January 2009 - Darwin-dev - lists.apple.com

Re: mach-o: why is code placed towards the end of a page?
by Andrew Myrick 09 Jan '09

09 Jan '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Joel, Is this layout preventing you from doing something? -Andrew _______________________________________________ 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 neglected to reply to the list. Here is my complete reply for context: The linker always places code on a power-of-2 boundary towards the end of the page. Is there a reason for this? The Mach-O headers are indeed at the beginning of the file, but when they are mapped into a program's address space, those headers are placed at the start of the __TEXT segment, and the actual text follows those headers. When the text is smaller than the available remaining space in the first page, it is placed at the end of the page as you have noticed. When the text is larger than the remaining space, it will begin at the end of the headers and continue on into the next page. I can't say for certain why it is implemented this way, but I can speculate that it is so that the text is adjacent to the __DATA segment on the following page. Can the code be placed at the beginning of the page instead? The headers are always placed at the beginning of the __TEXT segment, so the actual code will never be at the first byte of that segment, regardless of whether the text is placed at the end of the headers or the end of the page. This email sent to site_archiver(a)lists.apple.com

1 0

Re: mach-o: why is code placed towards the end of a page?
by Joel Reymont 09 Jan '09

09 Jan '09

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:cc:message-id:from:to :in-reply-to:content-type:content-transfer-encoding:mime-version :subject:date:references:x-mailer; bh=xSTQpwRu/3WiyN+B+XJhAQVUovDnuz6farI/UcSSq78=; b=M7Vb/XNvcHC+stR9KUbKUNM/X73PC5eavCX2axOcVTHRZLUKcTEH7mL6L0xo5WU2eO p/Jbw+IbL3fQsiwwv/xTncerZcdVvGz4JC2js7hp4r1E8/kfCQn1B89vUqO/U4wil75w PgPevurHhkoDjWxTHZXqa31qyPLrD6jAaAHQQ= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=cc:message-id:from:to:in-reply-to:content-type :content-transfer-encoding:mime-version:subject:date:references :x-mailer; b=EmLR71p9H2VFLZOVRpinalyDr+92XhvgC6pjdH6hVxHMkBwPAy6UY3Jmk42dR3XwHe zvSS5Y8w6jQseDtiD/bHyFn5fAQq1hNseNrsmXjraWCziGWOFDHZjFAPkrT6LRQ2PzsR pIsMEknwbMweFRZQ0peOEt1xNybpDdfpFBPqM= On Jan 8, 2009, at 11:56 PM, Andrew Myrick wrote: Thanks, Joel --- http://twitter.com/wagerlabs _______________________________________________ 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… The headers are always placed at the beginning of the __TEXT segment, so the actual code will never be at the first byte of that segment, regardless of whether the text is placed at the end of the headers or the end of the page. I apologize, I meant at the beginning of a new page. I'm placing the code right after headers, aligned on a 16-byte boundary (just in case). Is this layout preventing you from doing something? No, I don't think so. I was just wondering if pushing text towards the end of the page was a must. I see that it's not. This email sent to site_archiver(a)lists.apple.com

1 0

Re: mach-o: why is code placed towards the end of a page?
by Joel Reymont 08 Jan '09

08 Jan '09

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:cc:message-id:from:to :in-reply-to:content-type:content-transfer-encoding:mime-version :subject:date:references:x-mailer; bh=4k1YpDDDEdF1UnXa9cs2TuFDTGOvGqUNY2ARKhxndDI=; b=ZpYtcZYzvP04q/jw4os6ZZpdVga8GN0UdduDRZUswq25z0k9oDwb/+yrzcVBwCRir5 XosowZF/sul2VzNb6B2qGAT1hnxvKpPwPz8hfZUbDWEBv6L3xSdaS2Be/grdLl7mBKay 7nhe6ouIXVETpr04hs54erRq0m6AOv/+7JyCc= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=cc:message-id:from:to:in-reply-to:content-type :content-transfer-encoding:mime-version:subject:date:references :x-mailer; b=ZV2EncaACTnqxe4JDDLrbzKk0SuR/09YvPjfWc9LNzyJ49DxnntcbeZdeVR/8yHX3A gs+k1yPMa8/nWJZ6uNsqSxWyKPKXdWDRoRUGwudchybN1S/KaOPhCQdrJ3GpeerVeCeT KN7y4OXWVQJBGx2AxT3ifqkq6fuRh6Y3qRQmI= Ed, On Jan 8, 2009, at 10:31 PM, Ed Wynne wrote: The page is 4096 bytes but the size of all commands is just 772. According to the text section, sectname __text segname __TEXT addr 0x00001fc8 size 0x00000035 offset 4040 0x1fc8 = 8136, 0x35 = 53. 4096 - 4040 = 56, our code size (0x35, 53) rounded to a multiple of 4. Load commands 2, 3, etc. go after the text section, e.g. 4096 below. Load command 2 cmd LC_SEGMENT cmdsize 124 segname __DATA vmaddr 0x00002000 vmsize 0x00001000 fileoff 4096 I ran the test on a few more executables with similar results. --- http://twitter.com/wagerlabs _______________________________________________ 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… It should not be intentionally towards the end of the page. If you look closer, you'll see that the mach-o header, as well as any additional space reserved for library names, comes before the code in the TEXT segment. My understanding is that the Mach-O header always goes at the beginning of the file. According to the docs, the load commands go into page #1. Mach header magic cputype cpusubtype caps filetype ncmds sizeofcmds flags 0xfeedface 7 3 0x00 2 11 772 0x00000085 8136 - 4096 (pagezero) = 4040, exactly our offset above. This is fine, we are on the 2nd page in memory, right after PAGEZERO. This proves that the linker places the first page of code towards the end of the page. Would you agree? This email sent to site_archiver(a)lists.apple.com

1 0

mach-o: why is code placed towards the end of a page?
by Joel Reymont 08 Jan '09

08 Jan '09

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:message-id:from:to :content-type:content-transfer-encoding:mime-version:subject:date :x-mailer; bh=ECalL8DTWPJbECc9nuPZ8lV1aCkwDyis2mMUFbekb7Q=; b=VJZU4Z+Y4SV0/lk2QaSF2B61UtFAz+gK9hxLlPshKSix9KkTqAUX0maOT23c610m0d 91uY/Ier78IytWgrZLN/fSuEaZ5Qxzq1liT90OEADpwn65go380W40Hz/AV4myepEofN hrU+/R15aq5iCVxdAe5D0y2cxxxhdEd7Ujm3E= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=message-id:from:to:content-type:content-transfer-encoding :mime-version:subject:date:x-mailer; b=XfLVfBJVjqBDV5EhqdPN97VkdA7/c+O3CdhIm4cERjoQkMzUKCCmARifGqSbusPkab hFP8YlqqOuWYwkPPHaYTVUyxdAOl0zZBGpIBtb05X2ldNZuG6YyqtugqEdy1diLTA2aJ qsKt7sqn36KfMU5Hw57BL+WVlP1mWKRH/tF6Q= Can the code be placed at the beginning of the page instead? Thanks, Joel --- sectname __text segname __TEXT addr 0x00001fc8 size 0x00000035 offset 4040 align 2^0 (1) reloff 0 nreloc 0 flags 0x80000400 reserved1 0 reserved2 0 --- http://twitter.com/wagerlabs _______________________________________________ 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… The linker always places code on a power-of-2 boundary towards the end of the page. Is there a reason for this? 0x1fc8 below is 8136, exactly 4040 if you subtract the __PAGEZERO that goes before it. This email sent to site_archiver(a)lists.apple.com

1 0

Re: Get BSDName from drive name
by Terry Lambert 08 Jan '09

08 Jan '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Hi all, I have very small query. I want to get the BSDName of mass storage device like perndrive. Ex. Name of pendrive is MyPenDrive and path is /Volumes/MyPenDrive and now I want to get the BSDName (ex. disk1, disk2, disk1s1 etc) for it. getmntinfo(3) -- Terry _______________________________________________ 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 Jan 8, 2009, at 2:17 AM, Rakesh Singhal wrote: This email sent to site_archiver(a)lists.apple.com

1 0

Get BSDName from drive name
by Rakesh Singhal 08 Jan '09

08 Jan '09

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:message-id:date:from:to :subject:mime-version:content-type:content-transfer-encoding :content-disposition; bh=unOxT6ngOWM7UFXWcEBf/tZXXMsVYDAtAuStKHwiM7s=; b=uOCnS4uBCqQZF9mbYFLIC4zWkXTSUnbS51Cqay+BVLek/AB4mHQjXTiMNEWiqaPSLE xHs5RKb2qdECI80qYYDDNLRj62ZOZLmyrFNuhWs/9BzHKMdvQLQkgim4po9A5P27338w jJwV/eFE00xQ5mObvRpMgiIbzXIjTe3NwrhpE= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=message-id:date:from:to:subject:mime-version:content-type :content-transfer-encoding:content-disposition; b=jDtjwVL1LTmnlTcj/5my9bOAKVGIFe6YZl86E8QoyQk+I+4WuJSHQtftakjzehf50E rUfeTiA9xvgThz9nE9UICQHtQGabF16ipomMY11cNaL/sH/tzdHvl+XR0srsircIgtv8 UWRZNMw/0bjajdYEETv5AsBDFxWrJhrcx3/fk= Hi all, I have very small query. I want to get the BSDName of mass storage device like perndrive. Ex. Name of pendrive is MyPenDrive and path is /Volumes/MyPenDrive and now I want to get the BSDName (ex. disk1, disk2, disk1s1 etc) for it. Thanks in advance. Regards rksinghal _______________________________________________ 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: Has anything changed in 10.5.6 bootstrap contexts management?
by Kevin Van Vechten 08 Jan '09

08 Jan '09

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com - Kevin On Jan 7, 2009, at 12:06 AM, Stéphane Letz wrote: Hi, 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… No, this is not a change. Daemons running in the system bootstrap context do not have visibility of registrations in the user bootstrap context on any version of Mac OS X. Instead, I would recommend the user agent connect to the system daemon. With OSX 10.5.6, our audio user land driver cannot be activated as default input/output device anymore. It seems that the "coreaudiod" deamon (that runs under root) cannot connect anymore to our audio server (jackd) that defines a mach port entry point in the user account bootstrap context. Is the any fundamental change in bootstrap contexts management in the system? Stephane Letz This email sent to site_archiver(a)lists.apple.com

1 0

Re: Apple public CVS?
by Kevin Van Vechten 08 Jan '09

08 Jan '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com That's correct. On Jan 7, 2009, at 7:33 PM, Benjamin Huntsman wrote: $ export CVSROOT=:ext:apsl@anoncvs.opensource.apple.com:/cvs/apsl appears to still be present, but doesn't allow logins... _______________________________________________ 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… Does anyone here know, is there anything left on anoncvs.opensource.apple.com? Per the instructions on http://developer.apple.com/opensource/ cvs.html, I can use the following: but it would seem the only projects left there are StreamingServer and mDNSResponder. export CVSROOT=:pserver:<username>@anoncvs.opensource.apple.com:/cvs/ Darwin CVS versions of xnu (or Apple-*) are no longer available anywhere except in .tar.gz form, correct? Thanks in advance! This email sent to site_archiver(a)lists.apple.com

1 0

Apple public CVS?
by Benjamin Huntsman 08 Jan '09

08 Jan '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Thread-index: AclxQdGsbEwg5Eg/TaONJyZqpSHaCw== Thread-topic: Apple public CVS? Does anyone here know, is there anything left on anoncvs.opensource.apple.com? Per the instructions on http://developer.apple.com/opensource/cvs.html, I can use the following: $ export CVSROOT=:ext:apsl@anoncvs.opensource.apple.com:/cvs/apsl but it would seem the only projects left there are StreamingServer and mDNSResponder. export CVSROOT=:pserver:<username>@anoncvs.opensource.apple.com:/cvs/Darwin appears to still be present, but doesn't allow logins... CVS versions of xnu (or Apple-*) are no longer available anywhere except in .tar.gz form, correct? Thanks in advance! _______________________________________________ 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

Apple public CVS?
by Benjamin Huntsman 08 Jan '09

08 Jan '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Thread-index: AclxQZab03iFiHdoRzSaT0LcSVOJbA== Thread-topic: Apple public CVS? Does anyone here know, is there anything left on anoncvs.opensource.apple.com? Per the instructions on http://developer.apple.com/opensource/cvs.html, I can use the following: $ export CVSROOT=:ext:apsl@anoncvs.opensource.apple.com:/cvs/apsl but it would seem the only projects left there are StreamingServer and mDNSResponder. export CVSROOT=:pserver:<username>@anoncvs.opensource.apple.com:/cvs/Darwin appears to still be present, but doesn't allow logins... CVS versions of xnu (or Apple-*) are no longer available anywhere except in .tar.gz form, correct? Thanks in advance! _______________________________________________ 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