February 2006 - Darwin-dev - lists.apple.com

Re: mmap alignment.
by Shaun Wexler 13 Feb '06

13 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Feb 13, 2006, at 9:00 PM, Theo Schlossnagle wrote: _______________________________________________ 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… Using the vm_map() directly may be one approach. We already use pthread_mach_thread_np in part of the porting effort. As it will be allocating a tremendous amount of memory and efficient use is a goal, over-allocating and doing alignment ourselves isn't the right approach. All allocations larger than 4k are page-aligned by default. If you want blocks guaranteed to start on 64k virtual address boundaries, create a custom zone and factor all of your allocations thru one function, padding sizes to mod 64k, and throw away the first 1-15 pages (up to 60k) for the initial alignment. -- Shaun Wexler MacFOH http://www.macfoh.com "Problems cannot be solved by the same level of thinking that created them." - Albert Einstein This email sent to site_archiver(a)lists.apple.com

1 0

Re: mmap alignment.
by Theo Schlossnagle 13 Feb '06

13 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Domainkey-signature: q=dns; a=rsa-sha1; c=nofws; s=test; d=omniti.com; h=Received:In-Reply-To:References:Mime-Version:Content-Type:Message-Id:Cc:Content-Transfer-Encoding:From:Subject:Date:To:X-Mailer; b=QXN9J6kipwGqdzsuSVcDO5WK+SF2ncvhNQjLRd2TUC76jZk4+I6qFwsQMQawG7Or UlG1uE0wWdgDMjQmx1gnGmga/W2VreNwsTU4Ogp2oPahb9WmnKb6uy2UlhbgKsJu Domainkey-status: good On Feb 13, 2006, at 11:31 PM, Mike Smith wrote: On Feb 13, 2006, at 3:04 PM, darwin-dev-request(a)lists.apple.com wrote: I'm porting an app that uses mmap() and banks on its alignment behaviour on Solaris and Linux. I'd like to mmap a 64k and have the resulting address be aligned on a 64k boundary. I'm trying with: mmap((void *)65536, 65536, PROT_NONE, MAP_PRIVATE | MAP_ANON | MAP_NORESERVE | MAP_ALIGN, -1, 0); MAP_ALIGN is not supported on Mac OS X. Thanks everyone for the replies. // Theo Schlossnagle // Principal Engineer -- http://www.omniti.com/~jesus/ // OmniTI Computer Consulting, Inc. -- http://www.omniti.com/ // Ecelerity: Run with it. _______________________________________________ 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 is almost always guaranteed to fail, as you are trying to map *at* 64KB, and normally you will find your program's text mapped around there. However, the return address is not modulo 64k. How do I accomplish what I want? Well now, what is it, specifically, that you think you want? Why is a virtual address modulo 64K important? (Pointer overloading?) One approach would be to use vm_allocate to allocate an N + 64K sized chunk, round the base address up to the next boundary, free the chunk and then mmap with that address as a hint. Unless you have activity elsewhere (on e.g. another thread) in your app, that's pretty safe). What I am attempting to do it port Solaris' super-scalable slab sllocator (libumem) to Darwin. We have a product that runs ons on a variety of architecture that leverages a custom allocator for performance and uniformity. The Solaris would seems a bit better than ours and we wanted to adopt it while still having a consistent allocator across all platforms. Using the vm_map() directly may be one approach. We already use pthread_mach_thread_np in part of the porting effort. As it will be allocating a tremendous amount of memory and efficient use is a goal, over-allocating and doing alignment ourselves isn't the right approach. This email sent to site_archiver(a)lists.apple.com

1 0

Re: mmap alignment.
by Jim Magee 13 Feb '06

13 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Feb 13, 2006, at 6:52 PM, Chris Sarcone wrote: Theo -- --Jim _______________________________________________ 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'd like to mmap a 64k and have the resulting address be aligned on a 64k boundary. I'm trying with: mmap((void *)65536, 65536, PROT_NONE, MAP_PRIVATE | MAP_ANON | MAP_NORESERVE | MAP_ALIGN, -1, 0); Darwin does not appear to support MAP_ALIGN (A cursory glance at SUS does not reveal any requirement for it - though other OSes do support it). File an enhancement request via http:// bugreporter.apple.com for this feature to be considered for a future OS release. Such an enhancement request should be easy to implement - as the underlying Mach VM supports that kind of alignment via the vm_map() SPI. But unless you are already tied to Mach interfaces for some reason, I don't suggest using vm_map() directly. It's easy enough to do the alignment in your own software as a workaround. This email sent to site_archiver(a)lists.apple.com

1 0

Re: mmap alignment.
by Chris Sarcone 13 Feb '06

13 Feb '06

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: #includes in xcode
by Andrew White 13 Feb '06

13 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com User-agent: Mozilla Thunderbird 0.9 (Macintosh/20041103) As to where these are set? The various target build prefs. -- Andrew White -------------------------------------------------------------------------- This email and any attachments may be confidential. They may contain legally privileged information or copyright material. You should not read, copy, use or disclose them without authorisation. If you are not an intended recipient, please contact us at once by return email and then delete both messages. We do not accept liability in connection with computer virus, data corruption, delay, interruption, unauthorised access or unauthorised amendment. This notice should not be removed. _______________________________________________ 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… Roland Silver wrote: What algorithm does xcode use to find files specified by #import <foo.h> and #import "bar.h" ? If someone can tell me where in the documentation to find the answers to these questions, rather than answering them directly, so much the better. At the basic level, I'm pretty sure they're semantically identical to the #include search paths, which are documented in the gcc documentation. As to what paths XCode adds so that they get searched as intended, if you click the "lines of text" button in the divider bar in the XCode results window you'll get another text field listing the actual gcc commands. From there you can read off the exact search path flags (typically -I and -L). This email sent to site_archiver(a)lists.apple.com

1 0

Re: #includes in xcode
by David Fang 13 Feb '06

13 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com > What algorithm does xcode use to find files specified by > #import <foo.h> > and > #import "bar.h" > ? > If someone can tell me where in the documentation to find the answers > to these questions, rather than answering them directly, so much the If you make an empty file, say foo.c and run: gcc -v -save-temps -c foo.c you'll see a message that tells you something like: #include "..." search starts here: #include <...> search starts here: /usr/include/gcc/darwin/3.3/c++ /usr/include/gcc/darwin/3.3/c++/ppc-darwin /usr/include/gcc/darwin/3.3/c++/backward /usr/local/include /usr/include/gcc/darwin/3.3 /usr/include End of search list. I don't know how to generate this without passing a dummy source file. If you pass more -I(path) arguments on the command line, those will show up in the search list for "...". David Fang _______________________________________________ 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: mmap alignment.
by Steven Bytnar 13 Feb '06

13 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com User-agent: Mutt/1.4i mmap is returning 4KB aligned pages. See xnu/bsd/kern/kern_mman.c On FreeBSD, MAP_ALIGN is also defined to return a page-aligned hunk of memory. How about allocating twice the size (128KB), aligning the returned address to 64KB and internally tracking the difference between the allocated region and the aligned address that you need? Are you sure that you need 64KB alignment? Is 4KB alignment not good enough? --Steve > Hello all, > > I'm porting an app that uses mmap() and banks on its alignment > behaviour on Solaris and Linux. > > I'd like to mmap a 64k and have the resulting address be aligned on a > 64k boundary. I'm trying with: > mmap((void *)65536, 65536, PROT_NONE, MAP_PRIVATE | MAP_ANON | > MAP_NORESERVE | MAP_ALIGN, -1, 0); > > However, the return address is not modulo 64k. How do I accomplish > what I want? > > Thanks! > > // Theo Schlossnagle > // Principal Engineer -- http://www.omniti.com/~jesus/ > // OmniTI Computer Consulting, Inc. -- http://www.omniti.com/ > // Ecelerity: Run with it. > > > > > ------------------------------ > > _______________________________________________ > Darwin-dev mailing list > Darwin-dev(a)lists.apple.com > http://lists.apple.com/mailman/listinfo/darwin-dev > > End of Darwin-dev Digest, Vol 3, Issue 37 > ***************************************** _______________________________________________ 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

#includes in xcode
by Roland Silver 13 Feb '06

13 Feb '06

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… What algorithm does xcode use to find files specified by #import <foo.h> and #import "bar.h" ? If someone can tell me where in the documentation to find the answers to these questions, rather than answering them directly, so much the better. This email sent to site_archiver(a)lists.apple.com

1 0

mmap alignment.
by Theo Schlossnagle 12 Feb '06

12 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Domainkey-signature: q=dns; a=rsa-sha1; c=nofws; s=test; d=omniti.com; h=Received:Mime-Version:Content-Type:Message-Id:Cc:Content-Transfer-Encoding:From:Subject:Date:To:X-Mailer; b=gmSWozvDsaCV9CCh921U3hvyAcCrtsAu/CYHoupjtUQFLnW/WGKLHe14bf67FVWc yBm89TmeDb2c/ASA5LGDpPULKU1C+jr8OgrJi2avQ1wv4K4ARmwZPijoq5c1ZK4p Domainkey-status: good Hello all, Thanks! // Theo Schlossnagle // Principal Engineer -- http://www.omniti.com/~jesus/ // OmniTI Computer Consulting, Inc. -- http://www.omniti.com/ // Ecelerity: Run with it. _______________________________________________ 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 porting an app that uses mmap() and banks on its alignment behaviour on Solaris and Linux. I'd like to mmap a 64k and have the resulting address be aligned on a 64k boundary. I'm trying with: mmap((void *)65536, 65536, PROT_NONE, MAP_PRIVATE | MAP_ANON | MAP_NORESERVE | MAP_ALIGN, -1, 0); However, the return address is not modulo 64k. How do I accomplish what I want? This email sent to site_archiver(a)lists.apple.com

1 0

Re: Reading CTS from a USB serial adaptor
by Dean Reece 11 Feb '06

11 Feb '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Mark, Cheers, - Dean On Feb 11, 2006, at 3:12 PM, Mark Gilbert wrote: Folks. Any suggestions would be welcomed. Thanks _______________________________________________ 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… [Please don't cross-post - I've followed up just to darwin-dev since this isn't really a USB issue.] What driver are you using for this device? The serial family will pass along this info if the driver is tracking it and presenting it. It is up to the driver (and device) to support the flow control lines, and I know several don't bother. I have a special USB serial adaptor using the FTDI RS232 chip. The device has a sensor attached to the CTS line of the chip, and I need to be able to sense the status of the CTS line from my app. Ideally I would be able to setup a blocking call to only return when the status changes. I use Posix to open and configure the port, then I have tried ioctl using TIOCMGET to read the CTS status, but TIOCMGET always returns the same thing (102 decimal) regardless of the value on the CTS line. This denotes CTS is high. Another developer using the same box on Windows is able to read the CTS no problem, so I know the device is OK. Also, I see that in Linux there is a call TIOCMIWAIT which appears to do the blocking sense, but I can't find this in any Mac headers. This would be desirable but not essential. Could anyone suggest anything else to try ? I am not having any luck with the TIOCMGET which should be working according to the Apple docs I have found ('WorkingwithSerial.pdf') Mark Gilbert -- mark(a)gallery.co.uk Tel: +44 208 340 5677 fax: +44 870 055 7790 http://www.gallery.co.uk New Product ! - 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/dean.reece% 40apple.com This email sent to dean.reece(a)apple.com This email sent to site_archiver(a)lists.apple.com

1 0