Commit 57256089 authored by Vineet Gupta's avatar Vineet Gupta Committed by Luis Henriques

ARC: dw2 unwind: Remove falllback linear search thru FDE entries

commit 2e22502c 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: default avatarVineet Gupta <vgupta@synopsys.com>
Signed-off-by: default avatarLuis Henriques <luis.henriques@canonical.com>
parent 2682034a
...@@ -986,42 +986,13 @@ int arc_unwind(struct unwind_frame_info *frame) ...@@ -986,42 +986,13 @@ int arc_unwind(struct unwind_frame_info *frame)
(const u8 *)(fde + (const u8 *)(fde +
1) + 1) +
*fde, ptrType); *fde, ptrType);
if (pc >= endLoc) if (pc >= endLoc) {
fde = NULL; 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; cie = NULL;
break;
} }
if (cie == NULL } else {
|| cie == &not_fde fde = NULL;
|| (ptrType = fde_pointer_type(cie)) < 0) cie = NULL;
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;
} }
} }
} }
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment