Re: RE : [fpc-devel] serial under linux, SerOpen blocks
there must be a reason why minicom uses O_NDELAY (=O_NONBLOCK) in case
O_NDELAY ist defined. Now i remember that i had the same problem several
years ago with cat file /dev/ttyUSB0, i think it was Fedora Core 3 or 4.
I assume this does not only happen with a pl2303.
#if defined(O_NDELAY) defined(F_SETFL)
portfd = open(dial_tty, O_RDWR|O_NDELAY|O_NOCTTY);
if (portfd = 0) {
/* Cancel the O_NDELAY flag. */
n = fcntl(portfd, F_GETFL, 0);
fcntl(portfd, F_SETFL, n ~O_NDELAY);
}
#else
if (portfd 0)
portfd = open(dial_tty, O_RDWR|O_NOCTTY);
#endif
On 08/18/2011 06:49 PM, Mark Morgan Lloyd wrote:
Ludo Brands wrote:
Some serial ports seem to have clocal unset, for some reason, which
causes it to block on open (waiting for the modem status
lines). Like
you, I've never seen it with any of my serial ports, either.
Might have to do with the defaulting to autobaud of some usbserial
chips. ___
Apparently a bug in a recent usb serial kernel module pl2303:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/661321
Well spotted :-) I was going to speculate that it might be something
to do with the Prolific driver- I think that the device I use most is
FTDI.
___
fpc-devel maillist - fpc-devel@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-devel
[fpc-devel] serial under linux, SerOpen blocks
I need some hint regarding opening a serial (ttyUSB) port using the unit
serial. Minicom works fine under my user id. When i call
SerOpen('/dev/ttyUSB0'), fpopen blocks forever. When i first start
minicom on that port, SerOpen and all other functions of serial.pp works
fine.
ttyUSB0 is:
crw-rw 1 root dialout 188, 0 Aug 18 09:37 /dev/ttyUSB0
changing it to
crw-rw-rw- 1 root dialout 188, 0 Aug 18 09:38 ttyUSB0
makes no difference. (my user id has the supplemental group dialout
added) SerOpen will always block if minicom is not started on the same
port, the same applies if i try it wit root.
stty with minicom started:
speed 19200 baud; rows 0; columns 0; line = 0;
intr = ^C; quit = ^\; erase = ^?; kill = ^U; eof = ^A; eol = undef;
eol2 = undef; swtch = undef; start = ^Q; stop = ^S;
susp = ^Z; rprnt = ^R; werase = ^W; lnext = ^V; flush = ^O; min = 1;
time = 5;
-parenb -parodd cs8 hupcl -cstopb cread clocal crtscts
ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr -icrnl ixon
ixoff -iuclc -ixany -imaxbel -iutf8
-opost -olcuc -ocrnl -onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0
bs0 vt0 ff0
-isig -icanon -iexten -echo -echoe -echok -echonl -noflsh -xcase -tostop
-echoprt -echoctl -echoke
stty without mnicom started:
speed 19200 baud; rows 0; columns 0; line = 0;
intr = ^C; quit = ^\; erase = ^?; kill = ^U; eof = ^A; eol = undef;
eol2 = undef; swtch = undef; start = ^Q; stop = ^S;
susp = ^Z; rprnt = ^R; werase = ^W; lnext = ^V; flush = ^O; min = 1;
time = 0;
-parenb -parodd cs8 hupcl -cstopb cread -clocal -crtscts
-ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr -icrnl
-ixon -ixoff -iuclc -ixany -imaxbel -iutf8
-opost -olcuc -ocrnl -onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0
bs0 vt0 ff0
-isig -icanon -iexten -echo -echoe -echok -echonl -noflsh -xcase -tostop
-echoprt -echoctl -echoke
diff
time = 0; -clocal -crtscts -ixon -ixof
Fedora 14, 64 Bit
2.6.35.13-92.fc14.x86_64 #1 SMP Sat May 21 17:26:25 UTC 2011 x86_64
x86_64 x86_64 GNU/Linux
Thx
Armin
___
fpc-devel maillist - fpc-devel@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] serial under linux, SerOpen blocks
Armin Diehl wrote:
I need some hint regarding opening a serial (ttyUSB) port using the unit
serial. Minicom works fine under my user id. When i call
SerOpen('/dev/ttyUSB0'), fpopen blocks forever. When i first start
minicom on that port, SerOpen and all other functions of serial.pp works
fine.
ttyUSB0 is:
crw-rw 1 root dialout 188, 0 Aug 18 09:37 /dev/ttyUSB0
changing it to
crw-rw-rw- 1 root dialout 188, 0 Aug 18 09:38 ttyUSB0
makes no difference. (my user id has the supplemental group dialout
added) SerOpen will always block if minicom is not started on the same
port, the same applies if i try it wit root.
stty with minicom started:
speed 19200 baud; rows 0; columns 0; line = 0;
intr = ^C; quit = ^\; erase = ^?; kill = ^U; eof = ^A; eol = undef;
eol2 = undef; swtch = undef; start = ^Q; stop = ^S;
susp = ^Z; rprnt = ^R; werase = ^W; lnext = ^V; flush = ^O; min = 1;
time = 5;
-parenb -parodd cs8 hupcl -cstopb cread clocal crtscts
ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr -icrnl ixon
ixoff -iuclc -ixany -imaxbel -iutf8
-opost -olcuc -ocrnl -onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0
bs0 vt0 ff0
-isig -icanon -iexten -echo -echoe -echok -echonl -noflsh -xcase -tostop
-echoprt -echoctl -echoke
stty without mnicom started:
speed 19200 baud; rows 0; columns 0; line = 0;
intr = ^C; quit = ^\; erase = ^?; kill = ^U; eof = ^A; eol = undef;
eol2 = undef; swtch = undef; start = ^Q; stop = ^S;
susp = ^Z; rprnt = ^R; werase = ^W; lnext = ^V; flush = ^O; min = 1;
time = 0;
-parenb -parodd cs8 hupcl -cstopb cread -clocal -crtscts
-ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr -icrnl
-ixon -ixoff -iuclc -ixany -imaxbel -iutf8
-opost -olcuc -ocrnl -onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0
bs0 vt0 ff0
-isig -icanon -iexten -echo -echoe -echok -echonl -noflsh -xcase -tostop
-echoprt -echoctl -echoke
diff
time = 0; -clocal -crtscts -ixon -ixof
Fedora 14, 64 Bit
2.6.35.13-92.fc14.x86_64 #1 SMP Sat May 21 17:26:25 UTC 2011 x86_64
x86_64 x86_64 GNU/Linux
Assuming that you mean the standard serial.pp, that's not something I've
seen here /but/ I'm using a 32-bit kernel. I think that the first thing
I'd explore is whether you need the incoming control lines (cts, dsr,
cd) to be asserted before the function completes, but quite frankly if
I'm reading your
When i first start minicom on that port, SerOpen and all other
functions of serial.pp works fine.
correctly then it sounds like a kernel or driver issue.
Anybody: I've modified serial.pp to support Win-32 and fix a couple of
issues. Bug http://62.166.198.202/view.php?id=18946
--
Mark Morgan Lloyd
markMLl .AT. telemetry.co .DOT. uk
[Opinions above are the author's, not those of his employers or colleagues]
___
fpc-devel maillist - fpc-devel@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] serial under linux, SerOpen blocks
Armin Diehl wrote:
I need some hint regarding opening a serial (ttyUSB) port using the
unit serial. Minicom works fine under my user id. When i call
SerOpen('/dev/ttyUSB0'), fpopen blocks forever. When i first start
minicom on that port, SerOpen and all other functions of serial.pp
works fine.
I don't use serial.pp but here is how I open a serial port for basic
communication (no flow control, modem control or stuff like that) under
linux. I remember this taking some experimentation to get right.
procedure tlserial.open;
var
fd : longint;
config : termios;
baudrateos : longint;
begin
fd := fpopen(device,O_RDWR or O_NOCTTY);
if isatty(fd)=0 then begin
writeln('not a tty');
halt(1);
end;
fillchar(config,sizeof(config),#0);
config.c_cflag := CLOCAL or CREAD;
cfmakeraw(config);
case baudrate of
50: baudrateos := B50;
75: baudrateos := B75;
110:baudrateos := B110;
134:baudrateos := B134;
150:baudrateos := B150;
200:baudrateos := B200;
300:baudrateos := B300;
600:baudrateos := B600;
1200: baudrateos := B1200;
1800: baudrateos := B1800;
2400: baudrateos := B2400;
4800: baudrateos := B4800;
9600: baudrateos := B9600;
19200: baudrateos := B19200;
38400: baudrateos := B38400;
57600: baudrateos := B57600;
115200: baudrateos := B115200;
230400: baudrateos := B230400;
else raise exception.create('unrecognised baudrate');
end;
cfsetispeed(config,baudrateos);
cfsetospeed(config,baudrateos);
config.c_cc[VMIN] := 1;
config.c_cc[VTIME] := 0;
if tcsetattr(fd,TCSAFLUSH,config) 0 then begin
writeln('could not set termios attributes');
halt(3);
end;
dup(fd);
closehandles := true;
end;
___
fpc-devel maillist - fpc-devel@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] serial under linux, SerOpen blocks
Hi Peter,
that is the same as in the standard serial.pp:
Result := fpopen(DeviceName, O_RDWR or O_NOCTTY);
and this fpopen will block as long as minicom is not started on that
device.
On 08/18/2011 11:10 AM, peter green wrote:
Armin Diehl wrote:
I need some hint regarding opening a serial (ttyUSB) port using the
unit serial. Minicom works fine under my user id. When i call
SerOpen('/dev/ttyUSB0'), fpopen blocks forever. When i first start
minicom on that port, SerOpen and all other functions of serial.pp
works fine.
I don't use serial.pp but here is how I open a serial port for basic
communication (no flow control, modem control or stuff like that)
under linux. I remember this taking some experimentation to get right.
procedure tlserial.open;
var
fd : longint;
config : termios;
baudrateos : longint;
begin
fd := fpopen(device,O_RDWR or O_NOCTTY);
if isatty(fd)=0 then begin
writeln('not a tty');
halt(1);
end;
fillchar(config,sizeof(config),#0);
config.c_cflag := CLOCAL or CREAD;
cfmakeraw(config);
case baudrate of
50: baudrateos := B50;
75: baudrateos := B75;
110:baudrateos := B110;
134:baudrateos := B134;
150:baudrateos := B150;
200:baudrateos := B200;
300:baudrateos := B300;
600:baudrateos := B600;
1200: baudrateos := B1200;
1800: baudrateos := B1800;
2400: baudrateos := B2400;
4800: baudrateos := B4800;
9600: baudrateos := B9600;
19200: baudrateos := B19200;
38400: baudrateos := B38400;
57600: baudrateos := B57600;
115200: baudrateos := B115200;
230400: baudrateos := B230400;else raise
exception.create('unrecognised baudrate');
end;
cfsetispeed(config,baudrateos);
cfsetospeed(config,baudrateos);
config.c_cc[VMIN] := 1;
config.c_cc[VTIME] := 0;
if tcsetattr(fd,TCSAFLUSH,config) 0 then begin
writeln('could not set termios attributes');
halt(3);
end;
dup(fd);
closehandles := true;
end;
___
fpc-devel maillist - fpc-devel@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-devel
___
fpc-devel maillist - fpc-devel@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] serial under linux, SerOpen blocks
On 18/08/11 10:34, Armin Diehl wrote: Hi Peter, that is the same as in the standard serial.pp: Result := fpopen(DeviceName, O_RDWR or O_NOCTTY); and this fpopen will block as long as minicom is not started on that device. Try to add O_NONBLOCK to the flags with fpopen. Then set CLOCAL like Peter suggested. Then disable O_NONBLOCK again with fcntl if you want blocking I/O back. Henry ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] serial under linux, SerOpen blocks
Just tried it on a 32 bit system (Linux t41.armin.d
2.6.35.13-92.fc14.i686 #1 SMP Sat May 21 17:39:42 UTC 2011 i686 i686
i386 GNU/Linux) as root
uses baseunix;
var handle : longint;
begin
writeln('Start: fpopen');
handle := fpopen('/dev/ttyUSB0',O_RDWR or O_NOCTTY);
writeln ('Result: ',Handle);
if handle 0 then fpclose(handle);
end.
and exactly the same problem. fpopen blocks if minicom is not started on
the same port. I think i have to figure out how minicom (or stty) opens
the port.
btw, ttyUSB0 is a Prolific pl2303:
Aug 18 11:47:18 t41 kernel: [ 599.994081] usb 3-2: new full speed USB
device using uhci_hcd and address 2
Aug 18 11:47:18 t41 kernel: [ 600.140165] usb 3-2: New USB device
found, idVendor=067b, idProduct=2303
Aug 18 11:47:18 t41 kernel: [ 600.140180] usb 3-2: New USB device
strings: Mfr=1, Product=2, SerialNumber=0
Aug 18 11:47:18 t41 kernel: [ 600.140191] usb 3-2: Product: USB-Serial
Controller
Aug 18 11:47:18 t41 kernel: [ 600.140200] usb 3-2: Manufacturer:
Prolific Technology Inc.
Aug 18 11:47:18 t41 mtp-probe: checking bus 3, device 2:
/sys/devices/pci:00/:00:1d.1/usb3/3-2
Aug 18 11:47:18 t41 mtp-probe: bus: 3, device: 2 was not an MTP device
Aug 18 11:47:18 t41 kernel: [ 600.329041] usbcore: registered new
interface driver usbserial
Aug 18 11:47:18 t41 kernel: [ 600.329070] USB Serial support registered
for generic
Aug 18 11:47:18 t41 kernel: [ 600.329948] usbcore: registered new
interface driver usbserial_generic
Aug 18 11:47:18 t41 kernel: [ 600.329953] usbserial: USB Serial Driver core
Aug 18 11:47:18 t41 kernel: [ 600.345812] USB Serial support registered
for pl2303
Aug 18 11:47:18 t41 kernel: [ 600.346655] pl2303 3-2:1.0: pl2303
converter detected
Aug 18 11:47:18 t41 kernel: [ 600.358507] usb 3-2: pl2303 converter now
attached to ttyUSB0
Aug 18 11:47:18 t41 kernel: [ 600.359439] usbcore: registered new
interface driver pl2303
Aug 18 11:47:18 t41 kernel: [ 600.359445] pl2303: Prolific PL2303 USB
to serial adaptor driver
On 08/18/2011 10:22 AM, Mark Morgan Lloyd wrote:
Armin Diehl wrote:
I need some hint regarding opening a serial (ttyUSB) port using the
unit serial. Minicom works fine under my user id. When i call
SerOpen('/dev/ttyUSB0'), fpopen blocks forever. When i first start
minicom on that port, SerOpen and all other functions of serial.pp
works fine.
ttyUSB0 is:
crw-rw 1 root dialout 188, 0 Aug 18 09:37 /dev/ttyUSB0
changing it to
crw-rw-rw- 1 root dialout 188, 0 Aug 18 09:38 ttyUSB0
makes no difference. (my user id has the supplemental group dialout
added) SerOpen will always block if minicom is not started on the
same port, the same applies if i try it wit root.
stty with minicom started:
speed 19200 baud; rows 0; columns 0; line = 0;
intr = ^C; quit = ^\; erase = ^?; kill = ^U; eof = ^A; eol = undef;
eol2 = undef; swtch = undef; start = ^Q; stop = ^S;
susp = ^Z; rprnt = ^R; werase = ^W; lnext = ^V; flush = ^O; min = 1;
time = 5;
-parenb -parodd cs8 hupcl -cstopb cread clocal crtscts
ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr -icrnl
ixon ixoff -iuclc -ixany -imaxbel -iutf8
-opost -olcuc -ocrnl -onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0
bs0 vt0 ff0
-isig -icanon -iexten -echo -echoe -echok -echonl -noflsh -xcase
-tostop -echoprt -echoctl -echoke
stty without mnicom started:
speed 19200 baud; rows 0; columns 0; line = 0;
intr = ^C; quit = ^\; erase = ^?; kill = ^U; eof = ^A; eol = undef;
eol2 = undef; swtch = undef; start = ^Q; stop = ^S;
susp = ^Z; rprnt = ^R; werase = ^W; lnext = ^V; flush = ^O; min = 1;
time = 0;
-parenb -parodd cs8 hupcl -cstopb cread -clocal -crtscts
-ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr -icrnl
-ixon -ixoff -iuclc -ixany -imaxbel -iutf8
-opost -olcuc -ocrnl -onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0
bs0 vt0 ff0
-isig -icanon -iexten -echo -echoe -echok -echonl -noflsh -xcase
-tostop -echoprt -echoctl -echoke
diff
time = 0; -clocal -crtscts -ixon -ixof
Fedora 14, 64 Bit
2.6.35.13-92.fc14.x86_64 #1 SMP Sat May 21 17:26:25 UTC 2011 x86_64
x86_64 x86_64 GNU/Linux
Assuming that you mean the standard serial.pp, that's not something
I've seen here /but/ I'm using a 32-bit kernel. I think that the first
thing I'd explore is whether you need the incoming control lines (cts,
dsr, cd) to be asserted before the function completes, but quite
frankly if I'm reading your
When i first start minicom on that port, SerOpen and all other
functions of serial.pp works fine.
correctly then it sounds like a kernel or driver issue.
Anybody: I've modified serial.pp to support Win-32 and fix a couple of
issues. Bug http://62.166.198.202/view.php?id=18946
___
fpc-devel maillist - fpc-devel@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] serial under linux, SerOpen blocks
In our previous episode, Armin Diehl said: that is the same as in the standard serial.pp: Result := fpopen(DeviceName, O_RDWR or O_NOCTTY); and this fpopen will block as long as minicom is not started on that device. Yes. And seems that minicom turns of xon/xoff. You could try to verify that by disabling it with stty. ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] serial under linux, SerOpen blocks
Yes, found that a few minutes ago (strace is nice), micom does:
open(/dev/ttyUSB0, O_RDWR|O_NOCTTY|O_NONBLOCK) = 3
fcntl(3, F_GETFL) = 0x8802 (flags
O_RDWR|O_NONBLOCK|O_LARGEFILE)
fcntl(3, F_SETFL, O_RDWR|O_LARGEFILE) = 0
ioctl(3, SNDCTL_TMR_TIMEBASE or SNDRV_TIMER_IOCTL_NEXT_DEVICE or TCGETS,
{B19200 -opost -isig -icanon -echo ...}) = 0
ioctl(3, TIOCMGET, [TIOCM_DTR|TIOCM_RTS]) = 0
ioctl(3, SNDCTL_TMR_TIMEBASE or SNDRV_TIMER_IOCTL_NEXT_DEVICE or TCGETS,
{B19200 -opost -isig -icanon -echo ...}) = 0
ioctl(3, SNDCTL_TMR_START or SNDRV_TIMER_IOCTL_TREAD or TCSETS, {B19200
-opost -isig -icanon -echo ...}) = 0
ioctl(3, SNDCTL_TMR_TIMEBASE or SNDRV_TIMER_IOCTL_NEXT_DEVICE or TCGETS,
{B19200 -opost -isig -icanon -echo ...}) = 0
ioctl(3, TIOCMGET, [TIOCM_DTR|TIOCM_RTS]) = 0
ioctl(3, TIOCMSET, [TIOCM_DTR|TIOCM_RTS]) = 0
ioctl(3, SNDCTL_TMR_TIMEBASE or SNDRV_TIMER_IOCTL_NEXT_DEVICE or TCGETS,
{B19200 -opost -isig -icanon -echo ...}) = 0
ioctl(3, SNDCTL_TMR_START or SNDRV_TIMER_IOCTL_TREAD or TCSETS, {B19200
-opost -isig -icanon -echo ...}) = 0
ioctl(3, SNDCTL_TMR_TIMEBASE or SNDRV_TIMER_IOCTL_NEXT_DEVICE or TCGETS,
{B19200 -opost -isig -icanon -echo ...}) = 0
ioctl(3, SNDCTL_TMR_TIMEBASE or SNDRV_TIMER_IOCTL_NEXT_DEVICE or TCGETS,
{B19200 -opost -isig -icanon -echo ...}) = 0
ioctl(3, SNDCTL_TMR_START or SNDRV_TIMER_IOCTL_TREAD or TCSETS, {B19200
-opost -isig -icanon -echo ...}) = 0
ioctl(3, SNDCTL_TMR_TIMEBASE or SNDRV_TIMER_IOCTL_NEXT_DEVICE or TCGETS,
{B19200 -opost -isig -icanon -echo ...}) = 0
ioctl(3, SNDCTL_TMR_TIMEBASE or SNDRV_TIMER_IOCTL_NEXT_DEVICE or TCGETS,
{B19200 -opost -isig -icanon -echo ...}) = 0
ioctl(3, SNDCTL_TMR_START or SNDRV_TIMER_IOCTL_TREAD or TCSETS, {B19200
-opost -isig -icanon -echo ...}) = 0
ioctl(3, SNDCTL_TMR_TIMEBASE or SNDRV_TIMER_IOCTL_NEXT_DEVICE or TCGETS,
{B19200 -opost -isig -icanon -echo ...}) = 0
ioctl(3, TCFLSH, 0x2) = 0
should that be changed in the standard serial.pp ?
On 08/18/2011 11:51 AM, Henry Vermaak wrote:
On 18/08/11 10:34, Armin Diehl wrote:
Hi Peter,
that is the same as in the standard serial.pp:
Result := fpopen(DeviceName, O_RDWR or O_NOCTTY);
and this fpopen will block as long as minicom is not started on that
device.
Try to add O_NONBLOCK to the flags with fpopen. Then set CLOCAL like
Peter suggested. Then disable O_NONBLOCK again with fcntl if you want
blocking I/O back.
Henry
___
fpc-devel maillist - fpc-devel@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-devel
___
fpc-devel maillist - fpc-devel@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] serial under linux, SerOpen blocks
In our previous episode, Armin Diehl said: should that be changed in the standard serial.pp ? I'm not sure, since this leaves nonblocking behaviour on by default, moreover serial is pan unix, and non linuxes might react differently. Btw do you see minicom do a sequence of opening in nonblock, setting params and reopening without nonblock ? ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] serial under linux, SerOpen blocks
On 18/08/11 11:21, Marco van de Voort wrote: In our previous episode, Armin Diehl said: should that be changed in the standard serial.pp ? I'm not sure, since this leaves nonblocking behaviour on by default, moreover serial is pan unix, and non linuxes might react differently. No it doesn't, the fcntl call disables it again. Henry ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] serial under linux, SerOpen blocks
Hi Marco, minicom calls open with O_NONBLOCK and resets O_NONBLOCK directly after the open call: open(/dev/ttyUSB0, O_RDWR|O_NOCTTY|O_NONBLOCK) = 3 fcntl(3, F_GETFL) = 0x8802 (flags O_RDWR|O_NONBLOCK|O_LARGEFILE) fcntl(3, F_SETFL, O_RDWR|O_LARGEFILE) = 0 doing the same in serial.pp should not change the blocking behaviour of serial.pp. function SerOpen(const DeviceName: String): TSerialHandle; begin Result := fpopen(DeviceName, O_RDWR or O_NOCTTY or O_NONBLOCK); // AD: O_NONBLOCK was missing if result -1 then fpfcntl(Result,F_SETFL,O_RDWR or O_NOCTTY); //AD: non blocking off again end; On 08/18/2011 12:21 PM, Marco van de Voort wrote: In our previous episode, Armin Diehl said: should that be changed in the standard serial.pp ? I'm not sure, since this leaves nonblocking behaviour on by default, moreover serial is pan unix, and non linuxes might react differently. Btw do you see minicom do a sequence of opening in nonblock, setting params and reopening without nonblock ? ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] serial under linux, SerOpen blocks
On 18/08/11 12:00, Armin Diehl wrote: Hi Marco, minicom calls open with O_NONBLOCK and resets O_NONBLOCK directly after the open call: open(/dev/ttyUSB0, O_RDWR|O_NOCTTY|O_NONBLOCK) = 3 fcntl(3, F_GETFL) = 0x8802 (flags O_RDWR|O_NONBLOCK|O_LARGEFILE) fcntl(3, F_SETFL, O_RDWR|O_LARGEFILE) = 0 doing the same in serial.pp should not change the blocking behaviour of serial.pp. function SerOpen(const DeviceName: String): TSerialHandle; begin Result := fpopen(DeviceName, O_RDWR or O_NOCTTY or O_NONBLOCK); // AD: O_NONBLOCK was missing if result -1 then fpfcntl(Result,F_SETFL,O_RDWR or O_NOCTTY); //AD: non blocking off again Better use F_GETFL first, then flags = ~O_NONBLOCK, like minicom does. Henry ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] serial under linux, SerOpen blocks
yes, that is better function SerOpen(const DeviceName: String): TSerialHandle; var flags : cint; begin Result := fpopen(DeviceName, O_RDWR or O_NOCTTY or O_NONBLOCK); if result -1 then begin flags := fpfcntl(Result,F_GetFl); fpfcntl(Result,F_SETFL,flags and (not O_NONBLOCK)); end; end; results in open(/dev/ttyUSB0, O_RDWR|O_NOCTTY|O_NONBLOCK|O_LARGEFILE) = 10 fcntl(10, F_GETFL) = 0x8802 (flags O_RDWR|O_NONBLOCK|O_LARGEFILE) fcntl(10, F_SETFL, O_RDWR|O_LARGEFILE) = 0 i think this should be changed in serial.pp On 08/18/2011 01:18 PM, Henry Vermaak wrote: On 18/08/11 12:00, Armin Diehl wrote: Hi Marco, minicom calls open with O_NONBLOCK and resets O_NONBLOCK directly after the open call: open(/dev/ttyUSB0, O_RDWR|O_NOCTTY|O_NONBLOCK) = 3 fcntl(3, F_GETFL) = 0x8802 (flags O_RDWR|O_NONBLOCK|O_LARGEFILE) fcntl(3, F_SETFL, O_RDWR|O_LARGEFILE) = 0 doing the same in serial.pp should not change the blocking behaviour of serial.pp. function SerOpen(const DeviceName: String): TSerialHandle; begin Result := fpopen(DeviceName, O_RDWR or O_NOCTTY or O_NONBLOCK); // AD: O_NONBLOCK was missing if result -1 then fpfcntl(Result,F_SETFL,O_RDWR or O_NOCTTY); //AD: non blocking off again Better use F_GETFL first, then flags = ~O_NONBLOCK, like minicom does. Henry ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] serial under linux, SerOpen blocks
Armin Diehl wrote: yes, that is better function SerOpen(const DeviceName: String): TSerialHandle; var flags : cint; begin Result := fpopen(DeviceName, O_RDWR or O_NOCTTY or O_NONBLOCK); if result -1 then begin flags := fpfcntl(Result,F_GetFl); fpfcntl(Result,F_SETFL,flags and (not O_NONBLOCK)); end; end; results in open(/dev/ttyUSB0, O_RDWR|O_NOCTTY|O_NONBLOCK|O_LARGEFILE) = 10 fcntl(10, F_GETFL) = 0x8802 (flags O_RDWR|O_NONBLOCK|O_LARGEFILE) fcntl(10, F_SETFL, O_RDWR|O_LARGEFILE) = 0 i think this should be changed in serial.pp But why have I not seen this problem on Debian despite having used this unit extensively? And why have I not seen it when testing the modified unit on Solaris with the SerOpen() function being unchanged? -- Mark Morgan Lloyd markMLl .AT. telemetry.co .DOT. uk [Opinions above are the author's, not those of his employers or colleagues] ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] serial under linux, SerOpen blocks
On 18/08/11 14:11, Mark Morgan Lloyd wrote: But why have I not seen this problem on Debian despite having used this unit extensively? And why have I not seen it when testing the modified unit on Solaris with the SerOpen() function being unchanged? Some serial ports seem to have clocal unset, for some reason, which causes it to block on open (waiting for the modem status lines). Like you, I've never seen it with any of my serial ports, either. Henry ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: [fpc-devel] serial under linux, SerOpen blocks
In our previous episode, Henry Vermaak said: On 18/08/11 14:11, Mark Morgan Lloyd wrote: But why have I not seen this problem on Debian despite having used this unit extensively? And why have I not seen it when testing the modified unit on Solaris with the SerOpen() function being unchanged? Some serial ports seem to have clocal unset, for some reason, which causes it to block on open (waiting for the modem status lines). Like you, I've never seen it with any of my serial ports, either. Might have to do with the defaulting to autobaud of some usbserial chips. ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel
RE : [fpc-devel] serial under linux, SerOpen blocks
Some serial ports seem to have clocal unset, for some reason, which causes it to block on open (waiting for the modem status lines). Like you, I've never seen it with any of my serial ports, either. Might have to do with the defaulting to autobaud of some usbserial chips. ___ Apparently a bug in a recent usb serial kernel module pl2303: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/661321 Ludo ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel
Re: RE : [fpc-devel] serial under linux, SerOpen blocks
Ludo Brands wrote: Some serial ports seem to have clocal unset, for some reason, which causes it to block on open (waiting for the modem status lines). Like you, I've never seen it with any of my serial ports, either. Might have to do with the defaulting to autobaud of some usbserial chips. ___ Apparently a bug in a recent usb serial kernel module pl2303: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/661321 Well spotted :-) I was going to speculate that it might be something to do with the Prolific driver- I think that the device I use most is FTDI. -- Mark Morgan Lloyd markMLl .AT. telemetry.co .DOT. uk [Opinions above are the author's, not those of his employers or colleagues] ___ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel
