From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Long Subject: Re: [PATCH 1/2] Move the pt_regs_offset struct definition from arch to common include file Date: Tue, 23 Jun 2015 09:48:19 -0400 Message-ID: <558963A3.4070503@linaro.org> References: <1434386579-6045-1-git-send-email-dave.long@linaro.org> <1434386579-6045-2-git-send-email-dave.long@linaro.org> <1434687583.23771.5.camel@ellerman.id.au> <5584234E.2040701@linaro.org> <1435030328.28070.5.camel@ellerman.id.au> Mime-Version: 1.0 Content-Transfer-Encoding: base64 Return-path: In-Reply-To: <1435030328.28070.5.camel@ellerman.id.au> List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: linuxppc-dev-bounces+glppe-linuxppc-embedded-2=m.gmane.org@lists.ozlabs.org Sender: "Linuxppc-dev" Content-Type: text/plain; charset="us-ascii"; format="flowed" To: Michael Ellerman Cc: linux-sh@vger.kernel.org, Will Deacon , Jan Willeke , linux-kernel@vger.kernel.org, Nikolay Borisov , Behan Webster , "H. Peter Anvin" , Richard Kuo , linux-s390@vger.kernel.org, Russell King , linux-hexagon@vger.kernel.org, x86@kernel.org, Ingo Molnar , mahesh@linux.vnet.ibm.com, Robert Richter , Kees Cook , Roland McGrath , Heiko Carstens , Martin Schwidefsky , Anton Blanchard , linux390@de.ibm.com, Thomas Gleixner , linux-arm-kernel@lists.infradead.org, Oleg Nesterov , Eric Paris , Andy Lutomirski T24gMDYvMjIvMTUgMjM6MzIsIE1pY2hhZWwgRWxsZXJtYW4gd3JvdGU6Cj4gT24gRnJpLCAyMDE1 LTA2LTE5IGF0IDEwOjEyIC0wNDAwLCBEYXZpZCBMb25nIHdyb3RlOgo+PiBPbiAwNi8xOS8xNSAw MDoxOSwgTWljaGFlbCBFbGxlcm1hbiB3cm90ZToKPj4+IE9uIE1vbiwgMjAxNS0wNi0xNSBhdCAx Mjo0MiAtMDQwMCwgRGF2aWQgTG9uZyB3cm90ZToKPj4+PiBGcm9tOiAiRGF2aWQgQS4gTG9uZyIg PGRhdmUubG9uZ0BsaW5hcm8ub3JnPgo+Pj4+Cj4+Pj4gVGhlIHB0X3JlZ3Nfb2Zmc2V0IHN0cnVj dHVyZSBpcyB1c2VkIGZvciBIQVZFX1JFR1NfQU5EX1NUQUNLX0FDQ0VTU19BUEkKPj4+PiAgICBm ZWF0dXJlIGFuZCBoYXMgaWRlbnRpY2FsIGRlZmluaXRpb25zIGluIGZvdXIgZGlmZmVyZW50IGFy Y2ggcHRyYWNlLmgKPj4+PiBpbmNsdWRlIGZpbGVzLiBJdCBzZWVtcyB1bmxpa2VseSB0aGF0IGRl ZmluaXRpb24gd291bGQgZXZlciBuZWVkIHRvIGJlCj4+Pj4gY2hhbmdlZCByZWdhcmRsZXNzIG9m IGFyY2hpdGVjdHVyZSBzbyBsZXRzIG1vdmUgaXQgaW50bwo+Pj4+IGluY2x1ZGUvbGludXgvcHRy YWNlLmguCj4+Pj4KPj4+PiBTaWduZWQtb2ZmLWJ5OiBEYXZpZCBBLiBMb25nIDxkYXZlLmxvbmdA bGluYXJvLm9yZz4KPj4+PiAtLS0KPj4+PiAgICBhcmNoL3Bvd2VycGMva2VybmVsL3B0cmFjZS5j IHwgNSAtLS0tLQo+Pj4KPj4+IEJ1aWx0IGFuZCBib290ZWQgb24gcG93ZXJwYywgYnV0IGlzIHRo ZXJlIGFuIGVhc3kgd2F5IHRvIGFjdHVhbGx5IHRlc3QgdGhlIGNvZGUKPj4+IHBhdGhzIGluIHF1 ZXN0aW9uPwo+Pj4KPj4KPj4gVGhlcmUgaXMgYW4gZWFzeSB3YXkgdG8gInNtb2tlIHRlc3QiIGl0 IG9uIGFsbCBhcmNoaWVjdHVyZXMgdGhhdCBhbHNvCj4+IGltcGxlbWVudCBrcHJvYmVzICh3aGlj aCBwb3dlcnBjIGRvZXMpLiAgSWYgSSdtIHVuZGVyc3RhbmRpbmcgdGhlCj4+IHBvd2VycGMgY29k ZSBjb3JyZWN0bHkgKFdSVCByZWdpc3RlciBuYW1pbmcgY29udmVudGlvbnMpIGp1c3QgZG8gdGhl Cj4+IGZvbGxvd2luZzoKPj4KPj4gY2QgL3N5cy9rZXJuZWwvZGVidWcvdHJhY2luZwo+PiBlY2hv ICdwIGRvX2ZvcmsgJWdwcjAnID4ga3Byb2JlX2V2ZW50cwo+PiBlY2hvIDEgPiBldmVudHMva3By b2Jlcy9lbmFibGUKPj4gbHMKPj4gY2F0IHRyYWNlCj4+IGVjaG8gMCA+IGV2ZW50cy9rcHJvYmVz L2VuYWJsZQo+Pgo+PiBFdmVyeSBmb3JrKCkgY2FsbCBkb25lIG9uIHRoZSBzeXN0ZW0gYmV0d2Vl biB0aG9zZSB0d28gZWNobyBjb21tYW5kcwo+PiAoaGVuY2UgdGhlICJscyIpIHNob3VsZCBhcHBl bmQgYSBsaW5lIHRvIHRoZSB0cmFjZSBmaWxlLiAgRm9yIGEgbW9yZQo+PiBleGhhdXN0aXZlIHRl c3Qgb25lIGNvdWxkIHJlcGVhdCB0aGlzIHNlcXVlbmNlIGZvciBldmVyeSByZWdpc3RlciBpbiB0 aGUKPj4gYXJjaGl0ZWN0dXJlLgo+Cj4gT0ssIHNvIEkgd2VudCB0aGUgd2hvbGUgaG9nIGFuZCBk aWQ6Cj4KPiAkIGVjaG8gJ3AgZG9fZm9yayAlZ3ByMCAlZ3ByMSAlZ3ByMiAlZ3ByMyAlZ3ByNCAl Z3ByNSAlZ3ByNiAlZ3ByNyAlZ3ByOCAlZ3ByOSAlZ3ByMTAgJWdwcjExICVncHIxMiAlZ3ByMTMg JWdwcjE0ICVncHIxNSAlZ3ByMTYgJWdwcjE3ICVncHIxOCAlZ3ByMTkgJWdwcjIwICVncHIyMSAl Z3ByMjIgJWdwcjIzICVncHIyNCAlZ3ByMjUgJWdwcjI2ICVncHIyNyAlZ3ByMjggJWdwcjI5ICVn cHIzMCAlZ3ByMzEgJW5pcCAlbXNyICVjdHIgJWxpbmsgJXhlciAlY2NyICVzb2Z0ZSAldHJhcCAl ZGFyICVkc2lzcicgPiBrcHJvYmVfZXZlbnRzCj4KPiBBbmQgSSBnZXQ6Cj4KPiAgICAgICAgICAg ICAgYmFzaC0yMDU3ICBbMDAxXSBkLi4uICAgNTM1LjQzMzk0MTogcF9kb19mb3JrXzA6IChkb19m b3JrKzB4OC8weDQ5MCkgYXJnMT0weGMwMDAwMDAwMDAwMDk0ZDAgYXJnMj0weGMwMDAwMDAxZmJl OWJlMzAgYXJnMz0weGMwMDAwMDAwMDExMzNiYjggYXJnND0weDEyMDAwMTEgYXJnNT0weDAgYXJn Nj0weDAgYXJnNz0weDAgYXJnOD0weDNmZmY3Yzg4NTk0MCBhcmc5PTB4MSBhcmcxMD0weGMwMDAw MDAxZmJlOWJlYTAgYXJnMTE9MHgwIGFyZzEyPTB4YzAxIGFyZzEzPTB4YzAwMDAwMDAwMDAwOTRj OCBhcmcxND0weGMwMDAwMDAwMGZkYzA0ODAgYXJnMTU9MHgwIGFyZzE2PTB4MjIwMDAwMDAgYXJn MTc9MHgxMDE2ZDZlOCBhcmcxOD0weDAgYXJnMTk9MHg0NDAwMDAwMCBhcmcyMD0weDAgYXJnMjE9 MHgxMDAzN2M4MjIwOCBhcmcyMj0weDEwMTdiMDA4IGFyZzIzPTB4MTAxNDNkMTggYXJnMjQ9MHgx MDE3ODg1NCBhcmcyNT0weDEwMTQ0ZjkwIGFyZzI2PTB4MTAwMzdjODIxZTggYXJnMjc9MHgwIGFy ZzI4PTB4MCBhcmcyOT0weDAgYXJnMzA9MHgwIGFyZzMxPTB4ODA5IGFyZzMyPTB4M2ZmZmY3ODhj MDEwIGFyZzMzPTB4YzAwMDAwMDAwMDBhN2ZlOCBhcmczND0weDgwMDAwMDAwMDAwMjkwMzMgYXJn MzU9MHhjMDAwMDAwMDAwMDA5NGM4IGFyZzM2PTB4YzAwMDAwMDAwMDAwOTRkMCBhcmczNz0weDAg YXJnMzg9MHg0MjIyMjg0NCBhcmczOT0weDEgYXJnNDA9MHg3MDAgYXJnNDE9MHhjMDAwMDAwMWZi ZTliZDUwIGFyZzQyPTB4YzAwMDAwMDFmYmU5YmQzMAo+Cj4gV2hpY2ggaXMgdWdseSBhcyBoZWxs LCBidXQgYXBwZWFycyB1bmNoYW5nZWQgc2luY2UgYmVmb3JlIHlvdXIgcGF0Y2guCj4KCkV4Y2Vs bGVudC4gIE1hbnkgdGhhbmtzLgoKPiBJIHRha2UgaXQgaXQncyBleHBlY3RlZCB0aGF0IHRoZSBu YW1lcyBhcmUgbm90IGRlY29kZWQgaW4gdGhlIG91dHB1dD8KPgoKWWVzLgoKPiBBbHNvIEkgd29u ZGVyIHdoeSB3ZSBjaG9vc2UgImdwciIgd2hlbiAiciIgaXMgdGhlIG1vcmUgdXN1YWwgcHJlZml4 IG9uIHBvd2VycGMuCj4gSSBndWVzcyB3ZSBjYW4gYWRkIG5ldyBhbGlhc2VzIHRvIHRoZSB0YWJs ZS4KPgoKWWVhaCBJIGNhbid0IGFuc3dlciB0aGF0LCB0aGlzIGlzIGp1c3Qgd2hhdCB0aGUgcHJl ZXhpc3RpbmcgY29kZSBpcyAKd3JpdHRlbiB0byBkby4gSSBiZWxpZXZlIHlvdSBjb3VsZCBhZGQg YWxpYXNlcyB0byB0aGUgdGFibGUsIHBlcmhhcHMgYXMgCmEgc3RlcCBpbiBtaWdyYXRpbmcgdG8g c3VwcG9ydGluZyBvbmx5IHRoZSBtb3JlIGNvbW1vbiBuYW1pbmcuICBUaGUgCnJldmVyc2UgdHJh bnNsYXRpb24gd291bGQgaGF2ZSB0byByZXR1cm4gb25lIG9yIHRoZSBvdGhlciB0aG91Z2guCgot ZGwKCgpfX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fXwpMaW51 eHBwYy1kZXYgbWFpbGluZyBsaXN0CkxpbnV4cHBjLWRldkBsaXN0cy5vemxhYnMub3JnCmh0dHBz Oi8vbGlzdHMub3psYWJzLm9yZy9saXN0aW5mby9saW51eHBwYy1kZXY= From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Long Date: Tue, 23 Jun 2015 13:48:19 +0000 Subject: Re: [PATCH 1/2] Move the pt_regs_offset struct definition from arch to common include file Message-Id: <558963A3.4070503@linaro.org> List-Id: References: <1434386579-6045-1-git-send-email-dave.long@linaro.org> <1434386579-6045-2-git-send-email-dave.long@linaro.org> <1434687583.23771.5.camel@ellerman.id.au> <5584234E.2040701@linaro.org> <1435030328.28070.5.camel@ellerman.id.au> In-Reply-To: <1435030328.28070.5.camel@ellerman.id.au> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-arm-kernel@lists.infradead.org On 06/22/15 23:32, Michael Ellerman wrote: > On Fri, 2015-06-19 at 10:12 -0400, David Long wrote: >> On 06/19/15 00:19, Michael Ellerman wrote: >>> On Mon, 2015-06-15 at 12:42 -0400, David Long wrote: >>>> From: "David A. Long" >>>> >>>> The pt_regs_offset structure is used for HAVE_REGS_AND_STACK_ACCESS_API >>>> feature and has identical definitions in four different arch ptrace.h >>>> include files. It seems unlikely that definition would ever need to be >>>> changed regardless of architecture so lets move it into >>>> include/linux/ptrace.h. >>>> >>>> Signed-off-by: David A. Long >>>> --- >>>> arch/powerpc/kernel/ptrace.c | 5 ----- >>> >>> Built and booted on powerpc, but is there an easy way to actually test the code >>> paths in question? >>> >> >> There is an easy way to "smoke test" it on all archiectures that also >> implement kprobes (which powerpc does). If I'm understanding the >> powerpc code correctly (WRT register naming conventions) just do the >> following: >> >> cd /sys/kernel/debug/tracing >> echo 'p do_fork %gpr0' > kprobe_events >> echo 1 > events/kprobes/enable >> ls >> cat trace >> echo 0 > events/kprobes/enable >> >> Every fork() call done on the system between those two echo commands >> (hence the "ls") should append a line to the trace file. For a more >> exhaustive test one could repeat this sequence for every register in the >> architecture. > > OK, so I went the whole hog and did: > > $ echo 'p do_fork %gpr0 %gpr1 %gpr2 %gpr3 %gpr4 %gpr5 %gpr6 %gpr7 %gpr8 %gpr9 %gpr10 %gpr11 %gpr12 %gpr13 %gpr14 %gpr15 %gpr16 %gpr17 %gpr18 %gpr19 %gpr20 %gpr21 %gpr22 %gpr23 %gpr24 %gpr25 %gpr26 %gpr27 %gpr28 %gpr29 %gpr30 %gpr31 %nip %msr %ctr %link %xer %ccr %softe %trap %dar %dsisr' > kprobe_events > > And I get: > > bash-2057 [001] d... 535.433941: p_do_fork_0: (do_fork+0x8/0x490) arg1=0xc0000000000094d0 arg2=0xc0000001fbe9be30 arg3=0xc000000001133bb8 arg4=0x1200011 arg5=0x0 arg6=0x0 arg7=0x0 arg8=0x3fff7c885940 arg9=0x1 arg10=0xc0000001fbe9bea0 arg11=0x0 arg12=0xc01 arg13=0xc0000000000094c8 arg14=0xc00000000fdc0480 arg15=0x0 arg16=0x22000000 arg17=0x1016d6e8 arg18=0x0 arg19=0x44000000 arg20=0x0 arg21=0x10037c82208 arg22=0x1017b008 arg23=0x10143d18 arg24=0x10178854 arg25=0x10144f90 arg26=0x10037c821e8 arg27=0x0 arg28=0x0 arg29=0x0 arg30=0x0 arg31=0x809 arg32=0x3ffff788c010 arg33=0xc0000000000a7fe8 arg34=0x8000000000029033 arg35=0xc0000000000094c8 arg36=0xc0000000000094d0 arg37=0x0 arg38=0x42222844 arg39=0x1 arg40=0x700 arg41=0xc0000001fbe9bd50 arg42=0xc0000001fbe9bd30 > > Which is ugly as hell, but appears unchanged since before your patch. > Excellent. Many thanks. > I take it it's expected that the names are not decoded in the output? > Yes. > Also I wonder why we choose "gpr" when "r" is the more usual prefix on powerpc. > I guess we can add new aliases to the table. > Yeah I can't answer that, this is just what the preexisting code is written to do. I believe you could add aliases to the table, perhaps as a step in migrating to supporting only the more common naming. The reverse translation would have to return one or the other though. -dl From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-qk0-f171.google.com (mail-qk0-f171.google.com [209.85.220.171]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by lists.ozlabs.org (Postfix) with ESMTPS id D50FE1A001E for ; Tue, 23 Jun 2015 23:48:23 +1000 (AEST) Received: by qkbp125 with SMTP id p125so4939548qkb.2 for ; Tue, 23 Jun 2015 06:48:21 -0700 (PDT) Message-ID: <558963A3.4070503@linaro.org> Date: Tue, 23 Jun 2015 09:48:19 -0400 From: David Long MIME-Version: 1.0 To: Michael Ellerman CC: "H. Peter Anvin" , Andy Lutomirski , Anton Blanchard , Behan Webster , Benjamin Herrenschmidt , Eric Paris , Heiko Carstens , Ingo Molnar , Jan Willeke , Kees Cook , Martin Schwidefsky , Nikolay Borisov , Oleg Nesterov , Paul Mackerras , Richard Kuo , Robert Richter , Roland McGrath , Russell King , Tejun Heo , Thomas Gleixner , Will Deacon , linux-arm-kernel@lists.infradead.org, linux-hexagon@vger.kernel.org, linux-kernel@vger.kernel.org, linux-s390@vger.kernel.org, linux-sh@vger.kernel.org, linux390@de.ibm.com, linuxppc-dev@lists.ozlabs.org, x86@kernel.org, mahesh@linux.vnet.ibm.com Subject: Re: [PATCH 1/2] Move the pt_regs_offset struct definition from arch to common include file References: <1434386579-6045-1-git-send-email-dave.long@linaro.org> <1434386579-6045-2-git-send-email-dave.long@linaro.org> <1434687583.23771.5.camel@ellerman.id.au> <5584234E.2040701@linaro.org> <1435030328.28070.5.camel@ellerman.id.au> In-Reply-To: <1435030328.28070.5.camel@ellerman.id.au> Content-Type: text/plain; charset=UTF-8; format=flowed List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On 06/22/15 23:32, Michael Ellerman wrote: > On Fri, 2015-06-19 at 10:12 -0400, David Long wrote: >> On 06/19/15 00:19, Michael Ellerman wrote: >>> On Mon, 2015-06-15 at 12:42 -0400, David Long wrote: >>>> From: "David A. Long" >>>> >>>> The pt_regs_offset structure is used for HAVE_REGS_AND_STACK_ACCESS_API >>>> feature and has identical definitions in four different arch ptrace.h >>>> include files. It seems unlikely that definition would ever need to be >>>> changed regardless of architecture so lets move it into >>>> include/linux/ptrace.h. >>>> >>>> Signed-off-by: David A. Long >>>> --- >>>> arch/powerpc/kernel/ptrace.c | 5 ----- >>> >>> Built and booted on powerpc, but is there an easy way to actually test the code >>> paths in question? >>> >> >> There is an easy way to "smoke test" it on all archiectures that also >> implement kprobes (which powerpc does). If I'm understanding the >> powerpc code correctly (WRT register naming conventions) just do the >> following: >> >> cd /sys/kernel/debug/tracing >> echo 'p do_fork %gpr0' > kprobe_events >> echo 1 > events/kprobes/enable >> ls >> cat trace >> echo 0 > events/kprobes/enable >> >> Every fork() call done on the system between those two echo commands >> (hence the "ls") should append a line to the trace file. For a more >> exhaustive test one could repeat this sequence for every register in the >> architecture. > > OK, so I went the whole hog and did: > > $ echo 'p do_fork %gpr0 %gpr1 %gpr2 %gpr3 %gpr4 %gpr5 %gpr6 %gpr7 %gpr8 %gpr9 %gpr10 %gpr11 %gpr12 %gpr13 %gpr14 %gpr15 %gpr16 %gpr17 %gpr18 %gpr19 %gpr20 %gpr21 %gpr22 %gpr23 %gpr24 %gpr25 %gpr26 %gpr27 %gpr28 %gpr29 %gpr30 %gpr31 %nip %msr %ctr %link %xer %ccr %softe %trap %dar %dsisr' > kprobe_events > > And I get: > > bash-2057 [001] d... 535.433941: p_do_fork_0: (do_fork+0x8/0x490) arg1=0xc0000000000094d0 arg2=0xc0000001fbe9be30 arg3=0xc000000001133bb8 arg4=0x1200011 arg5=0x0 arg6=0x0 arg7=0x0 arg8=0x3fff7c885940 arg9=0x1 arg10=0xc0000001fbe9bea0 arg11=0x0 arg12=0xc01 arg13=0xc0000000000094c8 arg14=0xc00000000fdc0480 arg15=0x0 arg16=0x22000000 arg17=0x1016d6e8 arg18=0x0 arg19=0x44000000 arg20=0x0 arg21=0x10037c82208 arg22=0x1017b008 arg23=0x10143d18 arg24=0x10178854 arg25=0x10144f90 arg26=0x10037c821e8 arg27=0x0 arg28=0x0 arg29=0x0 arg30=0x0 arg31=0x809 arg32=0x3ffff788c010 arg33=0xc0000000000a7fe8 arg34=0x8000000000029033 arg35=0xc0000000000094c8 arg36=0xc0000000000094d0 arg37=0x0 arg38=0x42222844 arg39=0x1 arg40=0x700 arg41=0xc0000001fbe9bd50 arg42=0xc0000001fbe9bd30 > > Which is ugly as hell, but appears unchanged since before your patch. > Excellent. Many thanks. > I take it it's expected that the names are not decoded in the output? > Yes. > Also I wonder why we choose "gpr" when "r" is the more usual prefix on powerpc. > I guess we can add new aliases to the table. > Yeah I can't answer that, this is just what the preexisting code is written to do. I believe you could add aliases to the table, perhaps as a step in migrating to supporting only the more common naming. The reverse translation would have to return one or the other though. -dl From mboxrd@z Thu Jan 1 00:00:00 1970 From: dave.long@linaro.org (David Long) Date: Tue, 23 Jun 2015 09:48:19 -0400 Subject: [PATCH 1/2] Move the pt_regs_offset struct definition from arch to common include file In-Reply-To: <1435030328.28070.5.camel@ellerman.id.au> References: <1434386579-6045-1-git-send-email-dave.long@linaro.org> <1434386579-6045-2-git-send-email-dave.long@linaro.org> <1434687583.23771.5.camel@ellerman.id.au> <5584234E.2040701@linaro.org> <1435030328.28070.5.camel@ellerman.id.au> Message-ID: <558963A3.4070503@linaro.org> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On 06/22/15 23:32, Michael Ellerman wrote: > On Fri, 2015-06-19 at 10:12 -0400, David Long wrote: >> On 06/19/15 00:19, Michael Ellerman wrote: >>> On Mon, 2015-06-15 at 12:42 -0400, David Long wrote: >>>> From: "David A. Long" >>>> >>>> The pt_regs_offset structure is used for HAVE_REGS_AND_STACK_ACCESS_API >>>> feature and has identical definitions in four different arch ptrace.h >>>> include files. It seems unlikely that definition would ever need to be >>>> changed regardless of architecture so lets move it into >>>> include/linux/ptrace.h. >>>> >>>> Signed-off-by: David A. Long >>>> --- >>>> arch/powerpc/kernel/ptrace.c | 5 ----- >>> >>> Built and booted on powerpc, but is there an easy way to actually test the code >>> paths in question? >>> >> >> There is an easy way to "smoke test" it on all archiectures that also >> implement kprobes (which powerpc does). If I'm understanding the >> powerpc code correctly (WRT register naming conventions) just do the >> following: >> >> cd /sys/kernel/debug/tracing >> echo 'p do_fork %gpr0' > kprobe_events >> echo 1 > events/kprobes/enable >> ls >> cat trace >> echo 0 > events/kprobes/enable >> >> Every fork() call done on the system between those two echo commands >> (hence the "ls") should append a line to the trace file. For a more >> exhaustive test one could repeat this sequence for every register in the >> architecture. > > OK, so I went the whole hog and did: > > $ echo 'p do_fork %gpr0 %gpr1 %gpr2 %gpr3 %gpr4 %gpr5 %gpr6 %gpr7 %gpr8 %gpr9 %gpr10 %gpr11 %gpr12 %gpr13 %gpr14 %gpr15 %gpr16 %gpr17 %gpr18 %gpr19 %gpr20 %gpr21 %gpr22 %gpr23 %gpr24 %gpr25 %gpr26 %gpr27 %gpr28 %gpr29 %gpr30 %gpr31 %nip %msr %ctr %link %xer %ccr %softe %trap %dar %dsisr' > kprobe_events > > And I get: > > bash-2057 [001] d... 535.433941: p_do_fork_0: (do_fork+0x8/0x490) arg1=0xc0000000000094d0 arg2=0xc0000001fbe9be30 arg3=0xc000000001133bb8 arg4=0x1200011 arg5=0x0 arg6=0x0 arg7=0x0 arg8=0x3fff7c885940 arg9=0x1 arg10=0xc0000001fbe9bea0 arg11=0x0 arg12=0xc01 arg13=0xc0000000000094c8 arg14=0xc00000000fdc0480 arg15=0x0 arg16=0x22000000 arg17=0x1016d6e8 arg18=0x0 arg19=0x44000000 arg20=0x0 arg21=0x10037c82208 arg22=0x1017b008 arg23=0x10143d18 arg24=0x10178854 arg25=0x10144f90 arg26=0x10037c821e8 arg27=0x0 arg28=0x0 arg29=0x0 arg30=0x0 arg31=0x809 arg32=0x3ffff788c010 arg33=0xc0000000000a7fe8 arg34=0x8000000000029033 arg35=0xc0000000000094c8 arg36=0xc0000000000094d0 arg37=0x0 arg38=0x42222844 arg39=0x1 arg40=0x700 arg41=0xc0000001fbe9bd50 arg42=0xc0000001fbe9bd30 > > Which is ugly as hell, but appears unchanged since before your patch. > Excellent. Many thanks. > I take it it's expected that the names are not decoded in the output? > Yes. > Also I wonder why we choose "gpr" when "r" is the more usual prefix on powerpc. > I guess we can add new aliases to the table. > Yeah I can't answer that, this is just what the preexisting code is written to do. I believe you could add aliases to the table, perhaps as a step in migrating to supporting only the more common naming. The reverse translation would have to return one or the other though. -dl