From: Alexey Kardashevskiy <aik@ozlabs.ru>
To: David Gibson <david@gibson.dropbear.id.au>
Cc: armbru@redhat.com, qemu-ppc@nongnu.org, qemu-devel@nongnu.org,
Suraj Jitindar Singh <sjitindarsingh@gmail.com>,
dgilbert@redhat.com
Subject: Re: [Qemu-devel] [PATCH 2/2] ppc: Add dump-stack implementation
Date: Thu, 2 May 2019 13:47:32 +1000 [thread overview]
Message-ID: <c0917aef-79cc-f42c-2f4c-6d26da9657d6@ozlabs.ru> (raw)
Message-ID: <20190502034732.GTyA9yJSADPSCHln9vCbprOeyIW21uY2wxbXsEqwOc8@z> (raw)
In-Reply-To: <20190502004346.GC13618@umbus.fritz.box>
On 02/05/2019 10:43, David Gibson wrote:
> On Wed, May 01, 2019 at 07:48:48PM +1000, Alexey Kardashevskiy wrote:
>>
>>
>> On 01/05/2019 15:35, Suraj Jitindar Singh wrote:
>>> The monitor function dump-stack is used to dump the stack for a cpu.
>>> This can be useful for debugging purposes when the stack cannot be
>>> dumped by another means.
>>>
>>> Add a ppc implementation ppc_cpu_dump_stack().
>>> The stack pointer is stored in R1 with the back pointer at offset 0 and
>>> the link register at offset 2.
>>> Also dump the registers from the stack frame if the marker "regshere" is
>>> found.
>>
>> Is this a Linux only marker? ABI does not mentioned this.
>>
>>> This only dumps the kernel stack, stopping if a non-kernel address is
>>> found in the stack.
>>
>> Why enforce this limit?
>
> It's also making a Linux specific assumption about addresses.
It does not have to if it used ppc_cpu_get_phys_page_debug(), the only
linux specific left is that "regshere" marker, otherwise it could work
for AIX or FreeBSD (obviously without the exception frame).
>
>>
>>>
>>> Sample output:
>>> (qemu) dump-stack
>>> sp: 0xc00000007bfc5690 lr: 0xc0000000000974b8
>>> sp: 0xc00000007bfc56f0 lr: 0xc00000000065aab4
>>> sp: 0xc00000007bfc5720 lr: 0xc00000000065ab04
>>> sp: 0xc00000007bfc5740 lr: 0xc0000000000c29b8
>>> sp: 0xc00000007bfc57b0 lr: 0xc0000000000bc9e8
>>> sp: 0xc00000007bfc57e0 lr: 0xc0000000000bd584
>>> sp: 0xc00000007bfc5800 lr: 0xc0000000000bee14
>>> sp: 0xc00000007bfc5ac0 lr: 0xc0000000000c2100
>>> sp: 0xc00000007bfc5c60 lr: 0xc000000000029460
>>> sp: 0xc00000007bfc5ca0 lr: 0xc00000000010b5e8
>>> sp: 0xc00000007bfc5d00 lr: 0xc000000000105f34
>>> trap : 0x0000000000000700
>>> pc : 0xc000000000104490
>>> msr : 0x9000000002843003
>>> lr : 0xc000000000103ffc
>>> gpr 0: 0x0000000000000001
>>> gpr 1: 0xc00000005051f530
>>> gpr 2: 0xc000000001088200
>>> gpr 3: 0x0000000000000001
>>> gpr 4: 0xc000000032d60000
>>> gpr 5: 0xc0000000014b8f00
>>> gpr 6: 0x0000000000c835e0
>>> gpr 7: 0x0000000000000000
>>> gpr 8: 0x0000000000000000
>>> gpr 9: 0xc000000032f00000
>>> gpr10: 0x9000000002803033
>>> gpr11: 0xc000000000b60f00
>>> gpr12: 0x0000000000002000
>>> gpr13: 0xc000000001250000
>>> gpr14: 0x0000000000000000
>>> gpr15: 0x0000000000000008
>>> gpr16: 0x0000000000000000
>>> gpr17: 0xc00000000114f790
>>> gpr18: 0x00000000ffffffff
>>> gpr19: 0xc00000005051f8e8
>>> gpr20: 0x0000000000000001
>>> gpr21: 0x0000000000000000
>>> gpr22: 0x0000000000000001
>>> gpr23: 0x0000000000000001
>>> gpr24: 0x0000000000000001
>>> gpr25: 0xc0000000014b8f70
>>> gpr26: 0x0000000000000000
>>> gpr27: 0x0000000000000001
>>> gpr28: 0x0000000000000001
>>> gpr29: 0x0000000000000000
>>> gpr30: 0xc0000000014b8f00
>>> gpr31: 0xc0000000014b8f00
>>
>> Looks bulky, using the "info registers" format would make sense here.
>>
>>
>>> sp: 0xc00000005051f530 lr: 0x0000000000000000
>>> sp: 0xc00000005051f600 lr: 0xc000000000103ffc
>>> sp: 0xc00000005051f670 lr: 0xc0000000000f60a8
>>> sp: 0xc00000005051f850 lr: 0xc0000000000f18c0
>>> sp: 0xc00000005051fa10 lr: 0xc0000000000f5184
>>> sp: 0xc00000005051fae0 lr: 0xc0000000000ddf54
>>> sp: 0xc00000005051fb00 lr: 0xc0000000000dab9c
>>> sp: 0xc00000005051fb90 lr: 0xc0000000000cbf88
>>> sp: 0xc00000005051fd00 lr: 0xc0000000003e7480
>>> sp: 0xc00000005051fdb0 lr: 0xc0000000003e7ce4
>>> sp: 0xc00000005051fe00 lr: 0xc0000000003e7d88
>>> sp: 0xc00000005051fe20 lr: 0xc00000000000b3a4
>>> trap : 0x0000000000000c01
>>> pc : 0x00007fffa6c9d8d0
>>> msr : 0x900000000280f033
>>> lr : 0x0000000010090f40
>>> gpr 0: 0x0000000000000036
>>> gpr 1: 0x00007fffa62fdd70
>>> gpr 2: 0x00007fffa6d57300
>>> gpr 3: 0x000000000000000d
>>> gpr 4: 0x000000002000ae80
>>> gpr 5: 0x0000000000000000
>>> gpr 6: 0x0000000000000537
>>> gpr 7: 0x0000000000000000
>>> gpr 8: 0x000000000000000d
>>> gpr 9: 0x0000000000000000
>>> gpr10: 0x0000000000000000
>>> gpr11: 0x0000000000000000
>>> gpr12: 0x0000000000000000
>>> gpr13: 0x00007fffa6306380
>>> gpr14: 0x0000000000000000
>>> gpr15: 0x0000000000000001
>>> gpr16: 0x0000000039ba6928
>>> gpr17: 0x0000000000000000
>>> gpr18: 0x0000000000000000
>>> gpr19: 0x00007fffa6d702f0
>>> gpr20: 0x00007fffa62fddf0
>>> gpr21: 0x0000000000000080
>>> gpr22: 0x0000000000000004
>>> gpr23: 0x0000000000000000
>>> gpr24: 0x0000000010ac85c0
>>> gpr25: 0x0000000000000008
>>> gpr26: 0x00007fffa62fde10
>>> gpr27: 0x0000000000000000
>>> gpr28: 0x0000000000000002
>>> gpr29: 0x0000000000000000
>>> gpr30: 0x0000000039ba6900
>>> gpr31: 0x0000000010ac85c0
>>> sp: 0x00007fffa62fdd70
>>>
>>> Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
>>> ---
>>> target/ppc/cpu.h | 1 +
>>> target/ppc/translate.c | 60 +++++++++++++++++++++++++++++++++++++++++
>>> target/ppc/translate_init.inc.c | 1 +
>>> 3 files changed, 62 insertions(+)
>>>
>>> diff --git a/target/ppc/cpu.h b/target/ppc/cpu.h
>>> index 5e7cf54b2f..28c4dffca1 100644
>>> --- a/target/ppc/cpu.h
>>> +++ b/target/ppc/cpu.h
>>> @@ -1284,6 +1284,7 @@ struct PPCVirtualHypervisorClass {
>>> void ppc_cpu_do_interrupt(CPUState *cpu);
>>> bool ppc_cpu_exec_interrupt(CPUState *cpu, int int_req);
>>> void ppc_cpu_dump_state(CPUState *cpu, FILE *f, int flags);
>>> +void ppc_cpu_dump_stack(CPUState *cpu, FILE *f);
>>> void ppc_cpu_dump_statistics(CPUState *cpu, int flags);
>>> hwaddr ppc_cpu_get_phys_page_debug(CPUState *cpu, vaddr addr);
>>> int ppc_cpu_gdb_read_register(CPUState *cpu, uint8_t *buf, int reg);
>>> diff --git a/target/ppc/translate.c b/target/ppc/translate.c
>>> index 8d08625c33..b162998ce7 100644
>>> --- a/target/ppc/translate.c
>>> +++ b/target/ppc/translate.c
>>> @@ -7705,6 +7705,66 @@ void ppc_cpu_dump_state(CPUState *cs, FILE *f, int flags)
>>> #undef RFPL
>>> }
>>>
>>> +struct ppc_pt_regs {
>>> + unsigned long gpr[32];
>>> + unsigned long nip;
>>> + unsigned long msr;
>>> + unsigned long orig_gpr3;
>>> + unsigned long ctr;
>>> + unsigned long link;
>>> + unsigned long xer;
>>> + unsigned long ccr;
>>> + unsigned long softe;
>>> + unsigned long trap;
>>> + unsigned long dar;
>>> + unsigned long dsisr;
>>> + unsigned long result;
>>> +};
>>> +
>>> +void ppc_cpu_dump_stack(CPUState *cs, FILE *f)
>>> +{
>>> +#if defined(TARGET_PPC64)
>>> + PowerPCCPU *cpu = POWERPC_CPU(cs);
>>> + CPUPPCState *env = &cpu->env;
>>> + uint64_t sp, next_sp, lr, buf[4];
>>
>> These are hwaddr really.
>>
>>> +
>>> + /* stack pointer stored in r1 */
>>> + sp = env->gpr[1];
>>> +
>>> + while (sp && (sp & (0xCUL << 60))) {
>>> + uint64_t marker = 0UL;
>>
>> sp = ppc_cpu_get_phys_page_debug(cs, sp) | (sp & ~TARGET_PAGE_MASK);
>>
>> and finish the loop when ppc_cpu_get_phys_page_debug returns -1?
>>
>>> +
>>> + /* read and print LR */
>>> + cpu_physical_memory_read(sp & ~(0xCUL << 60), buf, sizeof(*buf) * 4);
>>
>> and s/ & ~(0xCUL << 60)//
>>
>>> + next_sp = buf[0];
>>> + lr = buf[2];
>>
>> These two need to be converted from guest endian. For a BE guest, I see:
>>
>> sp: 0x000000007e582ff0 lr: 0xe4e60a00000000c0
>> sp: 0xffffffffffffffff lr: 0x0000600000006000
>> sp: 0x0000600000006000
>>
>>
>>> + qemu_fprintf(f, "sp: 0x%.16lx lr: 0x%.16lx\n", sp, lr);
>>
>> HWADDR_PRIx. Or at least PRIx64, otherwise it won't compile on 32bit or
>> x86 or somewhere else.
>>
>>
>>> + sp &= ~(0xCUL << 60);
>>
>> and remove this line. And now you can dump
>>
>>
>>
>>> +
>>> + /* Does the stackframe contain regs? */
>>> + cpu_physical_memory_read(sp + 96, &marker, sizeof(marker));
>>
>> I suspect the marker needs byteswap as well.
>
> Yeah, best to use the cpu_ldl() etc wrappers if possible, which
> include byteswaps. Urgh... except this depends on the cpu mode which
> complicates things.
>
>>
>> What is that 96?
>>
>>> + if (marker == 0x7265677368657265) { /* regshere */
>>> + struct ppc_pt_regs regs;
>>> + int i;
>>> +
>>> + cpu_physical_memory_read(sp + 112, ®s, sizeof(regs));
>>
>> and the regs.
>>
>> What is that 112?
>>
>> I'd copy from arch/powerpc/include/asm/ptrace.h:
>>
>> #define STACK_FRAME_OVERHEAD 112 /* size of minimum stack frame */
>> #define STACK_FRAME_REGS_MARKER ASM_CONST(0x7265677368657265)
>>
>> and whatever that 96 is.
>>
>> Sadly, scripts/update-linux-headers.sh cannot copy
>> arch/powerpc/include/asm/ptrace.h as kernel's "make headers_install"
>> does not install it (it installs the "uapi" header which does not have
>> these symbols) so you'll have to define them.
>>
>>
>>> +
>>> + qemu_fprintf(f, "\ttrap : 0x%.16lx\n", regs.trap);
>>> + qemu_fprintf(f, "\tpc : 0x%.16lx\n", regs.nip);
>>> + qemu_fprintf(f, "\tmsr : 0x%.16lx\n", regs.msr);
>>> + qemu_fprintf(f, "\tlr : 0x%.16lx\n", regs.link);
>>> + for (i = 0; i < 32; i++)
>>> + qemu_fprintf(f, "\tgpr%2d: 0x%.16lx\n", i,
>>> + regs.gpr[i]);
>>> + }
>>> +
>>> + sp = next_sp;
>>> + }
>>> +
>>> + qemu_fprintf(f, "sp: 0x%.16lx\n", sp);
>>
>>
>> and this is "sp: 0x%"HWADDR_PRIx"\n".
>>
>>
>>
>>> +#endif
>>> +}
>>> +
>>> void ppc_cpu_dump_statistics(CPUState *cs, int flags)
>>> {
>>> #if defined(DO_PPC_STATISTICS)
>>> diff --git a/target/ppc/translate_init.inc.c b/target/ppc/translate_init.inc.c
>>> index 0394a9ddad..3fd24f85cc 100644
>>> --- a/target/ppc/translate_init.inc.c
>>> +++ b/target/ppc/translate_init.inc.c
>>> @@ -10587,6 +10587,7 @@ static void ppc_cpu_class_init(ObjectClass *oc, void *data)
>>> cc->do_interrupt = ppc_cpu_do_interrupt;
>>> cc->cpu_exec_interrupt = ppc_cpu_exec_interrupt;
>>> cc->dump_state = ppc_cpu_dump_state;
>>> + cc->dump_stack = ppc_cpu_dump_stack;
>>> cc->dump_statistics = ppc_cpu_dump_statistics;
>>> cc->set_pc = ppc_cpu_set_pc;
>>> cc->gdb_read_register = ppc_cpu_gdb_read_register;
>>>
>>
>
--
Alexey
next prev parent reply other threads:[~2019-05-02 3:48 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-05-01 5:35 [Qemu-devel] [PATCH 1/2] monitor: Add dump-stack command Suraj Jitindar Singh
2019-05-01 5:35 ` Suraj Jitindar Singh
2019-05-01 5:35 ` [Qemu-devel] [PATCH 2/2] ppc: Add dump-stack implementation Suraj Jitindar Singh
2019-05-01 5:35 ` Suraj Jitindar Singh
2019-05-01 9:48 ` Alexey Kardashevskiy
2019-05-01 9:48 ` Alexey Kardashevskiy
2019-05-02 0:43 ` David Gibson
2019-05-02 0:43 ` David Gibson
2019-05-02 3:47 ` Alexey Kardashevskiy [this message]
2019-05-02 3:47 ` Alexey Kardashevskiy
2019-05-06 3:39 ` David Gibson
2019-05-02 13:59 ` [Qemu-devel] [Qemu-ppc] " Greg Kurz
2019-05-02 13:59 ` Greg Kurz
2019-05-07 11:24 ` [Qemu-devel] " Markus Armbruster
2019-05-01 10:44 ` [Qemu-devel] [PATCH 1/2] monitor: Add dump-stack command Dr. David Alan Gilbert
2019-05-01 10:44 ` Dr. David Alan Gilbert
2019-05-02 0:44 ` David Gibson
2019-05-02 0:44 ` David Gibson
2019-05-02 2:15 ` [Qemu-devel] [Qemu-ppc] " Alexey Kardashevskiy
2019-05-02 2:15 ` Alexey Kardashevskiy
2019-05-07 11:21 ` Markus Armbruster
2019-05-07 11:09 ` [Qemu-devel] " Markus Armbruster
2019-05-08 10:26 ` Dr. David Alan Gilbert
2019-05-08 13:10 ` Markus Armbruster
2019-05-08 13:15 ` Dr. David Alan Gilbert
2019-06-21 0:51 ` Suraj Jitindar Singh
2019-06-24 8:57 ` Markus Armbruster
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=c0917aef-79cc-f42c-2f4c-6d26da9657d6@ozlabs.ru \
--to=aik@ozlabs.ru \
--cc=armbru@redhat.com \
--cc=david@gibson.dropbear.id.au \
--cc=dgilbert@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=qemu-ppc@nongnu.org \
--cc=sjitindarsingh@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).