August 2005 - Darwin-dev - lists.apple.com

Re: Slightly Off-Topic: Help on PCI Drivers wanted
by Alexander v. Below 31 Aug '05

31 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.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: Slightly Off-Topic: Help on PCI Drivers wanted
by Sean McBride 31 Aug '05

31 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Alexander von Below (below(a)mac.com) on 2005-08-30 10:29 said: >We are developing a Kernel Driver for a PCI Card, and are stuck with >a problem. (namely, using IOConnectMapMemory) We have filed a DTS >incident, but time is running out quickly. Is your problem a state secret? :) I don't think you'll get many answers if you don't describe the problem... or are you looking to hire someone? -- "First they ignore you, then they laugh at you, then they fight you, then you win." - Gandhi _______________________________________________ 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: Non-canonical mode on OS X?
by Gregory Weston 30 Aug '05

30 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Justin C. Walker wrote: Your code asks for one character, so the VMIN requirement will be satisfied before the VTIME requirement starts (VTIME is the number of time periods to wait after the last character is received; until the first character is received, there is no time requirement). I think that is why you get the return before the second burst. G _______________________________________________ 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… At first I was sure this wasn't the case, so I went back and re-read the interaction between VMIN and VTIME. Several times. Finally I figured out that I had, in fact, misread. Thanks for shoving me back onto the right track. This email sent to site_archiver(a)lists.apple.com

1 0

Slightly Off-Topic: Help on PCI Drivers wanted
by Alexander von Below 30 Aug '05

30 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Thanks Alex _______________________________________________ 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… I know this is not exactly the right list to post this, but I wanted to get a somewhat broader audience. This has been posted on Darwin- Drivers, but as of now without the success we had hoped for. We are developing a Kernel Driver for a PCI Card, and are stuck with a problem. (namely, using IOConnectMapMemory) We have filed a DTS incident, but time is running out quickly. If anyone here thinks they could help us, or knows someone who could, please contact me. This email sent to site_archiver(a)lists.apple.com

1 0

Storage access is compromising my function
by Mark Gilbert 30 Aug '05

30 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Folks. I am having a performance problem with a blocking Posix serial thread. Thanks Mark Gilbert -- mark(a)gallery.co.uk Tel: +44 208 340 5677 fax: +44 870 055 7790 http://www.gallery.co.uk New Product ! - Zen MetaCorder Location Sound Field Recorder http://www.metacorder.info _______________________________________________ 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… - My Posix serial thread is designed to QUICKLY reply to an incoming serial message. it blocks waiting for the start of the incoming message, then reads the remainder when it wakes, and quickly replies. - At any time there can be an extremely heavy load on the attached storage (SATA or SAN), since the application is dealing with Video material. - We observe that the ability for our Posix serial thread to reply _promptly_ is compromised by the load on the storage - ie. if the storage is being hit hard, at the same time the serial message is coming in, our thread seems to be delayed somewhat, by a matter of 10mSecs or more. I have experimented with thread priority, and setting the thread to realtime, but nothing seems to help. Using XSan seems to make the problem worse than using local SATA storage. The problem is worse at the point when video playback starts, and there is a massive rush on the storage. Its not clear exactly where the holdup is. I guess its possible that the serial driver is being held up rather than my thread itself - perhaps delaying the incoming bytes, or the outgoing reply. Can anyone advise how to raise my thread's status above that of the storage subsystem ? Any other tips for no-compromise performance on my thread (the thread doesn't do very much, but it needs to respond very quickly to the incoming data). This email sent to site_archiver(a)lists.apple.com

1 0

Re: Non-canonical mode on OS X?
by Justin C. Walker 29 Aug '05

29 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Aug 29, 2005, at 10:42 , Gregory Weston wrote: static int set_local_mode(int inFileDescriptor) { int theResult = 0; struct termios theOptions = {}; Regards, Justin -- Justin C. Walker, Curmudgeon-At-Large Institute for General Semantics -------- Some people have a mental horizon of radius zero, and call it their point of view. -- David Hilbert -------- _______________________________________________ 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… I'm trying to read a Mac's built-in modem in raw mode (after some help from this lists months ago, then a long period of downtime) and I'm running into something annoying. When data shows up, it comes through as a small burst (100 characters or so) followed by a brief pause (.02 seconds, say) and then another small burst (32 characters). The difficulty I'm having is that the read() I eventually issue returns before the second burst shows up. Here's the setup routine. if(tcgetattr(inFileDescriptor, &theOptions) == -1) {theResult = errno;} cfmakeraw(&theOptions); //theOptions.c_lflag &= ~(ICANON | ECHO); /* Clear ICANON and ECHO. */ theOptions.c_cc[VMIN] = 1; theOptions.c_cc[VTIME] = 5; theOptions.c_cflag |= (CS8 | CLOCAL); if(tcsetattr(inFileDescriptor, TCSANOW, &theOptions) == -1) {theResult = errno;} return theResult; } My understanding is that if this completes successfully, the read() should return after once there has been a half-second interval since the last character received. The buffer passed to read() is much larger than the returned data, and the commented line doesn't have any impact (c_lflag is already zero by the time we get there). Any advice? Your code asks for one character, so the VMIN requirement will be satisfied before the VTIME requirement starts (VTIME is the number of time periods to wait after the last character is received; until the first character is received, there is no time requirement). I think that is why you get the return before the second burst. Also, if at the time of the call, there are bytes queued up, they will be copied and if the VMIN requirement is thereby satisfied, return is immediate. This email sent to site_archiver(a)lists.apple.com

1 0

Non-canonical mode on OS X?
by Gregory Weston 29 Aug '05

29 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com static int set_local_mode(int inFileDescriptor) { int theResult = 0; struct termios theOptions = {}; I'm trying to read a Mac's built-in modem in raw mode (after some help from this lists months ago, then a long period of downtime) and I'm running into something annoying. When data shows up, it comes through as a small burst (100 characters or so) followed by a brief pause (.02 seconds, say) and then another small burst (32 characters). The difficulty I'm having is that the read() I eventually issue returns before the second burst shows up. Here's the setup routine. if(tcgetattr(inFileDescriptor, &theOptions) == -1) {theResult = errno;} cfmakeraw(&theOptions); //theOptions.c_lflag &= ~(ICANON | ECHO); /* Clear ICANON and ECHO. */ theOptions.c_cc[VMIN] = 1; theOptions.c_cc[VTIME] = 5; theOptions.c_cflag |= (CS8 | CLOCAL); if(tcsetattr(inFileDescriptor, TCSANOW, &theOptions) == -1) {theResult = errno;} return theResult; } My understanding is that if this completes successfully, the read() should return after once there has been a half-second interval since the last character received. The buffer passed to read() is much larger than the returned data, and the commented line doesn't have any impact (c_lflag is already zero by the time we get there). Any advice? _______________________________________________ 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

(no subject)
by plumber 27 Aug '05

27 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.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: libintl
by Martin Costabel 25 Aug '05

25 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com User-agent: Mozilla Thunderbird 1.0.2 (Macintosh/20050317) -- Martin _______________________________________________ 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… Rebecca Bendick wrote: Hi. Where can I get libintl? I have a version in /sw/lib, but it seems like I should have a native version in /usr/lib. Is the fink one OK? If not, how can I get and install a Darwin library? What's wrong with the one in /sw/lib?. Gnu gettext libraries installed by Fink or by darwinports are as "native" as GNU software can get. The one (or two actually) provided by Fink is at least sufficiently OK to be used in several hundred other packages. AFAIK the darwin sources do not contain gettext (the source of libintl). This email sent to site_archiver(a)lists.apple.com

1 0

Re: change MXUNIT f2c gcc 3.3 or 4.0
by Sam Hart 25 Aug '05

25 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Markus wrote: You should consider picking up this path again, perhaps with other versions. While building (and using) gcc from FSF sources isn't done as often as on other platforms, it should work anyways. S a m _______________________________________________ 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… not neck-of-celery, try building GCC 2.95.3 with GCC 4 :O) it builds fine with GCC 3.3 though. This email sent to site_archiver(a)lists.apple.com

1 0