, do _not_ explictly
disable the space we decide to not use, as this confuses some versions
of Alpha console software (which are arguably buggy because of this
problem).
Also, fix a logic error pointed out by Ross Harvey <ross@teraflop.com>
that would cause memory-mapped access to never be enabled.
by default if it's usable, but falling back to I/O space if mem isn't usable.
If TULIP_IOMAPPED is defined (default on the x86), prefer I/O space
then fall back to mem.
