Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1347726
| From | Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
|---|---|
| Newsgroups | linux.kernel |
| Subject | [PATCH 3.10 32/80] ARC: dw2 unwind: Remove falllback linear search thru FDE entries |
| Date | 2016-03-02 03:40 +0100 |
| Message-ID | <r84RJ-1dc-53@gated-at.bofh.it> (permalink) |
| References | <r82dd-7Iy-17@gated-at.bofh.it> |
| Organization | linux.* mail to news gateway |
3.10-stable review patch. If anyone has any objections, please let me know.
------------------
From: Vineet Gupta <vgupta@synopsys.com>
commit 2e22502c080f27afeab5e6f11e618fb7bc7aea53 upstream.
Fixes STAR 9000953410: "perf callgraph profiling causing RCU stalls"
| perf record -g -c 15000 -e cycles /sbin/hackbench
|
| INFO: rcu_preempt self-detected stall on CPU
| 1: (1 GPs behind) idle=609/140000000000002/0 softirq=2914/2915 fqs=603
| Task dump for CPU 1:
in-kernel dwarf unwinder has a fast binary lookup and a fallback linear
search (which iterates thru each of ~11K entries) thus takes 2 orders of
magnitude longer (~3 million cycles vs. 2000). Routines written in hand
assembler lack dwarf info (as we don't support assembler CFI pseudo-ops
yet) fail the unwinder binary lookup, hit linear search, failing
nevertheless in the end.
However the linear search is pointless as binary lookup tables are created
from it in first place. It is impossible to have binary lookup fail while
succeed the linear search. It is pure waste of cycles thus removed by
this patch.
This manifested as RCU stalls / NMI watchdog splat when running
hackbench under perf with callgraph profiling. The triggering condition
was perf counter overflowing in routine lacking dwarf info (like memset)
leading to patheic 3 million cycle unwinder slow path and by the time it
returned new interrupts were already pending (Timer, IPI) and taken
rightaway. The original memset didn't make forward progress, system kept
accruing more interrupts and more unwinder delayes in a vicious feedback
loop, ultimately triggering the NMI diagnostic.
Signed-off-by: Vineet Gupta <vgupta@synopsys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
arch/arc/kernel/unwind.c | 37 ++++---------------------------------
1 file changed, 4 insertions(+), 33 deletions(-)
--- a/arch/arc/kernel/unwind.c
+++ b/arch/arc/kernel/unwind.c
@@ -984,42 +984,13 @@ int arc_unwind(struct unwind_frame_info
(const u8 *)(fde +
1) +
*fde, ptrType);
- if (pc >= endLoc)
+ if (pc >= endLoc) {
fde = NULL;
- } else
- fde = NULL;
- }
- if (fde == NULL) {
- for (fde = table->address, tableSize = table->size;
- cie = NULL, tableSize > sizeof(*fde)
- && tableSize - sizeof(*fde) >= *fde;
- tableSize -= sizeof(*fde) + *fde,
- fde += 1 + *fde / sizeof(*fde)) {
- cie = cie_for_fde(fde, table);
- if (cie == &bad_cie) {
cie = NULL;
- break;
}
- if (cie == NULL
- || cie == ¬_fde
- || (ptrType = fde_pointer_type(cie)) < 0)
- continue;
- ptr = (const u8 *)(fde + 2);
- startLoc = read_pointer(&ptr,
- (const u8 *)(fde + 1) +
- *fde, ptrType);
- if (!startLoc)
- continue;
- if (!(ptrType & DW_EH_PE_indirect))
- ptrType &=
- DW_EH_PE_FORM | DW_EH_PE_signed;
- endLoc =
- startLoc + read_pointer(&ptr,
- (const u8 *)(fde +
- 1) +
- *fde, ptrType);
- if (pc >= startLoc && pc < endLoc)
- break;
+ } else {
+ fde = NULL;
+ cie = NULL;
}
}
}
Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread
[PATCH 3.10 00/80] 3.10.99-stable review Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:40 +0100
[PATCH 3.10 08/80] dm thin metadata: fix bug when taking a metadata snapshot Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:40 +0100
[PATCH 3.10 02/80] drm/ast: Initialized data needed to map fbdev memory Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:40 +0100
[PATCH 3.10 23/80] dts: vt8500: Add SDHC node to DTS file for WM8650 Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:40 +0100
[PATCH 3.10 07/80] efi: Disable interrupts around EFI calls, not in the epilog/prolog calls Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:40 +0100
[PATCH 3.10 31/80] mac: validate mac_partition is within sector Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:40 +0100
[PATCH 3.10 14/80] proc: Fix ptrace-based permission checks for accessing task maps Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:40 +0100
[PATCH 3.10 29/80] megaraid_sas : SMAP restriction--do not access user memory from IOCTL code Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:40 +0100
[PATCH 3.10 32/80] ARC: dw2 unwind: Remove falllback linear search thru FDE entries Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:40 +0100
[PATCH 3.10 11/80] USB: cp210x: add IDs for GE B650V3 and B850V3 boards Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:40 +0100
[PATCH 3.10 41/80] dm btree: fix bufio buffer leaks in dm_btree_del() error path Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:40 +0100
[PATCH 3.10 27/80] wm831x_power: Use IRQF_ONESHOT to request threaded IRQs Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:40 +0100
[PATCH 3.10 10/80] can: ems_usb: Fix possible tx overflow Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:40 +0100
[PATCH 3.10 01/80] tracepoints: Do not trace when cpu is offline Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:40 +0100
[PATCH 3.10 06/80] drm/radeon: fix hotplug race at startup Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:40 +0100
[PATCH 3.10 26/80] devres: fix a for loop bounds check Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:40 +0100
[PATCH 3.10 30/80] mmc: remove bondage between REQ_META and reliable write Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:40 +0100
[PATCH 3.10 15/80] iw_cxgb3: Fix incorrectly returning error on success Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:40 +0100
[PATCH 3.10 35/80] virtio: fix memory leak of virtio ida cache layers Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:40 +0100
[PATCH 3.10 04/80] bcache: unregister reboot notifier if bcache fails to unregister device Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 03:50 +0100
Re: [PATCH 3.10 00/80] 3.10.99-stable review Guenter Roeck <linux@roeck-us.net> - 2016-03-02 15:40 +0100
Re: [PATCH 3.10 00/80] 3.10.99-stable review Willy Tarreau <w@1wt.eu> - 2016-03-02 16:50 +0100
Re: [PATCH 3.10 00/80] 3.10.99-stable review Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-03-02 18:30 +0100
csiph-web