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

Re: [NEVERMIND] what happened to rtredirect()?
by Chase 19 Jun '05

19 Jun '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com thanks. - chase _______________________________________________ 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.… nevermind. i figured out (finally) how to do this properly from user space. i have a couple minor questions regarding the new way (new to me) of doing this, but i'll bug the macnetworking folks about it. This email sent to site_archiver(a)lists.apple.com

1 0

what happened to rtredirect()?
by Chase 19 Jun '05

19 Jun '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com rtredirect() is no longer in route.h in tiger. spotlight isn't finding any headers that mention it either. what can be done to replace that functionality? - chase _______________________________________________ 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

getattrlist needed for Darwin 8?
by Brian Bergstrand 18 Jun '05

18 Jun '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Our filesystem implemented getattrlist under Darwin 7 to return some Volume information -- if the node was not the root we just returned ENOTSUP. Now under Darwin 8 there is getattr at the VFS level and it looks to handle all the stuff that getattrlist used to. So is getattrlist needed anymore? Will old userland clients be shunted through getattr as necessary? Thanks. Brian Bergstrand <http://www.bergstrand.org/brian/> PGP Key ID: 0xB6C7B6A2 -----BEGIN PGP SIGNATURE----- Version: PGP Desktop 9.0.1 (Build 2185) iQA/AwUBQrSexXnR2Fu2x7aiEQIeLwCdEjjZtrh2P2Sl4NvabITKEO3ssFYAn3Hv ZZ4NhEMgPn67SVhC2GxXx7GC =nIyd -----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: vm / physmem info
by Mike Smith 18 Jun '05

18 Jun '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Jun 17, 2005, at 11:24 AM, Glen Beane 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 just really need (actually, would like) to get a *rough* idea of how memory is fragmented - if I have a HPC application that allocates a large amount of memory, how fragmented is that in physical ram. It's not the end of the world if I can't, but it would definitely be helpful (for a number of reasons that are outside the scope of my question)... I'm really only concerned with one or two processes If you can make some minor changes to the application and write an I/ O kit driver, you could easily have the app issue an I/O against the driver and then have the driver use the IOKit functionality to do the virtual->physical mapping. That avoids trying to go backwards, which is what you're describing there. This email sent to site_archiver(a)lists.apple.com

1 0

Re: Processes running path
by Mike Smith 18 Jun '05

18 Jun '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Jun 17, 2005, at 4:35 PM, matt jaffa wrote: Well basically I am trying to associate outgoing traffic with the process that is sending it using a socket filter. It works great, and right now I have my kext sending the pid to a super user process in user space that gets the process path and sends it back. This is less than ideal, I would rather have the kext be able to get this info. I assume you want the path because you use the path to the application to associate the application with some rules regarding the traffic it can send. This isn't very robust though; what if someone moves the application? I just don't want my application to have all these processes have to be running for it to work. Being afraid of dividing responsibility along architectural lines loses you most of these benefits; you should design your application to work with the system, not try to bend the system to suit your prejudices. 8) = 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 would be much more inclined to use the proc_name KPI; in most cases this will remain unique. It's still not such a great idea to depend on the application name, but this is at least resistant to the application moving. I'm a little puzzled by this. The kernel makes it easy to run processes, it provides good mechanisms for communicating between them, and it makes them safe from one another. Having a bunch of processes cooperating can be a neat, tidy and robust way to architect a solution. This email sent to site_archiver(a)lists.apple.com

1 0

Re: Processes running path
by Terry Lambert 18 Jun '05

18 Jun '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com -- Terry Well basically I am trying to associate outgoing traffic with the process that is sending it using a socket filter. It works great, and right now I have my kext sending the pid to a super user process in user space that gets the process path and sends it back. This is less than ideal, I would rather have the kext be able to get this info. In Panther I wrote my own function in the kernel that would look at the process structure and look at the path from the process user space stack. But now Tiger hides the process structure definition and other definitions my function used so it has become useless in Tiger. I just don't want my application to have all these processes have to be running for it to work. Matt On 6/17/05, Terry Lambert <tlambert(a)apple.com> wrote: On Jun 16, 2005, at 11:18 PM, Josh Graessley wrote: There isn't a good, straightforward way of doing this. If you get in touch with DTS, they may be able to work with you towards a solution. We are aware of the issue and have a bug (enhancement request) filed to track it. -josh On Jun 16, 2005, at 9:31 PM, matt jaffa wrote: Does anyone on this list know of some api or hack of using a processes id (pid) in a network kernel extension to get that process running path? Thanks Matt Specifically, the path is stored into the strings area at the top of the user stack during the exec process for use by the "open(1)" command in obtaining bundle ID information. Because this is in memory belonging to the user process, it's possible for the user process to modify its contents after the application has started, so you cannot trust that the information is reliable or accurate to create a security association (as one example use someone might have for needing this information). If it's not for a security association, one method would be to have to program tell the kext. For bundle information, the most correct thing for the process to do is to write a file in /var/run with the information; this is probably not usesful for a kext, though, since you shouldn't be trying to do file I/O from a kext. If you can tell us why you need the information, we can probably tell you a different way to solve the problem you are trying to solve. -- 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.… You could probably do the same thing you were doing, if you link against the specific kernel, instead of linking against the KPI restricted symbol sets. If you did that, though, you would end up having to recompile your kext each time Aplle releases a software update that changes a minor version number - that's intentional, since the KPI's are intended to document what we think won't change between versions, and so it's safe to use and won't change between software updates. The user space daemon seems to me to be the correct approach, actually, since it's (A) more portable and (B) less likely to stop working following an OS update and (C) is likely more portable to Intel systems and (D) will work when it's run under Rosetta. The type of lookup you are trying to do is expensive enough that it should be decoupled from the functioning of the driver. If you just need the command name, and don't care about it's full path, you can get that easily enough from the visible information in the kernel. Most users of the full path tend to be programs in user space that use it to obtain bundle IDs (better done by popen'ing "ps" with the appropriate parameters to get just the process and data you are interested in, and reading only it). That's assuming they don't simply write out a "/var/run/bundle.XXX" file for the other programs that want to rendezvous with the program running from the bundle, which is an even better approach. On Jun 17, 2005, at 4:35 PM, matt jaffa wrote: This email sent to site_archiver(a)lists.apple.com

1 0

IOPMrootDomain and rebootSystem
by Herbert 18 Jun '05

18 Jun '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com I am trying to have my kext call rebootSystem from a driver. Does anyone have any idea how to call this and make it work? Or any other way to get the machine to reboot from a kext? Thanks. _______________________________________________ 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.… The driver is a subclass of IOBlockStorage, which ultimately inherits from IOStorage. IOStorage has a method called getPMrootDomain which should get the global IOPMrootDomain. I should them be able to call rebootSystem from this and have the machine reboot. Everything compiles and runs fine, but the call to rebootSystem returns an unimplemented error message. This email sent to site_archiver(a)lists.apple.com

1 0

Re: Kernel Controls Buffer Size Issue
by Vincent Lubet 18 Jun '05

18 Jun '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com It's a bug, please file a bug report at http://bugreporter.apple.com/ Vincent On Jun 17, 2005, at 5:38 PM, Raj Doshi wrote: Hello, Your assistance is appreciated. This email sent to vlubet(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.… As a workaround you can set the SO_RCVBUF option on kernel control socket once you've connected to it. I'm using kernel controls (Mac OS X 10.4) for communication between a kext and a user space process. The kern_ctl_reg structure that is passed to the ctl_register function has some fields (ctl_sendsize and ctl_recvsize) that can be set to adjust the buffer sizes for messages. When I set these fields to zero, then they are changed to default sizes of 2048 and 8192, respectively, after calling ctl_register. However, when I specify larger values for these fields, they are still set to these default values after the ctl_register call. When I establish a connection, a call to ctl_getenqueuespace confirms that the receive buffer is set to 8192 bytes, even though I specified a larger amount. For my purposes (sending network packets back and forth), I would like these buffers to be at least 32000 bytes. My questions are the following: 1. Am I doing something wrong, or is this a bug? 2. Is there another good method of kext/user space communication that I should consider using (for moderate latency and moderate bandwidth, a pointer to documentation/examples would be helpful)? Raj Doshi _______________________________________________ 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/vlubet% 40apple.com This email sent to site_archiver(a)lists.apple.com

1 0

Kernel Controls Buffer Size Issue
by Raj Doshi 18 Jun '05

18 Jun '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Hello, Your assistance is appreciated. Raj Doshi _______________________________________________ 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 using kernel controls (Mac OS X 10.4) for communication between a kext and a user space process. The kern_ctl_reg structure that is passed to the ctl_register function has some fields (ctl_sendsize and ctl_recvsize) that can be set to adjust the buffer sizes for messages. When I set these fields to zero, then they are changed to default sizes of 2048 and 8192, respectively, after calling ctl_register. However, when I specify larger values for these fields, they are still set to these default values after the ctl_register call. When I establish a connection, a call to ctl_getenqueuespace confirms that the receive buffer is set to 8192 bytes, even though I specified a larger amount. For my purposes (sending network packets back and forth), I would like these buffers to be at least 32000 bytes. My questions are the following: 1. Am I doing something wrong, or is this a bug? 2. Is there another good method of kext/user space communication that I should consider using (for moderate latency and moderate bandwidth, a pointer to documentation/examples would be helpful)? This email sent to site_archiver(a)lists.apple.com

1 0

Re: Processes running path
by matt jaffa 17 Jun '05

17 Jun '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:reply-to:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=SCRBEZrxKUpBdujkb4vgGRXRMkvWA7aSSgUhid1cUJB9r5MPhANYxZk7Byx6xjWw4sLMl1N+zYYrDN8lsSFRiL5b+Ndc/z/Ccz2EiDWfRGbmZjxvKrghbbHWODOA+1c1VarYGaXqF/hn7kT8Hq3dBh0zVTGCvaTzcb5Lmadr4eA= Well basically I am trying to associate outgoing traffic with the process that is sending it using a socket filter. It works great, and right now I have my kext sending the pid to a super user process in user space that gets the process path and sends it back. This is less than ideal, I would rather have the kext be able to get this info. In Panther I wrote my own function in the kernel that would look at the process structure and look at the path from the process user space stack. But now Tiger hides the process structure definition and other definitions my function used so it has become useless in Tiger. I just don't want my application to have all these processes have to be running for it to work. Matt On 6/17/05, Terry Lambert <tlambert(a)apple.com> wrote: > On Jun 16, 2005, at 11:18 PM, Josh Graessley wrote: > > There isn't a good, straightforward way of doing this. > > > > If you get in touch with DTS, they may be able to work with you > > towards a solution. We are aware of the issue and have a bug > > (enhancement request) filed to track it. > > > > -josh > > > > On Jun 16, 2005, at 9:31 PM, matt jaffa wrote: > > > > > >> Does anyone on this list know of some api or hack of using a > >> processes > >> id (pid) in a network kernel extension to get that process running > >> path? > >> > >> Thanks > >> Matt > > Specifically, the path is stored into the strings area at the top of > the user stack during the exec process for use by the "open(1)" > command in obtaining bundle ID information. Because this is in > memory belonging to the user process, it's possible for the user > process to modify its contents after the application has started, so > you cannot trust that the information is reliable or accurate to > create a security association (as one example use someone might have > for needing this information). > > If it's not for a security association, one method would be to have > to program tell the kext. > > For bundle information, the most correct thing for the process to do > is to write a file in /var/run with the information; this is probably > not usesful for a kext, though, since you shouldn't be trying to do > file I/O from a kext. > > If you can tell us why you need the information, we can probably tell > you a different way to solve the problem you are trying to solve. > > -- 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.… This email sent to site_archiver(a)lists.apple.com

1 0