Message ID | 45fbc3cd7f6ce404d1e25bf01695fb629d86189d.1289189560.git.michael@ellerman.id.au (mailing list archive) |
---|---|
State | Superseded |
Delegated to: | Benjamin Herrenschmidt |
Headers | show |
On Mon, 2010-11-08 at 15:12 +1100, Michael Ellerman wrote: > This patch adds support for the new "jump label" feature. Darn it! I wanted to port this!!! ;-) > > Built for 64 & 32 bit, tested on 64 bit. Survives a kernel build under > perf with all tracepoints enabled, and manual inspection of the patched > code looks right. But still, lightly tested, YMMV. > > Unlike x86 and sparc we just merrily patch the code with no locks etc, > as far as I know this is safe, but I'm not really sure what the x86/sparc > code is protecting against so maybe it's not. Thanks, I'll give it a test on my electra as soon as I get a gcc 4.5.1 cross compiler working. -- Steve
Hi Michael, Just a trivial note ... On Mon, 8 Nov 2010 15:12:46 +1100 (EST) Michael Ellerman <michael@ellerman.id.au> wrote: > > diff --git a/arch/powerpc/include/asm/jump_label.h b/arch/powerpc/include/asm/jump_label.h > new file mode 100644 > index 0000000..feffb86 > --- /dev/null > +++ b/arch/powerpc/include/asm/jump_label.h > @@ -0,0 +1,46 @@ > +#ifndef _ASM_POWERPC_JUMP_LABEL_H > +#define _ASM_POWERPC_JUMP_LABEL_H > + > +/* > + * Copyright 2010 Michael Ellerman, IBM Corp. > + * > + * This program is free software; you can redistribute it and/or > + * modify it under the terms of the GNU General Public License > + * as published by the Free Software Foundation; either version > + * 2 of the License, or (at your option) any later version. > + */ > + > +#ifdef __KERNEL__ You don't need this protection in a file that is not exported to user space.
On Mon, 2010-11-08 at 09:18 -0500, Steven Rostedt wrote: > On Mon, 2010-11-08 at 15:12 +1100, Michael Ellerman wrote: > > This patch adds support for the new "jump label" feature. > > Darn it! I wanted to port this!!! > > ;-) Sorry! I've been locked in a dark room for ages and this was a nice chance to work on some upstream code for a change :) > > Built for 64 & 32 bit, tested on 64 bit. Survives a kernel build under > > perf with all tracepoints enabled, and manual inspection of the patched > > code looks right. But still, lightly tested, YMMV. > > > > Unlike x86 and sparc we just merrily patch the code with no locks etc, > > as far as I know this is safe, but I'm not really sure what the x86/sparc > > code is protecting against so maybe it's not. > > Thanks, I'll give it a test on my electra as soon as I get a gcc 4.5.1 > cross compiler working. Thanks. cheers
On Tue, 2010-11-09 at 13:35 +1100, Stephen Rothwell wrote: > Hi Michael, > > Just a trivial note ... > > On Mon, 8 Nov 2010 15:12:46 +1100 (EST) Michael Ellerman <michael@ellerman.id.au> wrote: > > > > diff --git a/arch/powerpc/include/asm/jump_label.h b/arch/powerpc/include/asm/jump_label.h > > new file mode 100644 > > index 0000000..feffb86 > > --- /dev/null > > +++ b/arch/powerpc/include/asm/jump_label.h > > @@ -0,0 +1,46 @@ > > +#ifndef _ASM_POWERPC_JUMP_LABEL_H > > +#define _ASM_POWERPC_JUMP_LABEL_H > > + > > +/* > > + * Copyright 2010 Michael Ellerman, IBM Corp. > > + * > > + * This program is free software; you can redistribute it and/or > > + * modify it under the terms of the GNU General Public License > > + * as published by the Free Software Foundation; either version > > + * 2 of the License, or (at your option) any later version. > > + */ > > + > > +#ifdef __KERNEL__ > > You don't need this protection in a file that is not exported to user > space. Ah true. I'll respin after I've heard if anyone has tested it :) cheers
On Tue, 2010-11-09 at 14:33 +1100, Michael Ellerman wrote: > > You don't need this protection in a file that is not exported to user > > space. > > Ah true. I'll respin after I've heard if anyone has tested it :) Don't wait for me. I'm stuck debugging some other crap. Getting a 4.5 PPC compiler working is very low on my priority list, unfortunately. :-( -- Steve
On Mon, 2010-11-08 at 22:41 -0500, Steven Rostedt wrote: > On Tue, 2010-11-09 at 14:33 +1100, Michael Ellerman wrote: > > > > You don't need this protection in a file that is not exported to user > > > space. > > > > Ah true. I'll respin after I've heard if anyone has tested it :) > > Don't wait for me. I'm stuck debugging some other crap. Getting a 4.5 > PPC compiler working is very low on my priority list, unfortunately. :-( No worries, testing is for sissies anyway. There are 4.5.0 crosses on kernel.org, I think they're the same as the ones I'm using, which seem to work on powerpc at least. http://kernel.org/pub/tools/crosstool/files/bin/x86_64/4.5.0/ cheers
diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index b644719..301c5b7 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -136,6 +136,7 @@ config PPC select HAVE_PERF_EVENTS select HAVE_REGS_AND_STACK_ACCESS_API select HAVE_HW_BREAKPOINT if PERF_EVENTS && PPC_BOOK3S_64 + select HAVE_ARCH_JUMP_LABEL config EARLY_PRINTK bool diff --git a/arch/powerpc/include/asm/jump_label.h b/arch/powerpc/include/asm/jump_label.h new file mode 100644 index 0000000..feffb86 --- /dev/null +++ b/arch/powerpc/include/asm/jump_label.h @@ -0,0 +1,46 @@ +#ifndef _ASM_POWERPC_JUMP_LABEL_H +#define _ASM_POWERPC_JUMP_LABEL_H + +/* + * Copyright 2010 Michael Ellerman, IBM Corp. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + */ + +#ifdef __KERNEL__ + +#include <linux/types.h> +#include <asm/feature-fixups.h> + +#define JUMP_ENTRY_TYPE stringify_in_c(FTR_ENTRY_LONG) + +#define JUMP_LABEL_NOP_SIZE 4 + +#define JUMP_LABEL(key, label) \ +do { \ + asm goto("1:\n\t" \ + "nop\n\t" \ + ".pushsection __jump_table, \"a\"\n\t" \ + JUMP_ENTRY_TYPE "1b, %l[" #label "], %c0\n\t" \ + ".popsection \n\t" \ + : : "i" (key) : : label); \ +} while (0) + +#ifdef CONFIG_PPC64 +typedef u64 jump_label_t; +#else +typedef u32 jump_label_t; +#endif + +struct jump_entry { + jump_label_t code; + jump_label_t target; + jump_label_t key; +}; + +#endif /* __KERNEL__ */ + +#endif /* _ASM_POWERPC_JUMP_LABEL_H */ diff --git a/arch/powerpc/kernel/Makefile b/arch/powerpc/kernel/Makefile index 36c30f3..315e933 100644 --- a/arch/powerpc/kernel/Makefile +++ b/arch/powerpc/kernel/Makefile @@ -71,6 +71,7 @@ obj-$(CONFIG_MODULES) += module.o module_$(CONFIG_WORD_SIZE).o obj-$(CONFIG_44x) += cpu_setup_44x.o obj-$(CONFIG_PPC_FSL_BOOK3E) += cpu_setup_fsl_booke.o dbell.o obj-$(CONFIG_PPC_BOOK3E_64) += dbell.o +obj-$(CONFIG_JUMP_LABEL) += jump_label.o extra-y := head_$(CONFIG_WORD_SIZE).o extra-$(CONFIG_PPC_BOOK3E_32) := head_new_booke.o diff --git a/arch/powerpc/kernel/jump_label.c b/arch/powerpc/kernel/jump_label.c new file mode 100644 index 0000000..368d158 --- /dev/null +++ b/arch/powerpc/kernel/jump_label.c @@ -0,0 +1,23 @@ +/* + * Copyright 2010 Michael Ellerman, IBM Corp. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + */ + +#include <linux/kernel.h> +#include <linux/jump_label.h> +#include <asm/code-patching.h> + +void arch_jump_label_transform(struct jump_entry *entry, + enum jump_label_type type) +{ + u32 *addr = (u32 *)(unsigned long)entry->code; + + if (type == JUMP_LABEL_ENABLE) + patch_branch(addr, entry->target, 0); + else + patch_instruction(addr, PPC_INST_NOP); +}
This patch adds support for the new "jump label" feature. Built for 64 & 32 bit, tested on 64 bit. Survives a kernel build under perf with all tracepoints enabled, and manual inspection of the patched code looks right. But still, lightly tested, YMMV. Unlike x86 and sparc we just merrily patch the code with no locks etc, as far as I know this is safe, but I'm not really sure what the x86/sparc code is protecting against so maybe it's not. I also don't see any reason for us to implement the poke_early() routine, even though sparc does. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> --- arch/powerpc/Kconfig | 1 + arch/powerpc/include/asm/jump_label.h | 46 +++++++++++++++++++++++++++++++++ arch/powerpc/kernel/Makefile | 1 + arch/powerpc/kernel/jump_label.c | 23 ++++++++++++++++ 4 files changed, 71 insertions(+), 0 deletions(-)