I am using pilot-link 0.9.5 on a Linux 2.4.16 kernel. I find that I
can run the various pilot-link tools just fine with $PILOTRATE set at
or below 19200. However, faster rates (38400, 57600, 115200) simply
do not work at all.
I am using a very modern machine with a very modern motherboard.
"setserial" confirms that I've got a proper 16550A UART. Surely the
hardware itself is capable of sustaining faster rates? (Or am I
mistaken?) How should I go about trying to determine why faster rates
fail? Serial communication is a mystery to me, and I'm really not
sure where to begin.
(Please forgive me if this is a FAQ; I suspect it must be. Honestly,
I did hunt through mailing list archives, bug databases, and other web
resources before posting. All I could find was a cryptic comment in
the pilot-link README and man page: "You are welcome to try higher
baud rates ... but various machines have various limitations." That's
not very informative.)
Technical details on the Palm side:
hardware: Palm Vx
OS: PalmOS 3.5.2
connection: DirectSerial; Speed: 115,200, Flow Ctl: Automatic
hacks: none active
Technical details on the Linux side:
CPU: Athlon XP 1800+
motherboard: Abit KG7
north bridge: AMD-760 [Irongate]
south bridge: VIA VT82C686 [Apollo Super South]
kernel: 2.4.16 (self built)
pilot-link: 0.9.5 (built by Ximian)
connection: /dev/pilot symlinked to /dev/ttyS0
setserial -ga output:
/dev/pilot, Line 0, UART: 16550A, Port: 0x03f8, IRQ: 4
Baud_base: 115200, close_delay: 50, divisor: 0
closing_wait: 3000
Flags: spd_normal skip_test
Representative example of a failed command:
% PILOTRATE=38400 pilot-xfer -l
Port: /dev/pilot
Please press the HotSync button now...
Connected...
Exiting on cancel, stopped before listing databases.
There's a delay of several seconds immediately after "Connected...",
during which the Palm times out and presents an error dialog. On the
Linux side, strace reveals that pilot-xfer is in a loop where it
repeatedly executes the following three system calls:
write(3, "\276\357\355\3\3\2\0\6\2\252\1\300\0\2.\0008Z", 18) = 18
nanosleep({0, 28000}, NULL) = 0
select(4, [3], NULL, NULL, {2, 0}) = 0 (Timeout)
Looks like that loop runs about ten times.
For the record, neither "setserial" nor "stty" has been used to change
any port settings.
Thanks in advance for any hints.
_______________________________________________
Pilot-unix mailing list
[EMAIL PROTECTED]
http://hcirisc.cs.binghamton.edu/mailman/listinfo/pilot-unix
