May 2005 - Darwin-kernel - lists.apple.com

Re: KPI replacement for sodisconnect?
by Brian Bergstrand 18 May '05

18 May '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Thanks Josh. Does this also work for TCP sockets or will it return an error since there's an actual connection state? On May 17, 2005, at 6:42 PM, Josh Graessley wrote: > > You do this the same way in the kernel you would in user space. > Connect again but specify the INADDR_ANY address (0.0.0.0). Yeah, > it's sorta weird, but that's just the way it is. You can also > connect directly to a different address. > > -josh > > On May 17, 2005, at 3:32 PM, Brian Bergstrand wrote: > > >> -----BEGIN PGP SIGNED MESSAGE----- >> Hash: SHA1 >> >> Am I missing something, or is there no KPI replacement for >> sodisconnect()? We were using this on UDP sockets to disassociate an >> address set with soconnect(). >> > Brian Bergstrand <http://www.bergstrand.org/brian/> PGP Key ID: 0xB6C7B6A2 -----BEGIN PGP SIGNATURE----- Version: PGP Desktop 9.0.1 (Build 2130) iQA/AwUBQorFV3nR2Fu2x7aiEQJQcwCgnls3/rEqoMOvccDGZdPm+++8Qh0AoL5n zbTB85CC9Le7a73zdziSXPGs =zlt7 -----END PGP SIGNATURE----- _______________________________________________ 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: KPI replacement for sodisconnect?
by Josh Graessley 17 May '05

17 May '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com -josh On May 17, 2005, at 3:32 PM, Brian Bergstrand wrote: -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Am I missing something, or is there no KPI replacement for sodisconnect()? We were using this on UDP sockets to disassociate an address set with soconnect(). If you try to send data on a UDP socket to a diff address than the one used in soconnect() you get back EISCONN. (Yes, this is expected functionality in our product.) We were also using this on TCP sockets to close the channel w/o destroying the socket. I though sock_shutdown() might work, but it looks to be doing more than a disconnect. TIA. Brian Bergstrand <http://www.bergstrand.org/brian/> PGP Key ID: 0xB6C7B6A2 -----BEGIN PGP SIGNATURE----- Version: PGP Desktop 9.0.1 (Build 2130) This email sent to jgraessley(a)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 You do this the same way in the kernel you would in user space. Connect again but specify the INADDR_ANY address (0.0.0.0). Yeah, it's sorta weird, but that's just the way it is. You can also connect directly to a different address. iQA/AwUBQopxC3nR2Fu2x7aiEQK/zQCglAPCHPPJFyxFgQnklW5J1a9KeXUAn2Hu w6gOdzAbF21uBCkVCPNDUMmD =fiCa -----END PGP SIGNATURE----- _______________________________________________ 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/jgraessley% 40apple.com smime.p7s

1 0

KPI replacement for sodisconnect?
by Brian Bergstrand 17 May '05

17 May '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Am I missing something, or is there no KPI replacement for sodisconnect()? We were using this on UDP sockets to disassociate an address set with soconnect(). If you try to send data on a UDP socket to a diff address than the one used in soconnect() you get back EISCONN. (Yes, this is expected functionality in our product.) We were also using this on TCP sockets to close the channel w/o destroying the socket. I though sock_shutdown() might work, but it looks to be doing more than a disconnect. TIA. Brian Bergstrand <http://www.bergstrand.org/brian/> PGP Key ID: 0xB6C7B6A2 -----BEGIN PGP SIGNATURE----- Version: PGP Desktop 9.0.1 (Build 2130) iQA/AwUBQopxC3nR2Fu2x7aiEQK/zQCglAPCHPPJFyxFgQnklW5J1a9KeXUAn2Hu w6gOdzAbF21uBCkVCPNDUMmD =fiCa -----END PGP SIGNATURE----- _______________________________________________ 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: Tiger kextcache command
by Gilles Vayssié 17 May '05

17 May '05

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: Problems building xnu-792
by Jake Luck 17 May '05

17 May '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com bootstrap_cmds-49.tar.gz Libstreams-24.1.tar.gz cctools-576.tar.gz IOKitUser-271.tar.gz kext_tools-59.tar.gz http://10k.org/jake/xnu/xnu-compiletools-20050515.tar.bz2 What do your end executable look like? -rw-r--r-- 1 root wheel 4274080 May 16 01:46 /mach_kernel MD5(/mach_kernel) = d99c473950ea0c4410c0f8433230dab8 jake _______________________________________________ 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 too got it working both ways(standalone and darwinbuild) Kevin, thank you for updating darwinbuild to 0.5. For your convenience, here is an updated package containing all the missing tools needed to build a xnu-792 kernel on a retail Mac OS X 10.4 system in standalone. All binaries are compiled using gcc 4.0 from the following darwin 8.0 packages: Also, the RELEASE build of the kernel is slightly smaller in size than the one came on the 10.4 DVD. (could this be because the tools were built with gcc 4.0 and the kernel was built with gcc 3.3?) This email sent to site_archiver(a)lists.apple.com

1 0

Remote debugging traces
by Anshul Makkar 16 May '05

16 May '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com User-agent: Mozilla Thunderbird 1.0.2 (Macintosh/20050317) hello , I am doing remote debugging of my kext through gdb. Is there a way to achieve the above task , Please help. Thanking You _______________________________________________ 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.… In my kext I have put lot of IOLogs. When in my target system I load the kext , on remote host none of the IOLogs get printed. But if I put break points at different functions eg start, probe etc then IOLogs in the respective functions are printed on the remote host. I want that when I load the kext all the IOlogs that are encountered during the loading of the kext (ie during the calling of different functions) should be printed on the remote host (without putting any break point) . I also want to break in between , during the loading of the kext ie I dont want to set the break point in the beginning. This email sent to site_archiver(a)lists.apple.com

1 0

Re: receiving data with soreceive
by Bijoy Thomas 15 May '05

15 May '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com But i guess this would just loop if the server sent exactly 20 bytes. Regards, Bijoy Thomas On May 14, 2005, at 6:59 PM, Brian Bergstrand wrote: -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On May 14, 2005, at 5:39 PM, Justin Walker wrote: On May 14, 2005, at 15:22, Bijoy Thomas wrote: Are you calling soreceive in an so upcall? If so, try setting MSG_DONTWAIT in the flags, and that should give you everything that is in the so buf at that point. MSG_DONTWAIT tells the socket not to block. For the auio resid, use M_COPYALL. This is an alternative, but keep in mind that using this means that your code has to make sure all the data you need is in place, essentially by reconstructing the incoming packet byte-by-byte (i.e., you need to make sure enough bytes are present to form a valid header so you can read the length, so you can make sure enough bytes are present to process the rest of the packet, ....). It also means you are polling for data, which may not be the best idea. Cheers, Justin Hmm, don't know where you got the polling from. upcalls are event based, so when there is data to read (or the sock is ready for writing) your upcall (callback) is called. If all of the data is not in the socket at that point, you buffer it and wait for the next event. No polling involved. However it does mean more work on your part because everything is async. Bijoy: Definitely get Stevens' TCP book as Justin recommended -- this is all covered in detail. Brian Bergstrand <http://www.bergstrand.org/brian/> PGP Key ID: 0xB6C7B6A2 -----BEGIN PGP SIGNATURE----- Version: PGP Desktop 9.0.1 (Build 2130) This email sent to bijoyjth(a)bluebottle.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.… I tried using MSG_PEEK with soreceive and it does return the actual number of bytes that the server has sent. So i guess making a soreceive with MSG_PEEK followed by a soreceive call with uio_resid set should return whatever data the server sent. I also tried the following code do { auio.uio_resid = 20; result = soreceive(sock,NULL,&auio,&m,NULL,&flags); //flags = MSG_DONTWAIT } while ((auio.uio_resid == 0) || (auio.uio_resid == 20)); iQA/AwUBQoaC6nnR2Fu2x7aiEQJsJQCgjVA6tN8qdfqP9RS2KQiqiTrFvyUAnjcd AwNZPhnBaZUJWlvQIchE6kRa =zVVH -----END PGP SIGNATURE----- _______________________________________________ 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/ bijoyjth%40bluebottle.com This email sent to site_archiver(a)lists.apple.com

1 0

Re: receiving data with soreceive
by Justin Walker 14 May '05

14 May '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On May 14, 2005, at 15:59, Brian Bergstrand wrote: -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On May 14, 2005, at 5:39 PM, Justin Walker wrote: On May 14, 2005, at 15:22, Bijoy Thomas wrote: Are you calling soreceive in an so upcall? If so, try setting MSG_DONTWAIT in the flags, and that should give you everything that is in the so buf at that point. MSG_DONTWAIT tells the socket not to block. For the auio resid, use M_COPYALL. This is an alternative, but keep in mind that using this means that your code has to make sure all the data you need is in place, essentially by reconstructing the incoming packet byte-by-byte (i.e., you need to make sure enough bytes are present to form a valid header so you can read the length, so you can make sure enough bytes are present to process the rest of the packet, ....). It also means you are polling for data, which may not be the best idea. Cheers, Justin upcalls are event based, so when there is data to read (or the sock is ready for writing) your upcall (callback) is called. If all of the data is not in the socket at that point, you buffer it and wait for the next event. No polling involved. However it does mean more work on your part because everything is async. Cheers, Justin -- Justin C. Walker, Curmudgeon at Large Institute for General Semantics ----------- My wife 'n kids 'n dog are gone, I can't get Jesus on the phone, But Ol' Milwaukee's Best is my best friend. ----------- _______________________________________________ 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.… Hmm, don't know where you got the polling from. Sorry for the confusion; this was germane to MSG_DONTWAIT: if you combine it with event-based notifications (i.e., wait for the next little bit if there's not enough yet), you're correct - polling need not be an issue (unless you have a notification for each byte; then it could be painful). Except for the above...in general, you're right, but one does have to worry about failure modes and edge cases :-} This email sent to site_archiver(a)lists.apple.com

1 0

Re: receiving data with soreceive
by Brian Bergstrand 14 May '05

14 May '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On May 14, 2005, at 5:39 PM, Justin Walker wrote: > > On May 14, 2005, at 15:22, Bijoy Thomas wrote: > >>> Are you calling soreceive in an so upcall? If so, try setting >>> MSG_DONTWAIT in the flags, and that should give you everything that >>> is in the so buf at that point. MSG_DONTWAIT tells the socket not to >>> block. For the auio resid, use M_COPYALL. >>> > > This is an alternative, but keep in mind that using this means that > your code has to make sure all the data you need is in place, > essentially by reconstructing the incoming packet byte-by-byte > (i.e., you need to make sure enough bytes are present to form a > valid header so you can read the length, so you can make sure > enough bytes are present to process the rest of the packet, ....). > It also means you are polling for data, which may not be the best > idea. > > Cheers, > > Justin > Hmm, don't know where you got the polling from. upcalls are event based, so when there is data to read (or the sock is ready for writing) your upcall (callback) is called. If all of the data is not in the socket at that point, you buffer it and wait for the next event. No polling involved. However it does mean more work on your part because everything is async. Bijoy: Definitely get Stevens' TCP book as Justin recommended -- this is all covered in detail. Brian Bergstrand <http://www.bergstrand.org/brian/> PGP Key ID: 0xB6C7B6A2 -----BEGIN PGP SIGNATURE----- Version: PGP Desktop 9.0.1 (Build 2130) iQA/AwUBQoaC6nnR2Fu2x7aiEQJsJQCgjVA6tN8qdfqP9RS2KQiqiTrFvyUAnjcd AwNZPhnBaZUJWlvQIchE6kRa =zVVH -----END PGP SIGNATURE----- _______________________________________________ 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: receiving data with soreceive
by Justin Walker 14 May '05

14 May '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On May 14, 2005, at 15:33, Bijoy Thomas wrote: Regards, Justin -- Justin C. Walker, Curmudgeon-at-Large /~\ The ASCII \ / Ribbon Campaign X Help Cure HTML Email / \ _______________________________________________ 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 am using SOCK_STREAM. What i am trying to do is set up a connection with an ftp server. I call soconnect and then soreceive to get the welcome message. I then send the username and call soreceive to get the message asking for the passwd. Then i send the passwd and call soreceive to get the login message. Each time, the data send by the server varies. Since you are using SOCK_STREAM, you should be aware that there is *nothing* in the definition of either SOCK_STREAM or TCP that guarantees anything like message boundaries. This means that your application (or your kernel code) has the responsibility to make sure that you have received all the data you expect; nothing in the kernel will do that for you. Using MSG_PEEK is probably the best way to look at what is currently queued, but you still have to watch out for the Perils of Poling :-} This email sent to site_archiver(a)lists.apple.com

1 0