[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20220622105017.04630f12@rorschach.local.home>
Date: Wed, 22 Jun 2022 10:50:18 -0400
From: Steven Rostedt <rostedt@...dmis.org>
To: Peter Zijlstra <peterz@...radead.org>
Cc: x86@...nel.org, linux-kernel@...r.kernel.org,
Josh Poimboeuf <jpoimboe@...hat.com>,
christophe.leroy@...roup.eu, naveen.n.rao@...ux.vnet.ibm.com,
mbenes@...e.cz
Subject: Re: [RFC][PATCH] ftrace,objtool: PC32 based __mcount_loc
On Fri, 17 Jun 2022 13:24:53 +0200
Peter Zijlstra <peterz@...radead.org> wrote:
> Hi,
>
> I recently noticed that __mcount_loc is 64bit wide, containing absolute
> addresses. Since __mcount_loc is a permanent section (not one we drop
> after boot), this bloats the kernel memory usage for no real purpose.
Wait, it's not dropped? Nothing uses it after it is read. It should be
dropped when init data is dropped.
>From include/asm-generic/vmlinux.lds.h
/* init and exit section handling */
#define INIT_DATA \
KEEP(*(SORT(___kentry+*))) \
*(.init.data init.data.*) \
MEM_DISCARD(init.data*) \
KERNEL_CTORS() \
MCOUNT_REC() \ <<----
*(.init.rodata .init.rodata.*) \
FTRACE_EVENTS() \
TRACE_SYSCALLS() \
KPROBE_BLACKLIST() \
ERROR_INJECT_WHITELIST() \
MEM_DISCARD(init.rodata) \
So it should be dropped after boot.
-- Steve
>
> The below patch adds __mcount_loc_32 and objtool support to generate it.
> This saves, on an x86_64-defconfig + FTRACE, 23975*4 ~= 94K of permanent
> storage.
>
> XXX hobbled sorttable for now
> XXX compile tested only
>
Powered by blists - more mailing lists