[<prev] [next>] [day] [month] [year] [list]
Message-Id: <1256650350.5499.117.camel@falcon>
Date: Tue, 27 Oct 2009 21:32:30 +0800
From: Wu Zhangjin <wuzhangjin@...il.com>
To: Steven Rostedt <rostedt@...dmis.org>, rdsandiford@...glemail.com
Cc: David Daney <ddaney@...iumnetworks.com>,
Ralf Baechle <ralf@...ux-mips.org>, linux-mips@...ux-mips.org,
linux-kernel@...r.kernel.org
Subject: mipsel-linux-gnu-gcc: -pg and -fomit-frame-pointer are incompatible
Hi, all
FUNCTION_TRACER have selected FRAME_POINTER by default to avoid the
following "weird" error when using -pg and -fomit-frame-pointer
together:
"-pg and -fomit-frame-pointer are incompatible"
kernel/trace/Kconfig:
config FUNCTION_TRACER
bool "Kernel Function Tracer"
depends on HAVE_FUNCTION_TRACER
select FRAME_POINTER
and here is what FRAME_POINTER does in (linux)/Makefile:
ifdef CONFIG_FRAME_POINTER
KBUILD_CFLAGS += -fno-omit-frame-pointer -fno-optimize-sibling-calls
else
KBUILD_CFLAGS += -fomit-frame-pointer
endif
but in reality, from the manual of gcc:
"Don’t keep the frame pointer in a register for functions that don’t
need one. This avoids the instructions to save, set up and restore
frame pointers; it also makes an extra register available in many
functions. It also makes debugging impossible on some machines.
On some machines, such as the VAX, this flag has no effect, because the
standard calling sequence automatically handles the frame pointer and
nothing is saved by pretending it doesn’t exist. The
machine-description macro "FRAME_POINTER_REQUIRED" controls whether a
target machine supports this flag.
Enabled at levels -O, -O2, -O3, -Os."
-fomit-frame-pointer will be enabled by default for -O2, and If I
disable -fno-omit-frame-pointer, it will really not keep the frame
pointer in a register:
ffffffff80200400 <do_one_initcall>:
ffffffff80200400: 67bdffd0 daddiu sp,sp,-48
ffffffff80200404: ffbf0028 sd ra,40(sp)
ffffffff80200408: ffb40020 sd s4,32(sp)
ffffffff8020040c: ffb30018 sd s3,24(sp)
ffffffff80200410: ffb20010 sd s2,16(sp)
ffffffff80200414: ffb10008 sd s1,8(sp)
ffffffff80200418: ffb00000 sd s0,0(sp)
ffffffff8020041c: 3c038021 lui v1,0x8021
ffffffff80200420: 64630fb0 daddiu v1,v1,4016 <> with -pg
ffffffff80200424: 03e0082d move at,ra
ffffffff80200428: 0060f809 jalr v1
ffffffff8020042c: 00020021 nop
[...]
ffffffff80205b18 <au1k_wait>:
ffffffff80205b18: 3c038021 lui v1,0x8021
ffffffff80205b1c: 64630fb0 daddiu v1,v1,4016
ffffffff80205b20: 03e0082d move at,ra
ffffffff80205b24: 0060f809 jalr v1
ffffffff80205b28: 00020021 nop
And without -fno-omit-frame-pointer option, ftrace for MIPS also works
normally and can save some overhead for us!
But perhaps some archs need the frame pointer, so, remove the
-fno-omit-frame-pointer from (linux)/Makefile, and add it into the arch
specific Makefile?
Besides, should we clear the "weird" error in gcc when using -pg and
-fomit-frame-pinter together?
Thanks & Regards,
Wu Zhangjin
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists