Clean up some isp_attach time stuff- if ISP2100_FABRIC is defined try *really*
hard to make sure that we get the firmware state to FW_READY and see the
loop state where the Port Database is ready to be gathered- if we don't do
this it's unlikely we will be able to correctly query the nameserver because
we won't see that we're on a fabric.
Clean up the completely broken and stupid attempts to hot switch the
'slow' start routine out of the way. Sigh.
Turn speed announcements into CFGPRINTF functions (available only if DEBUG
defined).
On platforms where we can use weak aliases, use
__weak_alias(__errno, __errno_func);
and on platforms where we cannot, just compile the function as
__errno, since on those platforms we cannot make libc thread-safe,
at least using this trick.
Are there any platforms where we cannot do __weak_alias() anymore?
the WHOIS service, again. New target: whois.networksolutions.com
This program should be replaced with one that does WHOIS++ so that it
can hit all the registries for data.
- Use explicit assembly; don't want the compiler optimizing things too
much here.
- Subtract 2 for each iteration of the loop, rather than 3. The loop
consists of only 2 instructions (even the C version compiled to a
loop of 2 instructions).
The latter change has squashed a fairly annoying timing bug in the
mii_bitbang module as used by the `ex' driver on my AlphaStation 500.
XXX delay() should maybe be rewritten again, to use the rpcc instruction.
option is given to the compiler. It's silly to have to specify it twice
on a command line (e.g. -mcpu=ev56 -Wa,-mev56), especially considering
that if you don't, and the compiler emits e.g. a BWX instruction, the
assembler will treat it as a macro, open-coding an equivalent, which can
have some serious unwanted side-effects in some situations.
Also, don't treat e.g. 21164 and ev5 as equivalent; the assembler treats
them differently: 21164 enables PALcode-only instructions, while ev5
does not.
byte) boundaries don't work correctly. Make use of the SDR0/SDR1 registers
to adjust the buffer alignment when starting the DMA transfer. Block device
I/O done by the filesystems will usually be aligned correctly, but character
device I/O may not be aligned correctly. This should fix the problems with
dump(8) failing randomly on DECstations.
