I noticed that the traffic on the list is very low, so I'll throw something
in...
Why don't we put a request_mem in the kernel? After all, there are many
memory-mapped devices for which we have a driver, and having two devices which
map memory on the same range is guaranteed to be destructive... If we have a
function for IRQ and one for io ports, I ask why not for memory ranges?
Then: I'm using 2.6.4-950530 to compile the kernel, and I get tons of warnings.
Since I don't like that at all, I'll start looking around to stop them. But I
don't like doing useless work, so if somebody thinks it's not a good idea, tell
me now!
Lastly: I was thinking about creating a /proc/serial/* hierarchy like the new
/proc/scsi. I don't know you, but since I do consulting job, mainly to ISPs and
mainly remotely, I'd like to be able to find in a single place everything I
can. This would include:
Hardware port type
Resources used (irq, io port)
Buffer length/used/free
Hardware status of control lines (CD, RTS, CTS, DTR)
Software status of the port (closed, open for callin, open for callout)
...
and anything you will come up with. I'm open for comments on this...
Bye!
Andrea
--
main(){char*s="DQNO_DJUQQAJQYQQAQQQQOOQ_QQEA_QQQIAQQQNQ_Q";int i,j,k,l,m,n;
for(j=0;j<7;j++)for(l=0;m=l-6+j,i=m/6,n=j*6+i,k=1<<m%6,l<43-j;l++)putchar(l<6-
j?' ':l==42-j?'\n':s[n]&k?(j==6||!(s[n+6]&k)?'_':' '):k/2&&s[n]&k/2?'/':' ');}