November 2011 - Darwin-kernel

Re: Basic Kernel Extension Programming questions
by Eric Gorr 18 Nov '11

18 Nov '11

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Thank you for the reply. I had considered that and tried 4096 bytes -1 page. Still panics. Very frustrating especially since I know doing this is possible ( google MacMemoryReader ). On Nov 18, 2011, at 11:27 AM, Andrew Gallatin <gallatin(a)cs.duke.edu> wrote: > >>>> What I am attempting to do now is to see if I can loop over all 8gb >>>> of physical memory >>>> >>>> for ( x = 0; x < 1024 * 1024 * 1024 * 8; x++ ) >>>> { >>>> descriptor = IOMemoryDescriptor::withPhysicalAddress( x, 10, >>>> kIODirectionOut ); >>>> descriptor->prepare( kIODirectionOut ); >>>> descriptor->readBytes( 0, buffer, 10 ); >>>> descriptor->complete(); >>>> descriptor->release(); >>>> } > > You will almost certainly still run into problems, but > reading in 10 byte chunks is almost certainly causing > you more problems, since you're going to be reading > odd lengths. > > Try reading 4 or 8 bytes. > > I'd say the odds are good that you'll still crash, but > who knows.. > > Drew _______________________________________________ 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: Basic Kernel Extension Programming questions
by Andrew Gallatin 18 Nov '11

18 Nov '11

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=simple/simple; d=cs.duke.edu; s=mail; t=1321633665; bh=YbttjuKexpQRKQCf1Zqug7iDxe3El2tznp3+TO322cw=; h=Message-ID:Date:From:MIME-Version:To:CC:Subject:References: In-Reply-To:Content-Type:Content-Transfer-Encoding; b=uOuB2eZeBdeDxqgpO19Hax5Qjczvg+CSRMdoaLDHzyjUnHdCXhoUdRo9WjWFQYY3k AxLx+OWKzOH9anhdaZozTsS13bU8PPrKUFFru/KNkpSygBS1z7edscQY22BFzhG8JV JFbt0CELI5CH2F4xCo4wthWdhnK66eK7xXN9uQH21pDNauXwPpHhhk7eKZM9Ay5Wcp 77ide5dZUfI8s4b93gqqKpBUNz8RN8sXk7sEBtAtIkutMbhuzrOfdqFNWzwRiMXorM xPebDWNuLbT+geNJTI9ojBVUcSVaJEeB11tvZGOcmux8SCZj/i2LxKQ6S/7rSLrdUv jOUPjdIdCjiFA== User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:7.0.1) Gecko/20110929 Thunderbird/7.0.1 What I am attempting to do now is to see if I can loop over all 8gb of physical memory for ( x = 0; x < 1024 * 1024 * 1024 * 8; x++ ) { descriptor = IOMemoryDescriptor::withPhysicalAddress( x, 10, kIODirectionOut ); descriptor->prepare( kIODirectionOut ); descriptor->readBytes( 0, buffer, 10 ); descriptor->complete(); descriptor->release(); } You will almost certainly still run into problems, but reading in 10 byte chunks is almost certainly causing you more problems, since you're going to be reading odd lengths. Try reading 4 or 8 bytes. I'd say the odds are good that you'll still crash, but who knows.. Drew _______________________________________________ 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: Basic Kernel Extension Programming questions
by Eric Gorr 18 Nov '11

18 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: Basic Kernel Extension Programming questions
by Ed Wynne 18 Nov '11

18 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: Basic Kernel Extension Programming questions
by Eric Gorr 18 Nov '11

18 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: how do I enable MACH_KDB on 10.8.0 ?
by Shantonu Sen 17 Nov '11

17 Nov '11

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com A couple of years. Consider this evidence of why "what are you really trying to do" is useful. Shantonu On Nov 17, 2011, at 8:01 AM, Andrew Gallatin wrote: > On 11/17/11 10:57, Shantonu Sen wrote: >> I don't admit that :-) You could have gotten 95% of that by setting "keepsyms=1" in boot-args, which would have symbolicated the frames for you. And, it works with K64. > > Ah! I wish I'd have known about that option. How long has it existed? > > > > > Drew > _______________________________________________ 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: how do I enable MACH_KDB on 10.8.0 ?
by Andrew Gallatin 17 Nov '11

17 Nov '11

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=simple/simple; d=cs.duke.edu; s=mail; t=1321545674; bh=ghhI/Ip5esLOAK0B9MyGlkfbLSiYrPh3Thqm9Pu0NhE=; h=Message-ID:Date:From:MIME-Version:To:CC:Subject:References: In-Reply-To:Content-Type:Content-Transfer-Encoding; b=u8lOpW5kBnFLvtS1uTTleELje+dJsfaEXVznP87D788KlpuG74+teiy8dZM7jHtX8 0PZPmJp7lmulMj2s0OYbH4OSnlg87SmitoZHCaAHXDRgFClyGZa8ow2Eg2zWC9n6RC DHw/Fv9tKodCllX+yHKuURowm1c1oUCXDRbhwzE2NkRnmtgsYvB7Coz/jeIzBC5X2L 73C+3shs42FB+nqtWCs+v2hphM3EhC4b5pglFCtdkrm/Q5jMXWPogPmisTANSmB6Kf xnxHVW/zu7MD67osfi+6cOupjBNH9HkKCJT3ZpBHKZWKlqMY1nveNbKZZ3LXvmCBMx m3U6/bI2G0xdw== User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:7.0.1) Gecko/20110929 Thunderbird/7.0.1 Ah! I wish I'd have known about that option. How long has it existed? Drew _______________________________________________ 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 11/17/11 10:57, Shantonu Sen wrote: I don't admit that :-) You could have gotten 95% of that by setting "keepsyms=1" in boot-args, which would have symbolicated the frames for you. And, it works with K64. This email sent to site_archiver(a)lists.apple.com

1 0

Re: how do I enable MACH_KDB on 10.8.0 ?
by Shantonu Sen 17 Nov '11

17 Nov '11

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com I don't admit that :-) You could have gotten 95% of that by setting "keepsyms=1" in boot-args, which would have symbolicated the frames for you. And, it works with K64. Shantonu On Nov 17, 2011, at 7:20 AM, Andrew Gallatin wrote: > > For my purposes, it seems to work. And I wasn't having > much luck with the traditional remote dump methods in > my non-Mac environment (running darwin under qemu-kvm to > talk to a verilog simulator for a device which is still > under development). The one thing I do have is a working > "serial" console, so I figured I'd might as well use it ;) > > At any rate, for some reason, it wasn't > picking up the debug= flags, so I had to hack the trap > routine. I finally have a nice, somewhat meaningful > stack trace: > > > db{0}> tr > thread 0x368b3d4, current_thread() is 0x368b3d4, ebp is 0x1b2fb878, eip is 0x2c2d61 > 0x1b2fb878 _Debugger(5eccb4,1b2fb8ac,22fc76,0,13f2d7c)+170 > 0x1b2fb8c8 _panic(5e4c34,0,1b2fb908,514022,13f2d74)+1bd > 0x1b2fb948 _m_freem_list(1cf89b00,10,39e4604,d5324282,39e4600)+9a > 0x1b2fb968 0x2130a625(39e4600,10,1b2fb998,1,3113cc0) > 0x1b2fb988 0x2130a6fb(39e4600,866858,1,499,126b0000) > 0x1b2fb9a8 __ZNK8OSObject13taggedReleaseEPKv(39e4600,866858,0,5142b5,1)+20 > 0x1b2fb9d8 __ZN7OSArray15flushCollectionEv(3113cc0,0,1,0,268d2c0)+30 > 0x1b2fb9f8 __ZN7OSArray4freeEv(3113cc0,0,1,267ddc0,2761480)+36 > 0x1b2fba18 __ZNK8OSObject13taggedReleaseEPKv(3113cc0,0,268d2c4,514d2a,1)+20 > 0x1b2fba38 __ZNK8OSObject7releaseEv(3113cc0,0,1,0,268d2c0)+19 > 0x1b2fba58 __ZN5OSSet4freeEv(268d2c0,0,1,1,0)+3d > 0x1b2fba78 __ZNK8OSObject13taggedReleaseEPKv(268d2c0,0,1,1,0)+20 > 0x1b2fba98 __ZNK8OSObject7releaseEv(268d2c0,0,0,0,2614c40)+19 > 0x1b2fbb18 __ZN9IOService15terminateWorkerEm(7,8832fc,a034a068,167,0)+5f7 > 0x1b2fbb78 __ZN9IOService23scheduleTerminatePhase2Em(126b0000,7,7,ffffffff,268d2c0)+a1 > 0x1b2fbbf8 __ZN9IOService15terminatePhase1Em(126b0000,7,0,126b0000,310ea40)+38d > 0x1b2fbc18 __ZN9IOService9terminateEm(126b0000,3,30d3000,517d5d,30d3000)+1b > 0x1b2fbc48 __ZN11IOCatalogue5resetEv(30d3000,2675600,3a0e640,dc008006,3a0e640)+14c > 0x1b2fbc88 __ZN11IOCatalogue25terminateDriversForModuleEP8OSStringb(2675620,3a0e640,0,513af6,40b3e00)+8a > 0x1b2fbcb8 __ZN11IOCatalogue25terminateDriversForModuleEPKcb(2675620,3043800,0,5174e1,269a480)+41 > 0x1b2fbd08 __ZN6OSKext10removeKextEPS_b(40b3e00,1,5f1c38,3a154f0,114)+bc > 0x1b2fbd78 __ZN6OSKext13handleRequestEP4hostjPcjPS2_PjS3_S4_(84bb80,ff2,20e8d000,e0,1b2fbdcc)+3f2 > 0x1b2fbde8 _kext_request(84bb80,ff2,3192200,e0,3338688)+191 > 0x1b2fbe58 _host_priv_server_routine(2a40680,333866c,1ff178,2a406ac,26103d4)+1d9 > 0x1b2fbe98 _ipc_kobject_server(2a40600,1b2fbedc,30e3a40,1b2fbed0,2a40698)+f4 > 0x1b2fbef8 _ipc_kmsg_send(2a40600,0,0,0,3627964)+69 > 0x1b2fbf78 _mach_msg_overwrite_trap(3627968,0,0,0,560754)+112 > 0x1b2fbfc8 _thread_setuserstack(3627964,0,10,3627964,3627964)+195 > kdb_trap(): type 14, code 0, regs->eip 0x2124ac > Caught page fault trap, code = 0, pc = 2124ac > > > Which you have to admit beats this: > > > Debugger called: <panic> > Backtrace (CPU 0), Frame : Return Address (4 potential args on stack) > 0x1b2fb878 : 0x2277b7 (0x5eccb4 0x1b2fb8ac 0x22fc76 0x0) > 0x1b2fb8c8 : 0x4be8a3 (0x5e4c34 0x0 0x1b2fb908 0x514022) > 0x1b2fb948 : 0x2130a625 (0x1cf89b00 0x10 0x39e4604 0xd5324282) > 0x1b2fb968 : 0x2130a6fb (0x39e4600 0x10 0x1b2fb998 0x1) > 0x1b2fb988 : 0x513f32 (0x39e4600 0x866858 0x1 0x499) > 0x1b2fb9a8 : 0x514d88 (0x39e4600 0x866858 0x0 0x5142b5) > 0x1b2fb9d8 : 0x514ef0 (0x3113cc0 0x0 0x1 0x0) > 0x1b2fb9f8 : 0x513f32 (0x3113cc0 0x0 0x1 0x267ddc0) > 0x1b2fba18 : 0x513f4d (0x3113cc0 0x0 0x268d2c4 0x514d2a) > 0x1b2fba38 : 0x525b71 (0x3113cc0 0x0 0x1 0x0) > 0x1b2fba58 : 0x513f32 (0x268d2c0 0x0 0x1 0x1) > 0x1b2fba78 : 0x513f4d (0x268d2c0 0x0 0x1 0x1) > 0x1b2fba98 : 0x54f031 (0x268d2c0 0x0 0x0 0x0) > 0x1b2fbb18 : 0x54fc99 (0x7 0x8832fc 0xa034a068 0x167) > 0x1b2fbb78 : 0x55042f (0x126b0000 0x7 0x7 0xffffffff) > 0x1b2fbbf8 : 0x550457 (0x126b0000 0x7 0x0 0x126b0000) > 0x1b2fbc18 : 0x55d6c2 (0x126b0000 0x3 0x30d3000 0x517d5d) > 0x1b2fbc48 : 0x55da3c (0x30d3000 0x2675600 0x3a0e640 0xdc008006) > 0x1b2fbc88 : 0x55daf3 (0x2675620 0x3a0e640 0x0 0x513af6) > 0x1b2fbcb8 : 0x51e6d8 (0x2675620 0x3043800 0x0 0x5174e1) > 0x1b2fbd08 : 0x522cc8 (0x40b3e00 0x1 0x5f1c38 0x3a154f0) > 0x1b2fbd78 : 0x52abb0 (0x84bb80 0xff2 0x20e8d000 0xe0) > 0x1b2fbde8 : 0x24662e (0x84bb80 0xff2 0x3192200 0xe0) > 0x1b2fbe58 : 0x229b7f (0x2a40680 0x333866c 0x1ff178 0x2a406ac) > 0x1b2fbe98 : 0x21c4a4 (0x2a40600 0x1b2fbedc 0x30e3a40 0x1b2fbed0) > 0x1b2fbef8 : 0x222f04 (0x2a40600 0x0 0x0 0x0) > 0x1b2fbf78 : 0x2a41dc (0x3627968 0x0 0x0 0x0) > 0x1b2fbfc8 : 0x2b36cd (0x3627964 0x0 0x10 0x3627964) > Kernel Extensions in backtrace (with dependencies): > com.apple.iokit.IONetworkingFamily(1.8)@0x212ff000->0x21314fff > > > Drew _______________________________________________ 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: how do I enable MACH_KDB on 10.8.0 ?
by Andrew Gallatin 17 Nov '11

17 Nov '11

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=simple/simple; d=cs.duke.edu; s=mail; t=1321543216; bh=e3TQbMyTK0Xxc4hEYpfFlR8EuPSvMqZUKHirh6JQjQU=; h=Message-ID:Date:From:MIME-Version:To:CC:Subject:References: In-Reply-To:Content-Type:Content-Transfer-Encoding; b=myXVv6ind5s65WFXWUx2xdJht0floHNQx9Iyr/0BbLQd0FfsAAhifRMuED557gpiO Dv8tcHlUeQjBu1d5eZnVBVF4qVJ+Cs2z+n1RXx4ooOZZevN0XWLwSEqUX8As5o3Jn5 7RhjF8v2WhCx0u0UK7WzERQnGOcUnd09qodFo8TCDEnl5nOZzVIyXy8ZbWePb1KG1Q uI8XuJ856r52KwHrLMYmGth/k1d95NlpBIoMvQGBcfq43rYl69TUKpWc+sSCzd1XdF h39wEmyoFFW1qhCePqpuf3Y8UU/uplIssDinmyo+ii+UMrupi+ojCvChwObJ2+F59t c5mrE2oGEzFmg== User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:7.0.1) Gecko/20110929 Thunderbird/7.0.1 For my purposes, it seems to work. And I wasn't having much luck with the traditional remote dump methods in my non-Mac environment (running darwin under qemu-kvm to talk to a verilog simulator for a device which is still under development). The one thing I do have is a working "serial" console, so I figured I'd might as well use it ;) At any rate, for some reason, it wasn't picking up the debug= flags, so I had to hack the trap routine. I finally have a nice, somewhat meaningful stack trace: Which you have to admit beats this: Debugger called: <panic> Backtrace (CPU 0), Frame : Return Address (4 potential args on stack) 0x1b2fb878 : 0x2277b7 (0x5eccb4 0x1b2fb8ac 0x22fc76 0x0) 0x1b2fb8c8 : 0x4be8a3 (0x5e4c34 0x0 0x1b2fb908 0x514022) 0x1b2fb948 : 0x2130a625 (0x1cf89b00 0x10 0x39e4604 0xd5324282) 0x1b2fb968 : 0x2130a6fb (0x39e4600 0x10 0x1b2fb998 0x1) 0x1b2fb988 : 0x513f32 (0x39e4600 0x866858 0x1 0x499) 0x1b2fb9a8 : 0x514d88 (0x39e4600 0x866858 0x0 0x5142b5) 0x1b2fb9d8 : 0x514ef0 (0x3113cc0 0x0 0x1 0x0) 0x1b2fb9f8 : 0x513f32 (0x3113cc0 0x0 0x1 0x267ddc0) 0x1b2fba18 : 0x513f4d (0x3113cc0 0x0 0x268d2c4 0x514d2a) 0x1b2fba38 : 0x525b71 (0x3113cc0 0x0 0x1 0x0) 0x1b2fba58 : 0x513f32 (0x268d2c0 0x0 0x1 0x1) 0x1b2fba78 : 0x513f4d (0x268d2c0 0x0 0x1 0x1) 0x1b2fba98 : 0x54f031 (0x268d2c0 0x0 0x0 0x0) 0x1b2fbb18 : 0x54fc99 (0x7 0x8832fc 0xa034a068 0x167) 0x1b2fbb78 : 0x55042f (0x126b0000 0x7 0x7 0xffffffff) 0x1b2fbbf8 : 0x550457 (0x126b0000 0x7 0x0 0x126b0000) 0x1b2fbc18 : 0x55d6c2 (0x126b0000 0x3 0x30d3000 0x517d5d) 0x1b2fbc48 : 0x55da3c (0x30d3000 0x2675600 0x3a0e640 0xdc008006) 0x1b2fbc88 : 0x55daf3 (0x2675620 0x3a0e640 0x0 0x513af6) 0x1b2fbcb8 : 0x51e6d8 (0x2675620 0x3043800 0x0 0x5174e1) 0x1b2fbd08 : 0x522cc8 (0x40b3e00 0x1 0x5f1c38 0x3a154f0) 0x1b2fbd78 : 0x52abb0 (0x84bb80 0xff2 0x20e8d000 0xe0) 0x1b2fbde8 : 0x24662e (0x84bb80 0xff2 0x3192200 0xe0) 0x1b2fbe58 : 0x229b7f (0x2a40680 0x333866c 0x1ff178 0x2a406ac) 0x1b2fbe98 : 0x21c4a4 (0x2a40600 0x1b2fbedc 0x30e3a40 0x1b2fbed0) 0x1b2fbef8 : 0x222f04 (0x2a40600 0x0 0x0 0x0) 0x1b2fbf78 : 0x2a41dc (0x3627968 0x0 0x0 0x0) 0x1b2fbfc8 : 0x2b36cd (0x3627964 0x0 0x10 0x3627964) Kernel Extensions in backtrace (with dependencies): com.apple.iokit.IONetworkingFamily(1.8)@0x212ff000->0x21314fff Drew _______________________________________________ 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.… db{0}> tr thread 0x368b3d4, current_thread() is 0x368b3d4, ebp is 0x1b2fb878, eip is 0x2c2d61 0x1b2fb878 _Debugger(5eccb4,1b2fb8ac,22fc76,0,13f2d7c)+170 0x1b2fb8c8 _panic(5e4c34,0,1b2fb908,514022,13f2d74)+1bd 0x1b2fb948 _m_freem_list(1cf89b00,10,39e4604,d5324282,39e4600)+9a 0x1b2fb968 0x2130a625(39e4600,10,1b2fb998,1,3113cc0) 0x1b2fb988 0x2130a6fb(39e4600,866858,1,499,126b0000) 0x1b2fb9a8 __ZNK8OSObject13taggedReleaseEPKv(39e4600,866858,0,5142b5,1)+20 0x1b2fb9d8 __ZN7OSArray15flushCollectionEv(3113cc0,0,1,0,268d2c0)+30 0x1b2fb9f8 __ZN7OSArray4freeEv(3113cc0,0,1,267ddc0,2761480)+36 0x1b2fba18 __ZNK8OSObject13taggedReleaseEPKv(3113cc0,0,268d2c4,514d2a,1)+20 0x1b2fba38 __ZNK8OSObject7releaseEv(3113cc0,0,1,0,268d2c0)+19 0x1b2fba58 __ZN5OSSet4freeEv(268d2c0,0,1,1,0)+3d 0x1b2fba78 __ZNK8OSObject13taggedReleaseEPKv(268d2c0,0,1,1,0)+20 0x1b2fba98 __ZNK8OSObject7releaseEv(268d2c0,0,0,0,2614c40)+19 0x1b2fbb18 __ZN9IOService15terminateWorkerEm(7,8832fc,a034a068,167,0)+5f7 0x1b2fbb78 __ZN9IOService23scheduleTerminatePhase2Em(126b0000,7,7,ffffffff,268d2c0)+a1 0x1b2fbbf8 __ZN9IOService15terminatePhase1Em(126b0000,7,0,126b0000,310ea40)+38d 0x1b2fbc18 __ZN9IOService9terminateEm(126b0000,3,30d3000,517d5d,30d3000)+1b 0x1b2fbc48 __ZN11IOCatalogue5resetEv(30d3000,2675600,3a0e640,dc008006,3a0e640)+14c 0x1b2fbc88 __ZN11IOCatalogue25terminateDriversForModuleEP8OSStringb(2675620,3a0e640,0,513af6,40b3e00)+8a 0x1b2fbcb8 __ZN11IOCatalogue25terminateDriversForModuleEPKcb(2675620,3043800,0,5174e1,269a480)+41 0x1b2fbd08 __ZN6OSKext10removeKextEPS_b(40b3e00,1,5f1c38,3a154f0,114)+bc 0x1b2fbd78 __ZN6OSKext13handleRequestEP4hostjPcjPS2_PjS3_S4_(84bb80,ff2,20e8d000,e0,1b2fbdcc)+3f2 0x1b2fbde8 _kext_request(84bb80,ff2,3192200,e0,3338688)+191 0x1b2fbe58 _host_priv_server_routine(2a40680,333866c,1ff178,2a406ac,26103d4)+1d9 0x1b2fbe98 _ipc_kobject_server(2a40600,1b2fbedc,30e3a40,1b2fbed0,2a40698)+f4 0x1b2fbef8 _ipc_kmsg_send(2a40600,0,0,0,3627964)+69 0x1b2fbf78 _mach_msg_overwrite_trap(3627968,0,0,0,560754)+112 0x1b2fbfc8 _thread_setuserstack(3627964,0,10,3627964,3627964)+195 kdb_trap(): type 14, code 0, regs->eip 0x2124ac Caught page fault trap, code = 0, pc = 2124ac This email sent to site_archiver(a)lists.apple.com

1 0

Re: IOMalloc / IOFree while holding a mutex
by chas williams - CONTRACTOR 17 Nov '11

17 Nov '11

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Wed, 16 Nov 2011 09:35:44 -0800 Shantonu Sen <ssen(a)apple.com> wrote: > There's some layers of sublety here. You can do a lot of things at a technical >level. Certainly there are no assertions or consistency checks that >prevent you from holding an IOLock/lck_mtx_t. However if you are >writing a storage driver that is on the paging path, consider what >happens if you take the lock (or run under a command gate), attempt the >IOMalloc, which in turn needs to free up pages by paging out user data, >which issues an I/O, which tries to acquire the mutex previously >acquired? You can try to massage the code to accept an asyncReadWrite() >without taking any locks, so you don't end up deadlocking, but it >requires some thought. This is one of the reasons why memory >allocations on the I/O path is frowned upon. as a side note, you should be able to avoid this (aleast for an IOSCSIParallelInterfaceController) by using ReportHBASpecificTaskDataSize and GetHBADataPointer. atleast i gather that is why these members exist -- to avoid this situation described above. _______________________________________________ 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