Re: [PATCH V3 0/3] perf/powerpc:Add ability to sample intr machine state in powerpc

From: Anju T
Date: Fri Nov 06 2015 - 01:59:35 EST


Hi Denis,

On Wednesday 04 November 2015 02:26 PM, Denis Kirjanov wrote:
On 11/3/15, Anju T <anju@xxxxxxxxxxxxxxxxxx> wrote:
This short patch series adds the ability to sample the interrupted
machine state for each hardware sample.

To test this patchset,
Eg:

$perf record -I ls // record machine state at interrupt
$perf script -D //read the perf.data file
Uncovered the following warning with the series applied. Looks like
that it's not
directly related to your patches but anyway...

May I know the config you used while testing?
Did this warning appeared on boot time?

Thanks

Anju


[ 507.655197] DEBUG_LOCKS_WARN_ON(!current->hardirqs_enabled)
[ 507.655220] ------------[ cut here ]------------
[ 507.655226] WARNING: at kernel/locking/lockdep.c:3523
[ 507.655230] Modules linked in: ipv6 binfmt_misc ehea
[ 507.655242] CPU: 12 PID: 3746 Comm: ls Tainted: G S
4.3.0-rc3-00103-g3b0e21e-dirty #11
[ 507.655249] task: c0000005b607b290 ti: c0000005b62e8000 task.ti:
c0000005b62e8000
[ 507.655255] NIP: c00000000010c944 LR: c00000000010c940 CTR: c000000000659380
[ 507.655261] REGS: c0000005b62eb5c0 TRAP: 0700 Tainted: G S
(4.3.0-rc3-00103-g3b0e21e-dirty)
[ 507.655266] MSR: 8000000000029032 <SF,EE,ME,IR,DR,RI> CR: 22088422
XER: 0000000e
[ 507.655284] CFAR: c0000000008977c4 SOFTE: 0
GPR00: c00000000010c940 c0000005b62eb840 c00000000102e600 000000000000002f
GPR04: 0000000000000001 c0000000001208d8 0000000000000000 0000000000000001
GPR08: c000000000eee600 c0000005b607b290 0000000000000000 0000000000003fef
GPR12: 0000000042088428 c00000000e956600 00000000001fffff 00003fffd546c4b0
GPR16: 00000000001fffff 0000000000000013 c000000000b198a8 0000000000000000
GPR20: c0000005acbb2f80 c0000005b5c5dd00 00003fffd546c500 0000000000000001
GPR24: c00000000024af54 0000000000000000 0000000000000001 0000000000000001
GPR28: 0000000000000000 0000000000000000 c0000005acbb2ea0 c000000001e0cf78
[ 507.655376] NIP [c00000000010c944] .check_flags.part.36+0xd4/0x240
[ 507.655382] LR [c00000000010c940] .check_flags.part.36+0xd0/0x240
[ 507.655387] Call Trace:
[ 507.655391] [c0000005b62eb840] [c00000000010c940]
.check_flags.part.36+0xd0/0x240 (unreliable)
[ 507.655400] [c0000005b62eb8c0] [c0000000001112b8] .lock_acquire+0x208/0x2a0
[ 507.655407] [c0000005b62eb990] [c00000000024af80] .__might_fault+0xb0/0xf0
[ 507.655415] [c0000005b62eba10] [c0000000004d5d38] .strnlen_user+0x1d8/0x200
[ 507.655422] [c0000005b62ebad0] [c00000000032fa0c]
.load_elf_binary+0x103c/0x1650
[ 507.655430] [c0000005b62ebc10] [c0000000002bac54]
.search_binary_handler+0xc4/0x260
[ 507.655437] [c0000005b62ebcb0] [c0000000002bcd54]
.do_execveat_common.isra.22+0x7d4/0xb40
[ 507.655444] [c0000005b62ebda0] [c0000000002bd4a8] .SyS_execve+0x38/0x50
[ 507.655451] [c0000005b62ebe30] [c00000000000916c] system_call+0x38/0xd0
[ 507.655456] Instruction dump:
[ 507.655461] 419e0034 3d4200e5 392a3280 81290000 2f890000 40fe0020
3c62ffad 3c82ffad
[ 507.655475] 3863c038 38841f88 4878adfd 60000000 <0fe00000> 3c62ffad
38632010 4878ade9
[ 507.655490] ---[ end trace 47284e8c92efaa7e ]---
[ 507.655494] possible reason: unannotated irqs-on.
[ 507.655498] irq event stamp: 2324
[ 507.655501] hardirqs last enabled at (2323): [<c00000000088ed24>]
._raw_spin_unlock_irqrestore+0x54/0xd0
[ 507.655510] hardirqs last disabled at (2324): [<c000000000009924>]
restore_irq_off+0x24/0x28
[ 507.655518] softirqs last enabled at (2184): [<c00000000009cb90>]
.__do_softirq+0x500/0x670
[ 507.655526] softirqs last disabled at (2169): [<c00000000009d198>]
.irq_exit+0xd8/0x120

Sample output obtained for this patchset/ output looks like as follows:

331557004666 0x1988 [0x188]: PERF_RECORD_SAMPLE(IP, 0x1): 4807/4807:
0xc0000000001ddf60 period: 1 addr: 0
... intr regs: mask 0x7ffffffffff ABI 64-bit
.... gpr0 0xc0000000001e6a74
.... gpr1 0xc0000000ff33b9a0
.... gpr2 0xc000000001523000
.... gpr3 0xc000000ffa9deb60
.... gpr4 0xc0000000ff971e00
.... gpr5 0x4d32564532
.... gpr6 0x1e000000
.... gpr7 0x0
.... gpr8 0x0
.... gpr9 0x0
.... gpr10 0x1
.... gpr11 0x0
.... gpr12 0x24022822
.... gpr13 0xc00000000feeaf80
.... gpr14 0x0
.... gpr15 0xc0000000fbc21000
.... gpr16 0x0
.... gpr17 0xc000000ffa9c5000
.... gpr18 0xc0000000ff33b8a0
.... gpr19 0xc000000001523000
.... gpr20 0xc0000000000a097c
.... gpr21 0xc000000fcac65600
.... gpr22 0xc0000000001e55a8
.... gpr23 0xc000000001523000
.... gpr24 0xc0000000ff33b850
.... gpr25 0xc000000fcac65600
.... gpr26 0xc000001e4b378210
.... gpr27 0xfffffffffffffead
.... gpr28 0x100000000
.... gpr29 0xc000000fcac65600
.... gpr30 0x1
.... gpr31 0x0
.... nip 0xc0000000001ddf68
.... msr 0x9000000000009032
.... orig_r3 0xc0000000001e5fcc
.... ctr 0xc00000000009e1b0
.... link 0xc0000000001e6a74
.... xer 0x0
.... ccr 0x84022882
.... softe 0x0
.... trap 0xf01
.... dar 0x0
.... dsisr 0xf00040060000004
... thread: :4807:4807
...... dso:
/root/.debug/.build-id/1c/011201a1082e91b8449e6dd528f224d7a16535
:4807 4807 331.557004: 1 cycles: c0000000001ddf60
.perf_ctx_unlock (/boot/vmlinux)

0x1b10 [0x188]: event: 9


Changes from V2:

- tools/perf/config/Makefile is moved to the patch tools/perf.
- The patchset is reordered.
- perf_regs_load() function is used for the dwarf unwind test.Since it is
not required here,
it is removed from tools/perf/arch/powerpc/include/perf_regs.h
- PERF_REGS_POWERPC_RESULT is removed.





Anju T (3):
perf/powerpc:add ability to sample intr machine state in power
perf/powerpc :add support for sampling intr machine state
tools/perf:Map the ID values with register names

arch/powerpc/Kconfig | 1 +
arch/powerpc/include/uapi/asm/perf_regs.h | 54 +++++++++++++
arch/powerpc/perf/Makefile | 2 +
arch/powerpc/perf/perf_regs.c | 92 ++++++++++++++++++++++
tools/perf/arch/powerpc/include/perf_regs.h | 114
++++++++++++++++++++++++++++
tools/perf/config/Makefile | 5 ++
6 files changed, 268 insertions(+)
create mode 100644 arch/powerpc/include/uapi/asm/perf_regs.h
create mode 100644 arch/powerpc/perf/perf_regs.c
create mode 100644 tools/perf/arch/powerpc/include/perf_regs.h

--
2.1.0

_______________________________________________


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/