May 2019 - Darwin-kernel - lists.apple.com

Re: Darwin-kernel Digest, Vol 16, Issue 10
by Prokash Sinha via Darwin-kernel 30 May '19

30 May '19

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com I think the problem could be Enet over VM…, in my case. In local subnet, I will try two real machines over internet to see how it respond. I know GDB and KGDB are not true sense kernel debugger. So I don’t expect LLDB going to be a real kernel debugger either, but system gets locked after two to three hard breaks ( __asm int3;) even though I respond fairly fast is pain… For example, Window kernel debugger I can set a break point, once it breaks into - I can take a vacation and start debugging from there after few days… Since KDP protocol does not send much across the line, even a NULL serial could be fine… /r, prokash > > ---------------------------------------------------------------------- > > Message: 1 > Date: Thu, 30 May 2019 17:21:32 +0200 > From: Phil Dennis-Jordan <lists(a)philjordan.eu> > To: joe.lowe(a)pismotechnic.com > Cc: "darwin-kernel(a)lists.apple.com Kernel" > <darwin-kernel(a)lists.apple.com> > Subject: Re: Been bothering ever since 10.12+ > Message-ID: > <CAGCz3vt9A58zteGcamV=m_=4L1y=9SA9XmHErVNSqAvWhon6rA(a)mail.gmail.com> > Content-Type: text/plain; CHARSET=US-ASCII > > On Fri, May 17, 2019 at 12:29 AM <joe.lowe(a)pismotechnic.com> wrote: >> In my experience: The mac kernel intermittently locks up during >> single-step kernel debugging. This has been an issue since well before >> 10.12, at least going back to 10.6 . >> >> Using breakpoints instead of single-step mostly avoids the kernel >> debugger lock-ups, but still happens occasionally. For issues where >> breakpoints don't cut it, things can be pretty frustrating. > > I can confirm all of the above and will add: it seems to be a little > more reliable over Firewire versus Ethernet. Plus, commands with large > output (showallthreads and such) tend to run much faster. These days, > using Firewire means a multitude of dongles, but what can you do. It'd > be nice if you could do it via a direct Thunderbolt connection or if > macOS started supporting the special USB 3.0 A-to-A debug cable that > can apparently be used for debugging the Linux and Windows kernels. > > Phil > > > ------------------------------ > > Subject: Digest Footer > > _______________________________________________ > Darwin-kernel mailing list > Darwin-kernel(a)lists.apple.com > https://lists.apple.com/mailman/listinfo/darwin-kernel > > > ------------------------------ > > End of Darwin-kernel Digest, Vol 16, Issue 10 > ********************************************* _______________________________________________ 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: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Been bothering ever since 10.12+
by Phil Dennis-Jordan via Darwin-kernel 30 May '19

30 May '19

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=philjordan-eu.20150623.gappssmtp.com; s=20150623; h=mime-version:references:in-reply-to:from:date:message-id:subject:to :cc; bh=YnfFxSQXULZrsAOIQCOwNwbQh5aqD962Axd1sM1uIsE=; b=1yWQItXBJyN1dCK+yeZgmColKUGyktH2/nBYCnR/dL5Jt1Wb5tFpL4eQytg0ResXOJ 9jRy5vpHZq3ba7F8s7GTwQZM+otbsiRtw37tp1JBzUJJtuYf5qwtUgu7wc2KLPuuhjB9 sVtc/5DMxEpYhKOk7ko8Ozm5hpLGoZTyZAiZiJ22XsPIJ1aMaUX1I+r6/wwOEgyqR+rW hamfcfdFEaeZxorAB9thqZog2Jx3n0Zd27vYGZYOtn/FIlh5XcFGfnCYBQyoUTlc+e6H j5TPvd6jky/xEUh79Snsa5XlO5VW9VDjKt0bZ3GpV21yv50HiJZJq9VuTNnK18l/vT5I I0Kg== On Fri, May 17, 2019 at 12:29 AM <joe.lowe(a)pismotechnic.com> wrote: > In my experience: The mac kernel intermittently locks up during > single-step kernel debugging. This has been an issue since well before > 10.12, at least going back to 10.6 . > > Using breakpoints instead of single-step mostly avoids the kernel > debugger lock-ups, but still happens occasionally. For issues where > breakpoints don't cut it, things can be pretty frustrating. I can confirm all of the above and will add: it seems to be a little more reliable over Firewire versus Ethernet. Plus, commands with large output (showallthreads and such) tend to run much faster. These days, using Firewire means a multitude of dongles, but what can you do. It'd be nice if you could do it via a direct Thunderbolt connection or if macOS started supporting the special USB 3.0 A-to-A debug cable that can apparently be used for debugging the Linux and Windows kernels. Phil _______________________________________________ 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: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Is there a command to recursively search for kernel source in lldb
by Prokash Sinha 18 May '19

18 May '19

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com For some reason, more generic paths don’t work for me. For the source mapping, in most cases the sources are not that different, and debugger is blissfully ignorant while map it. Once I get to an approximate point, disassembling points to the right places. -p > On May 16, 2019, at 4:18 PM, Jason Molenda <jmolenda(a)apple.com> wrote: > > target.source-map should work the way you're imagining already. The first > name ("/BuildRoot/Library/Caches/" etc) is substituted in the debug > information for the second name that you provide. You can provide a less > specific remapping. I'm not sure about pointing the debugger at > xnu-4903.221.2 sources when you're debugging a xnu-4903.240.8 kernel - are > there many differences in the source code? It seems like it could be > confusing. > > >> On May 8, 2019, at 11:30 AM, Prokash Sinha <prokash(a)garlic.com> wrote: >> >> Hi All, >> >> Instead of — >> >> settings append target.source-map >> /BuildRoot/Library/Caches/com.apple.xbs/Sources/xnu/xnu-4903.240.8/bsd/kern >> ~/xnu_srcs/xnu_src/xnu-4903.221.2/bsd/kern >> >> Is it possible to do like - >> >> settings append target.source-map >> /BuildRoot/Library/Caches/com.apple.xbs/Sources/xnu/xnu-4903.240.8/bsd >> ~/xnu_srcs/xnu_src/xnu-4903.221.2/bsd >> >> So all the source could be resolved from symbols ??? >> >> >> Thanks, >> Prokash >> _______________________________________________ >> 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: >> https://lists.apple.com/mailman/options/darwin-kernel/jmolenda%40apple.com >> >> This email sent to jmolenda(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: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Is there a command to recursively search for kernel source in lldb
by Jason Molenda 16 May '19

16 May '19

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com target.source-map should work the way you're imagining already. The first name ("/BuildRoot/Library/Caches/" etc) is substituted in the debug information for the second name that you provide. You can provide a less specific remapping. I'm not sure about pointing the debugger at xnu-4903.221.2 sources when you're debugging a xnu-4903.240.8 kernel - are there many differences in the source code? It seems like it could be confusing. > On May 8, 2019, at 11:30 AM, Prokash Sinha <prokash(a)garlic.com> wrote: > > Hi All, > > Instead of — > > settings append target.source-map > /BuildRoot/Library/Caches/com.apple.xbs/Sources/xnu/xnu-4903.240.8/bsd/kern > ~/xnu_srcs/xnu_src/xnu-4903.221.2/bsd/kern > > Is it possible to do like - > > settings append target.source-map > /BuildRoot/Library/Caches/com.apple.xbs/Sources/xnu/xnu-4903.240.8/bsd > ~/xnu_srcs/xnu_src/xnu-4903.221.2/bsd > > So all the source could be resolved from symbols ??? > > > Thanks, > Prokash > _______________________________________________ > 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: > https://lists.apple.com/mailman/options/darwin-kernel/jmolenda%40apple.com > > This email sent to jmolenda(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: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Been bothering ever since 10.12+
by Ken Hornstein 16 May '19

16 May '19

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=nrl.navy.mil; h=message-id : from : to : subject : in-reply-to : references : mime-version : content-type : date; s=s1.dkim; bh=1aYImeYnbNK+9Roc1QTrs7CEFrtfVvtsLdIqI80Gh8Q=; b=xE0Z1Lhz/dVmutlxB7XdPE/6DM1FYGfVutbnZfZehC41XRYIBbwX/R159FT1xUqKgNVh 05zPc5xOn95HQDlDJAoZhHwVoyuzbjQD+glY6zX/XkLdFYUgypVln+j1AVjhAp0h3Xv+ NH/VIkFXkLw0mx2dlgEzSaiZTTAdk8EE4iw= >In my experience: The mac kernel intermittently locks up during >single-step kernel debugging. This has been an issue since well before >10.12, at least going back to 10.6 . > >Using breakpoints instead of single-step mostly avoids the kernel >debugger lock-ups, but still happens occasionally. For issues where >breakpoints don't cut it, things can be pretty frustrating. Totally agree; I ran into this problem as well. From memory, you get a limited number of single-steps before you should continue or you get a hang. I also recommend using as many breakpoints as possible. --Ken _______________________________________________ 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: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Been bothering ever since 10.12+
by joe.lowe＠pismotechnic.com 16 May '19

16 May '19

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com User-agent: Mozilla/5.0 (Windows NT 6.3; WOW64; rv:60.0) Gecko/20100101 Thunderbird/60.5.1 Joe On 2019-05-16 14:19, Prokash Sinha wrote: Hi All, I do kernel debugging using lldb and target being a fusion VM. It will sincerely help me save lots of annoying time. This, IMO is not a true sense kernel debugger… TIA, -p _______________________________________________ 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: https://lists.apple.com/mailman/options/darwin-kernel/joe.lowe%40pismotechn… This email sent to joe.lowe(a)pismotechnic.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: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com In my experience: The mac kernel intermittently locks up during single-step kernel debugging. This has been an issue since well before 10.12, at least going back to 10.6 . Using breakpoints instead of single-step mostly avoids the kernel debugger lock-ups, but still happens occasionally. For issues where breakpoints don't cut it, things can be pretty frustrating. When step thru the kext code, say for example the _start code, after few seconds, … kernel stopped messages I get. I was wondering if there is any additional param I need to put into my existing param — boot-args-v debug=0x146 pmuflags=1 watchdog=0 kcsuffix=development kext-dev-mode=1

1 0

Been bothering ever since 10.12+
by Prokash Sinha 16 May '19

16 May '19

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Hi All, I do kernel debugging using lldb and target being a fusion VM. When step thru the kext code, say for example the _start code, after few seconds, … kernel stopped messages I get. I was wondering if there is any additional param I need to put into my existing param — boot-args -v debug=0x146 pmuflags=1 watchdog=0 kcsuffix=development kext-dev-mode=1 It will sincerely help me save lots of annoying time. This, IMO is not a true sense kernel debugger… TIA, -p _______________________________________________ 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: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0

Is there a command to recursively search for kernel source in lldb
by Prokash Sinha 08 May '19

08 May '19

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Hi All, Instead of — settings append target.source-map /BuildRoot/Library/Caches/com.apple.xbs/Sources/xnu/xnu-4903.240.8/bsd/kern ~/xnu_srcs/xnu_src/xnu-4903.221.2/bsd/kern Is it possible to do like - settings append target.source-map /BuildRoot/Library/Caches/com.apple.xbs/Sources/xnu/xnu-4903.240.8/bsd ~/xnu_srcs/xnu_src/xnu-4903.221.2/bsd So all the source could be resolved from symbols ??? Thanks, Prokash _______________________________________________ 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: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0

Adding a 3rd party KEXT to a prelinkedkernel with safe boot and T2
by Slava Imameev 07 May '19

07 May '19

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=20161025; h=mime-version:from:date:message-id:subject:to; bh=NrpdlogOEdzyWcGK9iM1qxa5WO0VtdS3vBH/6zKRutg=; b=B5BSr6BA7ZreOQWZFYC9c8VfkFUvyLqcwRWDUDBWfxeb5ZWZHORH2mBIKO1PHJD8sZ ocQA/kFrAnfZGSLjH0RmHVhfDutd6yjDt5MmAsZSKb2G+HTJa3WAk+ibrZcZxyqSLGsG 2pmqYWlebo35fXmpe+JpPa7Rhg3gbVLx9uli42K88vLKwGcTB6nnLq9CrZwnbO/6Crz0 7oWLbfxrsICf7SZ6mHvo315IMqffmr4NTGm3bp1oZuD6VSgVv3h19UY2fJAjGrQlk+9p akEAiTr+cjevm99qo4ffl2qamZhhttSiHiYmgVZTNXJj8gVdmsoVgbWuLp7reYnawqMl WDjw== Hi, Can somebody confirm that it is not possible to add a 3rd party KEXT to a prelinked kernel on systems with the T2 chip and active safeboot mode? I need an IOKit object being probed against a device nub as early as possible. This requires a KEXT to be added to a prelinked kernel hosted at /System/Library/PrelinkedKernel . I noticed that the system uses /System/Library/PrelinkedKernel/immutablekernel instead of /System/Library/PrelinkedKernel/prelinkedkernel if safeboot is active and the system has the T2 chip . The immutablekernel is protected by the UEFI boot module which checks a prelinked kernel signature saved at the Preboot volume so there is no way to modify it with the kextcache command. The existing procedure when a 3rd party KEXT is being loaded by the kextd process after the system has been initialized might not be the best solution in my case. Is there a procedure for a 3rd party KEXT to be linked with immutablekernel or change the system behavior to load /System/Library/PrelinkedKernel/prelinkedkernel in safeboot mode ? Regards, Slava Imameev _______________________________________________ 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: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists… This email sent to site_archiver(a)lists.apple.com

1 0