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

Re: dlil_inject_if_input
by Justin Walker 08 Apr '05

08 Apr '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Apr 7, 2005, at 16:39, Carl Smith wrote: Is it not being initialized in the line ' pHeader = *frame_ptr;'? This is just an assignment of the data being passed in from the filter routine. D'oh! My bad. Back to your regularly scheduled question... Regards, Justin -- Justin C. Walker, Curmudgeon-At-Large Institute for General Semantics -------- Men are from Earth. Women are from Earth. Deal with it. -------- _______________________________________________ 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

FW: dlil_inject_if_input
by Carl Smith 07 Apr '05

07 Apr '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Thread-index: AcU7yBdHqfz3n4z2TGi2I4M23Dgu0AAAdmRAAABFjnA= Thread-topic: dlil_inject_if_input Is it not being initialized in the line ' pHeader = *frame_ptr;'? This is just an assignment of the data being passed in from the filter routine. Carl -----Original Message----- From: darwin-kernel-bounces+csmith=fortresstech.com(a)lists.apple.com [mailto:darwin-kernel-bounces+csmith=fortresstech.com@lists.apple.com] On Behalf Of Justin Walker Sent: Thursday, April 07, 2005 7:18 PM To: darwin-kernel Dev Subject: Re: dlil_inject_if_input On Apr 7, 2005, at 15:54, Carl Smith wrote: > Ok this is driving me crazy, as you probable know from my other > messages > I am filtering packets at the DLIL interface layer, my problem is that > when ever I call dlil_inject_if_input my Mac crashes. Below is the code > in my interface filter input routine: > > Int My_intr_filtr_input(caddr_t cookie, struct ifnet **ifnet_ptr, > struct > mbuf **m_orig, char **frame_ptr) > { > struct My_Attached_Filters *ifb = (struct > My_Attached_Filters *) cookie; > char *pHeader; > pHeader = *frame_ptr; > > struct mbuf *m; > m = *m_orig; > > u_int16_t etype = ntohs(*(u_int16_t*)&pHeader[12]); Just a guess, but isn't pHeader supposed to be initialized before you use it? Cheers, Justin -- Justin C. Walker, Curmudgeon-At-Large Institute for General Semantics -------- "Weaseling out of things is what separates us from the animals. Well, except the weasel." - Homer J Simpson -------- _______________________________________________ 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/csmith%40fortresste ch.com This email sent to csmith(a)fortresstech.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: dlil_inject_if_input
by Justin Walker 07 Apr '05

07 Apr '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Apr 7, 2005, at 15:54, Carl Smith wrote: struct mbuf *m; m = *m_orig; Cheers, Justin -- Justin C. Walker, Curmudgeon-At-Large Institute for General Semantics -------- "Weaseling out of things is what separates us from the animals. Well, except the weasel." - Homer J Simpson -------- _______________________________________________ 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.… Ok this is driving me crazy, as you probable know from my other messages I am filtering packets at the DLIL interface layer, my problem is that when ever I call dlil_inject_if_input my Mac crashes. Below is the code in my interface filter input routine: Int My_intr_filtr_input(caddr_t cookie, struct ifnet **ifnet_ptr, struct mbuf **m_orig, char **frame_ptr) { struct My_Attached_Filters *ifb = (struct My_Attached_Filters *) cookie; char *pHeader; pHeader = *frame_ptr; u_int16_t etype = ntohs(*(u_int16_t*)&pHeader[12]); Just a guess, but isn't pHeader supposed to be initialized before you use it? This email sent to site_archiver(a)lists.apple.com

1 0

dlil_inject_if_input
by Carl Smith 07 Apr '05

07 Apr '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Thread-index: AcU7xM5t1qEwPa2XTOGbdL7XdPtDpw== Thread-topic: dlil_inject_if_input _______________________________________________ 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: dlil_attah_interface_filter
by Justin Walker 07 Apr '05

07 Apr '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Apr 7, 2005, at 12:32, Carl Smith wrote: Yup. The logic in SharedIP is not for everyone :-} Cheers, Justin -- Justin C. Walker, Curmudgeon-At-Large Institute for General Semantics -------- It's not whether you win or lose... It's whether *I* win or lose. -------- _______________________________________________ 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.… OK thanks Justin I believe I understand. So as long as I am working only in X I should ignore the logic in SharedIP and in my filtered 'in' routine use dlil_inject_if_input and in my filtered 'out' routine continue to use dlil_inject_if_output? This email sent to site_archiver(a)lists.apple.com

1 0

FW: dlil_attah_interface_filter
by Carl Smith 07 Apr '05

07 Apr '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Thread-index: AcU7jXG/5j+nnBLvRMqf22H5ezdHMwAF+H6QAADNJXA= Thread-topic: dlil_attah_interface_filter OK thanks Justin I believe I understand. So as long as I am working only in X I should ignore the logic in SharedIP and in my filtered 'in' routine use dlil_inject_if_input and in my filtered 'out' routine continue to use dlil_inject_if_output? Carl -----Original Message----- From: darwin-kernel-bounces+csmith=fortresstech.com(a)lists.apple.com [mailto:darwin-kernel-bounces+csmith=fortresstech.com@lists.apple.com] On Behalf Of Justin Walker Sent: Thursday, April 07, 2005 12:18 PM To: darwin-kernel Dev Subject: Re: dlil_attah_interface_filter On Apr 6, 2005, at 13:33, Carl Smith wrote: >> In my NKE I have attached some filtering using >> 'dlil_attach_interface_filter' and pointed my call back functions via >> this call. >> >> The filtering works find and in my call back function, >> My_if_input(caddr_t cookie, struct ifnet **ifnet_ptr, struct mbuf >> **mbuf_ptr), I want to get/look at the filter ID of the calling > filter. >> This is the value that was obtained in the previous calling to >> 'dlil_attach_interface_filter'. >> > > Another question I have relating to the above statement, is that in > My_if_input I am trapping the mbuf, Ethernet frame, as it is coming in. [snip] Sorry for the delay; I haven't had time to look at this. Short answer: Yes :-} dlil_inject* injects into the indicated stream (input: World -> System; output: System -> World). The confusion arising from SharedIP (other than its twisted architecture) comes from the need to treat "World" on output as (World+X). Therefore, when Classic sends a frame to the "world", that includes X; the output processing then must (selectively) inject frames intended for X into the *input* stream (even though the frame is really going "out to the world"). HTH. Regards, Justin -- Justin C. Walker, Curmudgeon-At-Large Institute for General Semantics -------- When LuteFisk is outlawed, Only outlaws will have LuteFisk -------- _______________________________________________ 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/csmith%40fortresste ch.com This email sent to csmith(a)fortresstech.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: dlil_attah_interface_filter
by Justin Walker 07 Apr '05

07 Apr '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Apr 6, 2005, at 13:33, Carl Smith wrote: In my NKE I have attached some filtering using 'dlil_attach_interface_filter' and pointed my call back functions via this call. The filtering works find and in my call back function, My_if_input(caddr_t cookie, struct ifnet **ifnet_ptr, struct mbuf **mbuf_ptr), I want to get/look at the filter ID of the calling This is the value that was obtained in the previous calling to 'dlil_attach_interface_filter'. Another question I have relating to the above statement, is that in My_if_input I am trapping the mbuf, Ethernet frame, as it is coming in. [snip] HTH. Regards, Justin -- Justin C. Walker, Curmudgeon-At-Large Institute for General Semantics -------- When LuteFisk is outlawed, Only outlaws will have LuteFisk -------- _______________________________________________ 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.… filter. Sorry for the delay; I haven't had time to look at this. Short answer: Yes :-} dlil_inject* injects into the indicated stream (input: World -> System; output: System -> World). The confusion arising from SharedIP (other than its twisted architecture) comes from the need to treat "World" on output as (World+X). Therefore, when Classic sends a frame to the "world", that includes X; the output processing then must (selectively) inject frames intended for X into the *input* stream (even though the frame is really going "out to the world"). This email sent to site_archiver(a)lists.apple.com

1 0

FW: dlil_attah_interface_filter
by Carl Smith 07 Apr '05

07 Apr '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Thread-index: AcU6wWDsW6llAhYSTL6owBOmKVPVJwAJYTUAACjFJ4A= Thread-topic: dlil_attah_interface_filter Bumping this up in hopes of getting some replies. Cheers Carl -----Original Message----- From: darwin-kernel-bounces+csmith=fortresstech.com(a)lists.apple.com [mailto:darwin-kernel-bounces+csmith=fortresstech.com@lists.apple.com] On Behalf Of Carl Smith Sent: Wednesday, April 06, 2005 4:33 PM To: darwin-kernel(a)lists.apple.com Subject: FW: dlil_attah_interface_filter > In my NKE I have attached some filtering using > 'dlil_attach_interface_filter' and pointed my call back functions via > this call. > > The filtering works find and in my call back function, > My_if_input(caddr_t cookie, struct ifnet **ifnet_ptr, struct mbuf > **mbuf_ptr), I want to get/look at the filter ID of the calling filter. > This is the value that was obtained in the previous calling to > 'dlil_attach_interface_filter'. > Another question I have relating to the above statement, is that in My_if_input I am trapping the mbuf, Ethernet frame, as it is coming in. That is from the outside world into the stack/machine. Now if I want to change this packet and send it along it's way and I want to use the functions dlil_inject_if_input, dlil_inject_if_output, am I correct in assuming that in my filtering function My_if_input, I would use dlil_inject_if_input and in My_if_output filter function I would use dlili_inject_if_output to send these 'altered' packets on there way? The only reason I am asking that as I was looking at the sample SharedIP I see that in the 'send' calls, i.e. si_send_eth_ipv4 the code is calling dlil_inject_pr_input. I thought a send was a call to the outside world so I would have thought the proper dlil call would have been dlil_inject_pr_output. This last statement along with the fact that every time I use dlil_inject_if_input in my My_if_input function I am getting a kernel panic, I am wondering if my logic or thinking is wrong. I would appreciate any feedback Carl _______________________________________________ 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/csmith%40fortresste ch.com This email sent to csmith(a)fortresstech.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: Writing to file from Kernel Extension
by William Kucharski 07 Apr '05

07 Apr '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Apr 6, 2005, at 6:39 PM, Justin Walker wrote: William Kucharski kucharsk(a)mac.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.… Read Dean's message again. In fact, search the archives and read many of our responses to similar questions over the years :-}. Ironically enough, there's an article in the May, 2005 Linux Journal covering this very issue. The article is entitled "Things You Should Never Do in the Kernel," if yet another admonition is needed. :-) This email sent to site_archiver(a)lists.apple.com

1 0

Re: Writing to file from Kernel Extension
by Justin Walker 07 Apr '05

07 Apr '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Apr 6, 2005, at 17:32, Carl Smith wrote: Cheers, Justin -- Justin C. Walker, Curmudgeon-At-Large Institute for General Semantics -------- Some people have a mental horizon of radius zero, and call it their point of view. -- David Hilbert -------- _______________________________________________ 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.… That was what I was afraid of but I did see some stuff in sys/file.h, like fo_write, fo_close and so on, but could not find it used any where as an example. Are these not the functions I think they are? First, you need to take a deep breath, practice that Yoga you've been thinking about, and become at peace with the idea that the code practices you are used to using when writing user-mode code are *not* goint to work in the kernel. It is a very restricted environment whose main aim is to support that user-mode code. It needs to do the minimal amount of work necessary to let user applications do their work. File I/O is not one of the things a kernel entity needs to do. Read Dean's message again. In fact, search the archives and read many of our responses to similar questions over the years :-}. This email sent to site_archiver(a)lists.apple.com

1 0