September 2002 - Darwin-kernel

Re: What should we do when headers are missing in 10.2?
by Justin C. Walker 11 Sep '02

11 Sep '02

On Wednesday, September 11, 2002, at 09:38 AM, Stiphane Sudre wrote: On Wednesday, September 11, 2002, at 06:19 PM, Justin C. Walker wrote: On Wednesday, September 11, 2002, at 03:42 AM, Stiphane Sudre wrote: In 10.2, there are some headers that vanished. Yeah; it's part of Apple's plan to deal with 'changing APIs'. Jim's message describes the "private API" define, but doesn't mention that you really are missing the files. They should be in the Kernel Framework, though, so you can either reference them there (-I...) or copy them over. They are not :-( Yeah; I just saw another message with that info. Not sure what's up, since Allan's message indicates that they should be there (I assume DevSDK is not a previous install...). What does 'lsbom' on your system show? Regards, Justin -- Justin C. Walker, Curmudgeon-At-Large * Institute for General Semantics | When LuteFisk is outlawed | Only outlaws will have | LuteFisk *--------------------------------------*-------------------------------* _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: What should we do when headers are missing in 10.2?
by Andrew Gallatin 11 Sep '02

11 Sep '02

Jim Magee writes: > > On Wednesday, September 11, 2002, at 11:49 AM, Andrew Gallatin wrote: > > I could make all uses APPLE_API_PRIVATE headers go away in my driver > > (OS-Bypass cluster interconnect -- > > http://www.myri.com/scs/macosx.html) if somebody would just fix the > > VM system so that it wouldn't block a user process indefinatly if said > > process attempts to free memory locked via IOMemoryDescriptor's > > prepare function. But I couldn't seem to get anybody interested in > > changing this behaviour in the past... > > I think I even exchanged some mail with you on this subject last time. Yes, I think you did. To refresh, the problem happens when an app has registered memory for os-bypass IO via an ioctl into our driver. The memory is locked via prepare. In the normal course of operations, a task will deregister the memory (and our driver will unprepare it, thus unlocking it) prior to freeing it. The problem occurs when a buggy applications neglects to deregister the memory prior to freeing it. This is a bug in the application (they've violated our API). But the resulting behaviour (process is unkillable, sleeping waiting for memory to be unprepared; our driver won't unprepare the memory because it has no idea that it should) is sub optimal. > Doing what you ask would require "prepare" to create and hold its own > reference to the underlying memory objects rather than "glom-ing" on to > the user's reference (leaving the user free to release his own with > impunity). This could be done, but it would slow prepare down even > more in its current form (and it is in dire need of a performance boost > already). In fact, we are in the process of investigating possible > enhancements to the implementation. The leading design does involve > burrowing all the way down to the object, so we do have the opportunity > to think about taking our own references again. This would allow > unprepare to be much more efficient, and subsequent prepares in the > same object could also be quicker. But is also allows a possible > redress of your issue. Excellent! Also, memory locking on MacOSX is currently slower than on any other unix-like platform we support. It would be good to see it sped up. > The real question is: Does anything depend upon that portion of the > address space being locked down while I/O is in progress? It would be > difficult to see how they could, because the prepare gets released > before their notion of an I/O operation is marked as complete. But > it's just something to investigate. Nothing depends on that part of the address space being locked down. However, the physical memory (formerly) backing the portion of the apps virtual address space which it has just freed cannot be recycled. Otherwise we risk either corrupting memory if those pages end up belonging to some other task. Ideally, the memory would be removed from the app's address space when it called free(), and the physical pages would held in limbo, only freed back to the system when we unprepared it and some reference count dropped to zero. This is what our linux driver does (by hand, in our driver, since there is no api for it. 300+ lines of code which change frequently. Ick). Actually, I suppose it could be freed if you had a sophisticated IOMMU and could remap those addresses to the bit bucket when they were accessed via dma. ;) > --Jim > > PS: I think, last we left it, we also talked about changing the timing > between file-descriptor closing and VM cleaning. We intentionally > clean up the Mach port space (which IOKit UserClient interfaces > typically use) before the VM space so that UserClient memory > descriptors would be freed before attempting to free the memory in an > exiting process. Closing file descriptors should happen even before > that (to allow file-descriptor-based drivers to release their holds on > memory). While this may not be necessary, if we do the work above, it > is still the cleaner way to go. This like what I was asking for... Has the behaviour changed? I haven't looked into this since I wrote my reaper thread work around, which uses APPLE_API_PRIVATE apis. The problem I was running into earlier was due to the process's VM space being cleaned up _prior_ to its file-descriptors being closed. The effect would be similar to what I talk about above -- the VM cleanup code frees registered memory, leading to a deadlock. Closing the file-descriptors first avoids the deadlock, because our driver deregisters the app's memory at this point. Thanks! Drew _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: What should we do when headers are missing in 10.2?
by Vincent Lubet 11 Sep '02

11 Sep '02

We always recommend that you file a Radar bug report anytime you discover a problem with Darwin or Mac OS X. Vincent On Wednesday, September 11, 2002, at 03:42 AM, Stiphane Sudre wrote: In 10.2, there are some headers that vanished. This leads to the impossibility to compile some code which was compiling fine on 10.1. A (maybe bad) example is this kind of compilation error: /System/Library/Frameworks/Kernel.framework/Headers/netinet/ tcpip.h:66: field `ti_i' has incomplete type What is the recommended solution in this situation? Is there a darwin patch we can apply to get back these headers in a clean way? _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: What should we do when headers are missing in 10.2?
by Jim Magee 11 Sep '02

11 Sep '02

On Wednesday, September 11, 2002, at 12:21 PM, Brian Bergstrand wrote: On Wednesday, September 11, 2002, at 10:30 AM, Jim Magee wrote: I remember there was a problem with them being missing in earlier seeds of the developer tools, but I thought for sure that was addressed. I'm sure someone more knowledgeable than I will chime in with exactly which versions of things have the fix. Wasn't fixed. And there is an open bug on the problem. I'm surprised the August 2002 tools did not fix the problem. I'm sure this problem is well known to everyone but me. But to be sure, I'll take my foot out of my mouth and make sure. ;-( In the meantime: To fix this, I checked out xnu Apple-344 from CVS and copied the headers into Kernel.framework. Which is the best workaround I can think of as well. --Jim _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: What should we do when headers are missing in 10.2?
by Stéphane Sudre 11 Sep '02

11 Sep '02

On Wednesday, September 11, 2002, at 06:19 PM, Justin C. Walker wrote: On Wednesday, September 11, 2002, at 03:42 AM, Stiphane Sudre wrote: In 10.2, there are some headers that vanished. Yeah; it's part of Apple's plan to deal with 'changing APIs'. Jim's message describes the "private API" define, but doesn't mention that you really are missing the files. They should be in the Kernel Framework, though, so you can either reference them there (-I...) or copy them over. They are not :-( _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: What should we do when headers are missing in 10.2?
by Allan Nathanson 11 Sep '02

11 Sep '02

On Wednesday, September 11, 2002, at 12:05 PM, Michael Cashwell wrote: On Wednesday, September 11, 2002, at 11:30 AM, Jim Magee wrote: It should be in /usr/include/netinet/ip_var.h. It should also be in /System/Library/Frameworks/Kernel.framework/Headers/netinet/ip_var.h. That's because you're building a kernel extension which cannot use the headers in /usr/include. ... Disturbing. I did a completely clean install of the retail 10.2, its dev tools and the August patch to the dev tools and searching all local disks yields no files with "ip_var" in the name. If anyone learns what's up with this please let us know. Thanks. g3> lsbom -s /Library/Receipts/DevSDK.pkg/Contents/Resources/DevSDK.bom | grep ip_var ./System/Library/Frameworks/Kernel.framework/Versions/A/Headers/ netinet/ip_var.h ./usr/include/netinet/ip_var.h _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: What should we do when headers are missing in 10.2?
by Jim Magee 11 Sep '02

11 Sep '02

On Wednesday, September 11, 2002, at 11:49 AM, Andrew Gallatin wrote: I could make all uses APPLE_API_PRIVATE headers go away in my driver (OS-Bypass cluster interconnect -- http://www.myri.com/scs/macosx.html) if somebody would just fix the VM system so that it wouldn't block a user process indefinatly if said process attempts to free memory locked via IOMemoryDescriptor's prepare function. But I couldn't seem to get anybody interested in changing this behaviour in the past... I think I even exchanged some mail with you on this subject last time. Doing what you ask would require "prepare" to create and hold its own reference to the underlying memory objects rather than "glom-ing" on to the user's reference (leaving the user free to release his own with impunity). This could be done, but it would slow prepare down even more in its current form (and it is in dire need of a performance boost already). In fact, we are in the process of investigating possible enhancements to the implementation. The leading design does involve burrowing all the way down to the object, so we do have the opportunity to think about taking our own references again. This would allow unprepare to be much more efficient, and subsequent prepares in the same object could also be quicker. But is also allows a possible redress of your issue. The real question is: Does anything depend upon that portion of the address space being locked down while I/O is in progress? It would be difficult to see how they could, because the prepare gets released before their notion of an I/O operation is marked as complete. But it's just something to investigate. --Jim PS: I think, last we left it, we also talked about changing the timing between file-descriptor closing and VM cleaning. We intentionally clean up the Mach port space (which IOKit UserClient interfaces typically use) before the VM space so that UserClient memory descriptors would be freed before attempting to free the memory in an exiting process. Closing file descriptors should happen even before that (to allow file-descriptor-based drivers to release their holds on memory). While this may not be necessary, if we do the work above, it is still the cleaner way to go. _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: What should we do when headers are missing in 10.2?
by Brian Bergstrand 11 Sep '02

11 Sep '02

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Wednesday, September 11, 2002, at 10:30 AM, Jim Magee wrote: On Wednesday, September 11, 2002, at 11:09 AM, Stiphane Sudre wrote: In fact, my problem is not with the need to #include any file before or after. The real issue is that the files are missing from /usr/include in a complete Mac OS X 10.2 install (by complete, I mean everything checked, last dev Tools, etc...). Currently, I'm using the header files from a 10.1 backup but it's far from being a good solution. So I'm asking for a better/recommended/official solution. It should be in /usr/include/netinet/ip_var.h. It should also be in /System/Library/Frameworks/Kernel.framework/Headers/netinet/ip_var.h. That's because you're building a kernel extension which cannot use the headers in /usr/include. My 10.2 system, installed with only the contents of the released CDs on a new volume, seems to have it. I remember there was a problem with them being missing in earlier seeds of the developer tools, but I thought for sure that was addressed. I'm sure someone more knowledgeable than I will chime in with exactly which versions of things have the fix. Wasn't fixed. And there is an open bug on the problem. I'm surprised the August 2002 tools did not fix the problem. Anyway, these are the missing headers that I have found so far in Kernel.framework. sys/unistd.h netinet/in_var.h netinet6/in6_var.h sys/signalvar.h ufs/ netinet/ip_var.h netinet/tcp_var.h netkey/key_var.h These missing headers break Kernel.framework as well as third party code. To fix this, I checked out xnu Apple-344 from CVS and copied the headers into Kernel.framework. Brian Bergstrand Software Engineer, Prosoft Engineering <http://www.prosofteng.com/> There are no stupid questions, but there are a lot of inquisitive idiots. - ThinkGeek.com Poster -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.7 (Darwin) iD8DBQE9f213edHYW7bHtqIRAqcIAJ0ZD1di+vFHZDmwUkTmZjRaAqlVigCgxnk+ V6CzeRernjvM6+pKLvHLInk= =HHbz -----END PGP SIGNATURE----- _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: What should we do when headers are missing in 10.2?
by Justin C. Walker 11 Sep '02

11 Sep '02

On Wednesday, September 11, 2002, at 03:42 AM, Stiphane Sudre wrote: In 10.2, there are some headers that vanished. Yeah; it's part of Apple's plan to deal with 'changing APIs'. Jim's message describes the "private API" define, but doesn't mention that you really are missing the files. They should be in the Kernel Framework, though, so you can either reference them there (-I...) or copy them over. Regards, Justin -- Justin C. Walker, Curmudgeon-At-Large * Institute for General Semantics | Men are from Earth. | Women are from Earth. | Deal with it. *--------------------------------------*-------------------------------* _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: What should we do when headers are missing in 10.2?
by Stéphane Sudre 11 Sep '02

11 Sep '02

On Wednesday, September 11, 2002, at 05:30 PM, Jim Magee wrote: On Wednesday, September 11, 2002, at 11:09 AM, Stiphane Sudre wrote: In fact, my problem is not with the need to #include any file before or after. The real issue is that the files are missing from /usr/include in a complete Mac OS X 10.2 install (by complete, I mean everything checked, last dev Tools, etc...). Currently, I'm using the header files from a 10.1 backup but it's far from being a good solution. So I'm asking for a better/recommended/official solution. It should be in /usr/include/netinet/ip_var.h. It should also be in /System/Library/Frameworks/Kernel.framework/Headers/netinet/ip_var.h. That's because you're building a kernel extension which cannot use the headers in /usr/include. Hum, it's not on mine. It's not on another machine with 10.2 This the content of /usr/include/netinet Last login: Wed Sep 11 17:50:55 on ttyp2 Welcome to Darwin! [coyotte:~] stephane% cd /usr/include/netinet [coyotte:/usr/include/netinet] stephane% ls bootp.h in.h ip6.h tcp.h tcp_timer.h icmp6.h in_pcb.h ip_fw.h tcp_debug.h tcpip.h if_ether.h in_systm.h ip_icmp.h tcp_fsm.h udp.h igmp.h ip.h ip_mroute.h tcp_seq.h [coyotte:/usr/include/netinet] stephane% And it's not in the Kernel.framework either. I installed 6C125 and 6C115. August Dev tools. My 10.2 system, installed with only the contents of the released CDs on a new volume, seems to have it. I remember there was a problem with them being missing in earlier seeds of the developer tools, but I thought for sure that was addressed. I'm sure someone more knowledgeable than I will chime in with exactly which versions of things have the fix. If that's not your cup of tea (and it's certainly not the typical business model of any long-time Mac developer), we are in the process of creating a set of more sustainable APIs for such kernel extensions. But as good as we are at guessing your needs, we aren't psychic. You [all] need to let us know what services you need sustainable interfaces for. Well do our best to include those [or comparable equivalents]. This makes me think I'm currently wondering why memcmp is said not to be defined when strings.h is included in a C source of a Kernel extensions. This is where that distinction between /usr/include and Kernel.framework really kicks in. The kernel runtime environment is [purposely] MUCH smaller than what you have available at user space. As such, /System/Library/Frameworks/Kernel.framework/Headers/string.h is NOT the identical file you have available at user space through /usr/include/string.h. It is a greatly limited subset for what is available in-kernel. Ouch, I will investigate this. Thanks for the info. _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0