November 2011 - Darwin-kernel

Re: Safe Assumptions(?)
by Nick Blievers 09 Nov '11

09 Nov '11

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: Safe Assumptions(?)
by Shantonu Sen 09 Nov '11

09 Nov '11

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: Safe Assumptions(?)
by José Cornado 08 Nov '11

08 Nov '11

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :content-type; bh=snpn2BZ9nJ4Yh/JxJY0nFP0zKCwG/Q0pWx0qMBeEINM=; b=c5cp8Jyfl1cnVx7Agq6MIL9v4gFsXuKJ7rbFqJsj4h19IO30Z6/K5+ytdO8SHQrjng gHh9mAWSRaz4zCC75MJn+Xn9h+LOWt6ycdyxdSPPX7pbuBvvZdxO350McN5s/bWseitL AH30b7Ey8kXiVp3bDoR3CGitzxmr5WINSFTMA= _______________________________________________ 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: Safe Assumptions(?)
by Shantonu Sen 08 Nov '11

08 Nov '11

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

Safe Assumptions(?)
by José Cornado 08 Nov '11

08 Nov '11

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=mime-version:date:message-id:subject:from:to:content-type; bh=and1Jw8/S7kEkoXjXwlaUA7r8QRNGoFPszIksjMHpu8=; b=u/AhUk984wu1q5MLrcVyHB77huol4uCxwe4aYFZ3VjrLXQBUJ6O1A6jaYa9wGqVRgO nrVkf/FyBZ4w/zJ1ubUaiBS4eX2eKj7AZSEdJyIDQF7gBPaSwodCY8l3TUJwwrDUbUqe KwqOnb6bgJ6i3geH8lB6GIzchKVpYaPumJbuw= _______________________________________________ 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

KEXT <-> user communication
by Kieran Tyrrell 07 Nov '11

07 Nov '11

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=sender:from:content-type:content-transfer-encoding:subject:date :message-id:to:mime-version:x-mailer; bh=OCY1LyKeZ5WQVSPW0yrGkodtW+1KebxTO//9yBBsHTg=; b=jJflyhcVaUJ4PQge5YUXHQ4T9XVfH4mdYEOHe6ppu8ZVAZM0LD6pA7neYrD7xITq4E LMfZJT8sI7lvtXCzb5xcAiQo0mirq9n1G4iqFAaBHk0IST624wIdv3xzs8+FZHYgB5Fx ELwGDFOth6xwM6pwireC99bCG1pIhp4AKMoRg= Hi, I have a IP filter kext (based on the enetlognke sample code) which receives network packets and forwards them to a user daemon process running in the mach realtime scheduling class. Currently I have implemented this communication between the kext and the user process using a PF_SYSTEM / SYSPROTO_CONTROL socket, by calling ctl_enqueuedata from the kext (as illustrated in the tcplognke sample). The problem I am having is that the arriving network packets are processed promptly by the filter, and the user daemon process (running at priority 97) has the capacity to process them as soon as they are received, but there are sometimes delays between the filter calling ctl_enqueuedata and the user process receiving the data. The delays can be up to 16ms. So my question is, are there any options to improve the performance of PF_SYSTEM / SYSPROTO_CONTROL sockets and ctl_enqueuedata, and if not, as an alternative, how can I share a mach semaphore and shared memory region between the kext and the user process? (I have managed to create mach semaphores and shared memory regions in both the kext and the user process, but I can't figure out how to share them!) Is there a better way to achieve this low latency kext<->user process communication? Thanks, Kieran. _______________________________________________ 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: Setting up two-machine kernel debugging via firewire
by Eric Gorr 03 Nov '11

03 Nov '11

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Nov 3, 2011, at 3:30 PM, Brian Bechtel <brian.bechtel(a)gmail.com> wrote: > On Thu, Nov 3, 2011 at 5:48 AM, <mailist(a)ericgorr.net> wrote: > >> Ok. So, unless I can locate the 10.6.8 10k549 kernel debug kit (which may >> not even exist), I will need to downgrade the machines to 10.6.7 10J869, >> which I can find at https://developer.apple.com/downloads...correct? > > 10K549 and 10K540 use the same kernel debug kit, Perhaps I missed it, but where is it documented that an exact match is not required in this case? _______________________________________________ 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: Setting up two-machine kernel debugging via firewire
by Brian Bechtel 03 Nov '11

03 Nov '11

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; bh=3fp7J40rYnls2O1EFDSOfnLAiTGo1ResspPFSjyjHRc=; b=agUc3L6TsSWGnCu9jMOcy03wHmHOPV6T0wafU7ZKoSHaKLcrI/89qS9eWpP7FD4yu2 Ny76+EMT2NA/0fuRRsUiyvBWrNsmxOcTkiSiKxg/yCeU/Zt7F6HWfjrynDAcpJIZzLLF ETuNzwwsGGIkQ4ggPHarLJVgOqUapbQD7fTA0= On Thu, Nov 3, 2011 at 5:48 AM, <mailist(a)ericgorr.net> wrote: > Ok. So, unless I can locate the 10.6.8 10k549 kernel debug kit (which may > not even exist), I will need to downgrade the machines to 10.6.7 10J869, > which I can find at https://developer.apple.com/downloads...correct? 10K549 and 10K540 use the same kernel debug kit, available at https://developer.apple.com/downloads/download.action?path=hardware_and_dri… About Mac OS X 10.6.8 Supplemental Update The Mac OS X 10.6.8 Supplemental Update is recommended for all users running Mac OS X Snow Leopard 10.6.8 and resolves issues with: * Transferring personal data, settings, and compatible applications from a Mac running Mac OS X Snow Leopard to a new Mac running Mac OS X Lion * Certain network printers that pause print jobs immediately and fail to complete * System audio that stops working when using HDMI or optical audio out If you aren't working with audio kexts, the SDK should be identical. _______________________________________________ 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: Setting up two-machine kernel debugging via firewire
by mailist＠ericgorr.net 03 Nov '11

03 Nov '11

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com User-agent: Roundcube Webmail/0.5.4 Shantonu On Nov 3, 2011, at 5:48 AM, mailist(a)ericgorr.net wrote: Thank you for the explanation. It was useful. Are you sure your target is running the 32-bit kernel? What does uname -a print before you enter the kernel debugger. So, armed with this information, I invoked gdb correctly by doing: gdb /Volumes/KernelDebugKit/mach_kernel --arch=x86_64 and, now, showallkmods works. Yes, you need to have your kernel debug kit exactly match whatever build your target is running. There is no close enough. Will do - as soon as bugreport comes back up (appears to be down at the moment). Bug report filed: rdar://10390500 _______________________________________________ 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 Thu, 03 Nov 2011 07:59:47 -0500, mailist(a)ericgorr.net wrote: On Thu, 03 Nov 2011 05:57:04 -0700, Shantonu Sen wrote: You should file a bug report at bugreport.apple.com asking that it be made available. In this particular case the kernel didn't change between the builds (so it might kind of look like it works), but other content that contributes to the Kernel Debug Kit did change, so you might see strange behavior until you get it synchronized. I'd recommend using --arch=x86_64 at the start of your gdb invocation. On Thu, 03 Nov 2011 05:18:34 -0700, Shantonu Sen wrote: kdp-reattach lets you steal a connection in case someone else is currently attached. It can also help in cases where a debug session is interrupted and you need to get the panicked system to just reset its internal state machine. Brilliant. I was pretty sure I was running a 64-bit kernel and rechecked in the system profiler. In the entry for 64-bit Kernel and Extensions in the Software section, it said yes. Ok. So, unless I can locate the 10.6.8 10k549 kernel debug kit (which may not even exist), I will need to downgrade the machines to 10.6.7 10J869, which I can find at https://developer.apple.com/downloads...correct? This email sent to site_archiver(a)lists.apple.com

1 0

Re: Setting up two-machine kernel debugging via firewire
by mailist＠ericgorr.net 03 Nov '11

03 Nov '11

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com User-agent: Roundcube Webmail/0.5.4 Shantonu On Nov 3, 2011, at 5:48 AM, mailist(a)ericgorr.net wrote: Thank you for the explanation. It was useful. Are you sure your target is running the 32-bit kernel? What does uname -a print before you enter the kernel debugger. So, armed with this information, I invoked gdb correctly by doing: gdb /Volumes/KernelDebugKit/mach_kernel --arch=x86_64 and, now, showallkmods works. Yes, you need to have your kernel debug kit exactly match whatever build your target is running. There is no close enough. In the meantime, I'm off to downgrade the machines. _______________________________________________ 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 Thu, 03 Nov 2011 05:57:04 -0700, Shantonu Sen wrote: You should file a bug report at bugreport.apple.com asking that it be made available. In this particular case the kernel didn't change between the builds (so it might kind of look like it works), but other content that contributes to the Kernel Debug Kit did change, so you might see strange behavior until you get it synchronized. I'd recommend using --arch=x86_64 at the start of your gdb invocation. On Thu, 03 Nov 2011 05:18:34 -0700, Shantonu Sen wrote: kdp-reattach lets you steal a connection in case someone else is currently attached. It can also help in cases where a debug session is interrupted and you need to get the panicked system to just reset its internal state machine. Brilliant. I was pretty sure I was running a 64-bit kernel and rechecked in the system profiler. In the entry for 64-bit Kernel and Extensions in the Software section, it said yes. Ok. So, unless I can locate the 10.6.8 10k549 kernel debug kit (which may not even exist), I will need to downgrade the machines to 10.6.7 10J869, which I can find at https://developer.apple.com/downloads...correct? Will do - as soon as bugreport comes back up (appears to be down at the moment). This email sent to site_archiver(a)lists.apple.com

1 0