After all there's no momentum for a W5100-based Ethernet solution on the ATARI. So I revert to static linkage of the CS8900A driver as this improves program load times.
If the setup of socket 0 to 3 with 4+2+1+1KB is detected then the W5100 is _not_ initialized, otherwise it does set up socket 0 and 1 with 4KB each. Either way socket 0 is used - now with 4KB instead of 8KB as before.
CC_FASTCALL was introduced many years ago for the cc65 tool chain. It was never used for another tool chain. With a798b1d648 the cc65 tool chain doesn't need CC_FASTCALL anymore.
There are scenarios in which it is beneficial to search for an Etherne chip at several i/o locations. To do so the chip initialization is performed at several i/o locations until it succeeds. In order to allow for that operation model the i/o location fixup needs to be repeatable.
Note: This won't work with the RR-Net because the fixup bits overlap with the chip i/o bits.
Behave just like the CS8900A driver: Both the CS8900A and the LAN91C96 dynamically share a buffer for received packets and packets to be send. If the chip is exposed to a network with a lot of broadcasts the shared buffer might fill quicker with received packets than the 6502 reads them (via polling). So we might need to drop some received packets in order to be able to send anything at all.
The previous chip detection was inspired by the old IP65 driver code. For some reason it didn't work as expected. The new code is simpler and based on this statement in the chip datasheet: "The upper byte always reads as 33h and can be used to help determine the I/O location of the LAN91C96."
Made Ethernet drivers easier to consume by assembly programs.
* Replaced function pointers with JMP instructions.
* Provide return values additionally via Carry flag.
Reset Ethernet chips on initialization.
Both for the CS8900A and the W5100 the data sheets just say that
the RESET bit is automatically cleared after the RESET. This may
be interpreted in two ways:
1) There's no need to be afraid of reading the RESET bit as 1 and
unintentionally trigger a RESET by writing it back after ORing in
some other bit.
2) The RESET process isn't complete before the RESET bit hasn't
become 0 again.
It's impossible for me to empirically falsify the latter option
as the drivers are supposed to work on faster machines than the
ones I have access to. And if the RESET process includes things
like oscillators then the time to complete the RESET could differ
even between multiple exemplars of the same chip. Therefore I
opted to presume the latter option.
However that means a non-exsistent chip may cause an infinite
loop while waiting for the RESET bit to be cleared so I finally
added code to detect the presence of the Ethernet chips. There's
a risk of a chip being locked up in a way that makes the detection
fail - and therefore the RESET not being performed. This catch-22
needs to be solved by the user doing a hard RESET.
It cannot be ruled out that access to the address register triggers
an address auto-increment. Therefore a temporary address register
shadow is introduced to replace the access to the address regsiter.
Additionally there are several minor beautifications.
- Speed: The primary byte copy loops are reduzed to the bare minimum by adjusting the base pointer 'ptr' and loop register 'y' in such a way that the 'y' overflow matches the low byte of the loop size.
- Introduced a loop for setting the MAC address.
Additional minor fix:
- Properly start self modification with first location.
- Speed: The primary byte copy loops are reduzed to the bare minimum by adjusting the base pointer 'ptr' and loop register 'y' in such a way that the 'y' overflow matches the low byte of the loop size.
- Size: Factored out all repeated code into subroutines. Introduced a loop for setting the MAC address.
Additional minor changes:
- Activate frame reception as last step of initialization after CS8900A configuration.
- Properly set internal address bits used by the CS8900A.
functions for converting between host and network byte order. These
names are the de facto standard names for this functionality because
of the original BSD TCP/IP implementation. But they cause problems for
uIP/Contiki: some platforms define these names themselves (Mac OS,
most notably), causing compilation problems for Contiki on those
platforms.
This commit changes all htons to uip_htons instead. Same goes for
htonl, ntohs, and ntohl. All-caps versions as well.
