March 2007 - Darwin-kernel - lists.apple.com

Re: large, disposable files?
by Andrew Gallatin 13 Mar '07

13 Mar '07

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Andrew Gallatin writes: > I doesn't look like Darwin/MacOSX has an O_DIRECT option to disable > caching on the file at open time. Does MacOSX/Darwin have a way to > disabling caching on a per-mountpoint basis? > I got two, independant suggestions to try fcntl(fd, F_NOCACHE, 1). Execellent! But before I go trying to re-compile these applications, I'm still wondering if there is any way to limit the ubc like in Tru64... Drew _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Beginning with Kernel Extension Concepts
by Michael Crawford 13 Mar '07

13 Mar '07

site_archiver(a)lists.apple.com Delivered-To: Darwin-kernel(a)lists.apple.com Dkim-signature: a=rsa-sha1; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=cWZ/ck2h2EWSDXcp9Y36dIsgHQkBulmOAAWuaAupQVo4zEpS0cKXCHRv0FwUz4NqUg0u3Xjpvlg60cvhrWJM2cn7yM3bHKmLx4tay6CIyuwuLlxgbYOv6w+eL3ooKGLIMZox4QvFJAYs5XwMAaXyxQnjQfhwEk23wA7kVjldaJQ= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=DA5uv3MilMFrVB/c/MRqad0aZalwrdwwW1o/BzO6c6aXuu85cRe5Zkve8dFJQEhaDjSxqNXNAISTe5og0/faZQ3dNvov6mTYF6kaqiG46VVytQIwueM03zoS0zqYXk3RovQhGXwDNyaY2HCQlcSsbhLH0jz2Zm1JPOJvtOD5s0s= Could you be compiling the code as C and not C++? extern "C" is legal in C++, but not C. You definitely want to compile as C++ for I/O Kit drivers. Mike Crawford mdcrawford at gmail dot com Bonjour à tous, >From Apple Dev Documentation, I get an error : "expected unqualified-id before string constant" - on line 5 in the HelloIOKit.cpp file of this "Hello I/O Kit : Creating a Device Driver With XCode" chapter. I stare at it and can't think anything brilliant about it. First, I am not sure about where the error raised from : the call to a "C " construct, or the line before ? //--------------------- /* add your code here */ #include <IOKit/IOLib.h> #include "HelloIOKit.h" extern "C" { #include <pexpert/pexpert.h> } // For debugging purpose only [...] //------------------- Thanks. Francis. _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/mdcrawford%40gmail.com This email sent to mdcrawford(a)gmail.com -- Michael David Crawford mdcrawford at gmail dot com Enjoy my art, photography, music and writing at http://www.geometricvisions.com/ --- Free Music Downloads --- _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… On 3/13/07, Francis Derive <francisderive(a)mac.com> wrote: This email sent to site_archiver(a)lists.apple.com

1 0

Beginning with Kernel Extension Concepts
by Francis Derive 13 Mar '07

13 Mar '07

site_archiver(a)lists.apple.com Delivered-To: Darwin-kernel(a)lists.apple.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com

1 0

Re: large, disposable files?
by Arvind Dalvi 13 Mar '07

13 Mar '07

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Dkim-signature: a=rsa-sha1; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=VIy15gu7SqSEcmKUdhCdfKp8ruNx8rPyvI8wB1HjpEmw6YWiy24x5adH1colRXk+0Rn2p7M7NJpCtn622Rqar1OS/gEFbMAzCl8Nio9yV+pIOPu0IPdBim3GpXTJ999cJYWdCJEJqlBXCSqaoAoUrcbYxN3z39QExZHYEMAjRMw= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=bfVMnAR5PDdF6sQucnrEJRaWHqFQTYKhFylwRjddBi4JT/6OZofnBMZWnazcbx345ZnAzgUGEjNehaU7zF/9SVAE4/6rbDs06JMsAD39M+C/7CytJLex0dOTkWOp4ySgur26p6N4D81CQXbco9mXvn1pxEr6TBoida9s48TWRK8= Hello, I am not to sure my experience relates to the same problem. When i was trying to copy 400MB+ .iso from a USB stick to my Mac drive. The copy operation was crawling. I tried by rebooting the system, and then copying over again, but no luck. Thanks, a.d. _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com

1 0

large, disposable files?
by Andrew Gallatin 13 Mar '07

13 Mar '07

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com I'm dealing with applications which read large (many GB, much larger than physical memory) files at a moderate rate of speed (a few MB/sec). They generally make one pass through the file and delete it. There is no benefit at all to caching the file in the kernel. The caching which is done is extremely detrimental as it causes all kinds of unnecessary paging; the OS pages out other applications which happen to be idle at the time in favor of caching the files which will never be read again. I doesn't look like Darwin/MacOSX has an O_DIRECT option to disable caching on the file at open time. Does MacOSX/Darwin have a way to disabling caching on a per-mountpoint basis? Failing that, the last Mach based OS I used (DEC OSF/1) had all kinds of nice tuning parameters to prevent the unified buffer cache from taking over too much physical memory. See, for example, the description of ubc-maxpercent at http://h30097.www3.hp.com/docs/base_doc/DOCUMENTATION/HTML/AA-Q0R3E-TET1_ht… Is there some sort of option like ubc-maxpercent in Darwin/MacOSX that I'm not seeing? Thanks, Drew _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Equivalent of Linux kmap on Mac OS
by Terry Lambert 10 Mar '07

10 Mar '07

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Now that the deal's not off, please, I'm eagerly awaiting a suggestion that would really work in my scenario. I've been following this e-mail thread, and so far I think only you've truly understood what I would like to do :) I'm pretty darn sure I understand; I just don't approve of the idea. -- Terry _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… On Mar 9, 2007, at 9:56 AM, Bhavesh Davda wrote: Otherwise the deal's off. No more helping. This email sent to site_archiver(a)lists.apple.com

1 0

Re: Network Interface for a Given Destination Address
by Josh Graessley 09 Mar '07

09 Mar '07

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Mar 9, 2007, at 12:57 AM, Jones Curtis wrote: On Mar 9, 2007, at 03.43.27, Michael Smith wrote: Thanks for the reply. -josh _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com It's my understanding that advisory forms of the kernel's routing policy decisions are not exported to userspace. There are good reasons for this; not the least of which being that between obtaining such advice and actually committing to a decision based on that advice, the validity of the advice may expire due to changes in the system's topology. I realise this doesn't help you much, but telling the future is a tough business. 8( The conditions under which the system's topology might change (DHCP server says to, user unplugs ethernet cable, cordless phone/ microwave interferes with wireless network and causes loss of connectivity, etc.) are all possibilities I'll have to live with. Knowing the future would certainly be helpful, but I'm content with knowing how it would proceed if it were to happen presently. Is there no KPI interface for asking the routing table for the gateway ip address for a given destination address? I think you want a routing socket (AF_ROUTE). I'm not familiar with the details of how routing sockets work, but I'll bet there is some good code in the route command. Check to see how "route get" works. smime.p7s

1 0

Re: Equivalent of Linux kmap on Mac OS
by Bhavesh Davda 09 Mar '07

09 Mar '07

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Dkim-signature: a=rsa-sha1; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=q/aheEDBtFb//JtqdZkNhxAbFMZJmkbhOdDF7NRW9lqlFyf45uc7ioTXhDFgNPB7hPKwBKrd7J8uQNtBuRJ6J1ipj055d3eG3aHWKtt3Vu0tmrZgKs+my+CqrwifktT+GIy/09vsNjpGtTIZRxvjjuzkEuCzzhbmb33hl764eyQ= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=X/sB8CJpkQ99tpQVYVchxEoDHMINwBoUMo43NS9732PdHmzzK4NRWrQFNQW2iBcjArtIa+I4ueuttGaiEKNYzX8HRiBUCIJnPsucfTfsqi2vNE8fMPOgkgdhNb0B8M4fPHBz/cuFr1/EX6qPDWNWWAuvlY1xKqk772W3pZLQeGw= Whoa, hold it right there. Before we help you any more, I want a promise out of you. This chunk of memory is *small*, right? A couple of pages at most, right? And you set this up once and then keep using it for a long time, right? Right. Right. Right. That's why I don't want to get into the business of map/unmap in the kext every time I want to access this chunk of memory. I also don't want to use any reader/writer accessors to access the memory without mapping it, because I plan on using atomic operations on the chunk of memory. Now that the deal's not off, please, I'm eagerly awaiting a suggestion that would really work in my scenario. I've been following this e-mail thread, and so far I think only you've truly understood what I would like to do :) The kernel is *not* in the business of handing out chunks of its virtual address space to arbitrary processes. Understood. As Terry suggests, you are better off allocating the space in the kernel, and then vending mappings of that space to user tasks as they come and go. I understand, and I wish I could do that. But I'm dealing with a special requirement here. I'm still curious as to why you think you want to share memory like this; there are better ways to do most of the common things that I can think of. My lips are sealed :) = Mike -- Bhavesh P. Davda _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… Otherwise the deal's off. No more helping. This email sent to site_archiver(a)lists.apple.com

1 0

Re: Network Interface for a Given Destination Address
by Jones Curtis 09 Mar '07

09 Mar '07

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Dkim-signature: a=rsa-sha1; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:in-reply-to:references:mime-version:content-type:message-id:cc:from:subject:date:to:x-mailer; b=sCbaZPULQJ8dOHnh6d07b0YD6APuUm01ZsCzqQVOczwmYpsieW4YptoaBEa1DWL6frVgs6YCqEH+qMolhSoACjkAR7a8mdkefki6CaZLqMPsrO0w+2lJK+TLWPw7Y6Vvl1WXC7nna/s5QE+le/WMIwUkkSt9CdQ7al/uIM9swJ0= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:in-reply-to:references:mime-version:content-type:message-id:cc:from:subject:date:to:x-mailer; b=mvCha3qyNGxUP2NzLmUwDbf1CmBedk849j4hLh0Ddbi2wMTWK8oLQdbU7mp1Axw/IcUno6Qi6wgtMX4OUa6Nu1m2dC96J5STvPOBvvOweoMdLOdw07gcWNxjqo89zawHAhNcjgqCAv+nAfdoqcZ+K6zzgbp7cHclZ4Yxwi7XNDw= On Mar 9, 2007, at 03.43.27, Michael Smith wrote: Thanks for the reply. -- Curtis Jones curtis.jones(a)gmail.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com It's my understanding that advisory forms of the kernel's routing policy decisions are not exported to userspace. There are good reasons for this; not the least of which being that between obtaining such advice and actually committing to a decision based on that advice, the validity of the advice may expire due to changes in the system's topology. I realise this doesn't help you much, but telling the future is a tough business. 8( The conditions under which the system's topology might change (DHCP server says to, user unplugs ethernet cable, cordless phone/microwave interferes with wireless network and causes loss of connectivity, etc.) are all possibilities I'll have to live with. Knowing the future would certainly be helpful, but I'm content with knowing how it would proceed if it were to happen presently. Is there no KPI interface for asking the routing table for the gateway ip address for a given destination address? smime.p7s

1 0

Re: Network Interface for a Given Destination Address
by Michael Smith 09 Mar '07

09 Mar '07

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Mar 9, 2007, at 12:32 AM, Jones Curtis wrote: = Mike _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… I'm trying to determine the ethernet interface (en0, en1, etc.) that will be used for a given destination ip address. I've tried using ifaddr_withdstaddr() and ifaddr_withnet() and I can get the correct result for addresses that are local to the machine and for those that are on the same subnet as one of the ethernet interfaces for the machine. So, that's part of the solution, but not all of it. I looked at ifaddr_findbestforaddr(), thinking that it might return NULL if that the given interface wasn't useful for that address, but if I'm reading the source code for ifaof_ifpforaddr() correctly, it doesn't touch the routing tables and it simply returns the first address entry if none of them match. So, I'm guessing I need to specifically find the interface for the default route if _withdstaddr() / _withnet() don't help. Is this the correct approach? And how would I go about doing that? It's my understanding that advisory forms of the kernel's routing policy decisions are not exported to userspace. There are good reasons for this; not the least of which being that between obtaining such advice and actually committing to a decision based on that advice, the validity of the advice may expire due to changes in the system's topology. I realise this doesn't help you much, but telling the future is a tough business. 8( This email sent to site_archiver(a)lists.apple.com

1 0