January 2010 - Darwin-dev - lists.apple.com

Re: Is blocking I/O possible with 3-wire RS232?
by Stephen Hoffman 05 Jan '10

05 Jan '10

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Tue, 5 Jan 2010 09:49:34 -0500, Michael Ellis <mellis(a)ellissoftworks.com> writes: > So, my fundamental question is this: "Is there a way to configure a serial port to disregard the DSR input altogether and still have it perform blocking I/O?" Digital Equipment Corporation (DEC) boxes and VT-series terminals have used XON and XOFF in-band flow control for eons (DC1 and DC3, respectively), and one of the usual approach for a controller that required DSR/DTR hardware flow control or modem control was to wire the serial connector to feed a false signal into the controller. To jumper signal pins. Off the top and from some very dim memory and presuming DB25 connector pinouts, pins 2 and 3 are transmit and receive and wired either straight-through or crossover depending on the requirements, pins 4 and 5 are jumpered inside the connector, and (if you had enough wires for the end-to-end connection) 6, 8 and 20 were bridged within the connectors and with one wire between the two connectors. (Check that against EIA-232, too.) Three signal wires (rx, tx, gnd) would work with XON/XOFF in-band flow control, so long as you're not running full binary data down the serial line. I'm aware of a whole lot of serial wiring with four signal wires, and DEC sold a wiring system (DECconnect) based on 6-wire EIA-423 (DEC-423) wiring scheme and modified modular jacks (MMJs). With in-band flow-control, the size of your read buffer will be sized to and tied to how fast your code can issue an XOFF down the serial line, too. Your write speeds need not get too far ahead of the receiver, too, as you can need to shut off transmit very quickly lest data be lost. With some of these devices, you could dispense with XON/XOFF entirely, as some of the intelligent (cough, dumb, cough) controllers counted characters. A more recent solution tends to be a terminal server or analogous; to dedicate some hardware to deal with this serial stuff, and to avoid having the host have to deal with XON and XOFF. The host then connects to the hardware via a network path. This might be an Arduino or otherwise, or an IP terminal server. As for the Mac OS X I/O path here, I'll leave that to others. _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: 10.5.2 and -force_load
by Joel Reymont 05 Jan '10

05 Jan '10

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:received:received:subject:mime-version :content-type:from:in-reply-to:date:cc:content-transfer-encoding :message-id:references:to:x-mailer; bh=mQTvLSsFXevXyQnH6A1PGw29Q4KEZhdZ72rsSfOSGBU=; b=QQxtp4Ey2oK/Ma2KuNmMhdsubPxF2oPa62S1H6+3QHm2pudcKUUCNQsFsvGoIR1TFD 7hACCmPFdAFe/9JGEE7Xm2VxL4G7n/DQ7CTdTkZJRGVKNcpFsQLt6O994nXBPz+WBfxr sBIpHcttsLrB7QqBITiXg5kndB5GNYHcmJjBU= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=subject:mime-version:content-type:from:in-reply-to:date:cc :content-transfer-encoding:message-id:references:to:x-mailer; b=dW+GvkDkSQlw+bn2jQxN4dBgwfzl4gXe4aUQ1Oyt14cKbDTnr+dj9I0Wf6ZgS8mjLz gOgk9Sy+C/cy1wJNTFXbGZxYamkBWUisR44vZJPGQVDeOXB1TgjJQXercvq7NoUc8ykl /jn6cHlwbaZkrE2QGxFV0PdShYUd+JFxLZSB4= On Jan 5, 2010, at 4:56 PM, Peter O'Gorman wrote: > 1) Do what firefox/mozilla has been doing in this case - unpack the archives and add all the .o. Shame on me! I saw this being done for dylibs. Thanks! --- http://wagerlabs.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: 10.5.2 and -force_load
by Peter O'Gorman 05 Jan '10

05 Jan '10

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1.4pre) Gecko/20091014 Fedora/3.0-2.8.b4.fc11 Thunderbird/3.0b4 I'm trying to use -force_load with ld64-85.2.1 on 10.5.2 but there's no such option. This is with XCode 3.1 and gcc 4.2. It's my understanding that -noall_load does not disable -all_load either so I can't use these around static libraries. Peter -- Peter O'Gorman http://pogma.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… On 01/05/2010 10:46 AM, Joel Reymont wrote: What are my options? I updated to XCode 3.1.4 but it has the same version of ld. 1) Do what firefox/mozilla has been doing in this case - unpack the archives and add all the .o. 2) Build your own ld64 on 10.5.x with this patch: http://svn.macosforge.org/repository/odcctools/trunk/patches/force_load.diff I suggest (1), since that doesn't require everyone who wants to build mozilla to also rebuild their linker. Mac OS X 10.5 and earlier are not the only platforms lacking --whole-archive/--no-whole-archive equivalents, it is necessary to keep the unpacking of archives in place in order to be portable. This email sent to site_archiver(a)lists.apple.com

1 0

10.5.2 and -force_load
by Joel Reymont 05 Jan '10

05 Jan '10

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:received:received:from:content-type :content-transfer-encoding:subject:date:message-id:to:mime-version :x-mailer; bh=/FLxBtDC6y0nHBSp4Cf2H7APHa4Ke7y0TZgdLPW18j8=; b=nKKrbVdzg8SlPnkM3XOMTqfnwfDjOus+jgYZbAX4q3p0unJ4hH3QFfAs/20YnDlO1W Ti6IXTPBKsy00e7NumCsCjULrLrzQh25nhki74zp90Nm60J4EqzA3LQN2zNL9cYjtncT OWUZdB4CURicUn8Hj1/In0eahjAqexkpVmGvg= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=from:content-type:content-transfer-encoding:subject:date:message-id :to:mime-version:x-mailer; b=s1VZkL6NHQGo2DpcR5v6+0aE2xIe+gzs3zpNTRh+5L4fl/CzJGvc6BmbJNUHIbqNHF XcE+YVKZkOc4N2rLZeAqwvBL5wMpehbrUb3dn1uDhm1YlceC3hUCBLtLbRYsnNyifo10 BRYCSrz1SYX6ovmGREq4WebEHAJ9wT1shvhVs= I'm trying to use -force_load with ld64-85.2.1 on 10.5.2 but there's no such option. This is with XCode 3.1 and gcc 4.2. It's my understanding that -noall_load does not disable -all_load either so I can't use these around static libraries. What are my options? I updated to XCode 3.1.4 but it has the same version of ld. FYI, this is for a static build of Firefox, to ensure that loaded dylibs have the symbols they need. No, I cannot upgrade from 10.5.2 at the moment but I'm Thanks, Joel --- http://wagerlabs.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Is blocking I/O possible with 3-wire RS232?
by David Leimbach 05 Jan '10

05 Jan '10

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:in-reply-to:references :date:message-id:subject:from:to:cc:content-type; bh=AzOVDqoiqhOUWTP+7pcyB9Sqf2NVS5pNNstsWjQ0HJM=; b=oo2mEZ/WO3SO33jj2dV9xott6pPsMIwSnRvFXC9zXFfSGHANKDupyRA5y3qzWjymxe L770D+yxychN/fmCGGkMH1q2IEI/xjc4wwtiRxFY7NDhpcIahoHDeDMlhDWYd3+0FOiC 0UeJw05IjgF2OP4/w8PiCcCsA1SjuOBbAqFh8= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; b=sCamoU1vB53NQXAegC/r9gwegu0OWDDhk1OZ0s1pKpZstW3Z2Hwu/odPqmMPvWiNsC Zarx9kMhFWqhknZgLj4ZguklK/Bj6NzMSUBuisEU5NggBXgXld46xUhZNH2YLQc3OUgv NWkVnkrFMqQQnRcnLUkNHWYM6+cf4fy1dqSlI= _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Is blocking I/O possible with 3-wire RS232?
by Michael Ellis 05 Jan '10

05 Jan '10

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Hi folks. I am writing a program that communicates with an 8-bit, 64K computer. The communication protocol is similar enough to RS232 that it is possible to build a cable based on a USB-Serial project board and TTL-level signals. In my case, I have built such a cable, and I am able to use it to communicate with the remote machine successfully. The problem I am having is in configuring my software to work with other cables that wire the DSR input to an out-of-band signal. That is to say, some people are using cables that have an unrelated signal wired to the DSR input, and this seems to case trouble for my software. From what I can tell, the serial communication driver perceives that the remote device "comes and goes" as the signal level changes on the DSR input. The documentation for programming serial ports suggests that opening the tty device-file with the FNODELAY argument will cause DSR signal to be disregarded, but doing this requires aggressive polling (in the form of repeated read() calls) that negatively affects the performance of my application. Furthermore, I cannot simply call read(), sleep() for a quantum, and then rely on the serial buffer to catch any data the arrives in the meantime because the remote computer has very tight timing requirements. The remote computer requires that responses to commands arrive within a very narrow time window. So, my fundamental question is this: "Is there a way to configure a serial port to disregard the DSR input altogether and still have it perform blocking I/O?" Thanks for your thoughts. ========================= Michael F. Ellis President Ellis Softworks Inc. ---------- Phone: (941) 713-0361 Email: mellis(a)ellissoftworks.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Call for participation to the Alt-OS devroom at FOSDEM (Extended deadline)
by François Revol 02 Jan '10

02 Jan '10

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Hello, I have been unavailable for a week, I tried to contact the project on IRC but it seems the info wasn't picked up and forwarded... This year, the FOSDEM (in Brussels, 6-7th February 2010) will host the "Alt-OS" devroom on sunday from 9:00 to 17:00. This devroom is meant as a place for Free and OpenSource Alternative Operating System projects to share ideas and work on topics like: * distinguishing design and features, * device drivers and how to share them, * sharing code for internal components, * internal usage of other FOSS projects to build upon, * application portability, * interoperability, o file formats, o extended attributes... We are very interested to hear about your project on all those aspects, and of course we'd be happy to help you set up a hacking session or other form of gathering eventually, depending on the schedule. See the complete call for participation. http://groups.google.com/group/rosetta-os/browse_thread/thread/16964e21bef2… The deadline for the propositions for the Alt-OS devroom at FOSDEM has been extended to 2010-01-05, to allow at least one talk per project. The shedule must be sent rapidly to the FOSDEM team so we can't stretch it much more. François. _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: How to suppress crash reporter?
by James W. Walker 02 Jan '10

02 Jan '10

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com On Jan 1, 2010, at 5:18 PM, James W. Walker wrote: > If I run a command line tool that crashes, I don't get the crash report dialog, but a crash report file does get created, and a line appears in the console log saying so. Is there a way to disable all that? I tried the task_set_exception_ports technique described here <http://lists.apple.com/archives/unix-porting/2007/Mar/msg00009.html> but it didn't work. I also tried installing a signal handler for SIGSEGV, the particular kind of crash I tested, and that didn't make any difference either. Never mind, I figured it out. I needed to set up an alternate stack with sigaltstack, and pass the SA_ONSTACK flag to sigaction. _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

How to suppress crash reporter?
by James W. Walker 02 Jan '10

02 Jan '10

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com If I run a command line tool that crashes, I don't get the crash report dialog, but a crash report file does get created, and a line appears in the console log saying so. Is there a way to disable all that? I tried the task_set_exception_ports technique described here <http://lists.apple.com/archives/unix-porting/2007/Mar/msg00009.html> but it didn't work. I also tried installing a signal handler for SIGSEGV, the particular kind of crash I tested, and that didn't make any difference either. _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0