Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1317059 > unrolled thread

[PATCH v6 0/9] ftrace with regs + live patching for ppc64 LE (ABI v2)

Back to article view | Back to linux.kernel

Contents

  [PATCH v6 0/9] ftrace with regs + live patching for ppc64 LE (ABI v2) Torsten Duwe <duwe@lst.de> - 2016-01-25 18:10 +0100
    [PATCH v6 5/9] ppc64 ftrace_with_regs: spare early boot and low level Torsten Duwe <duwe@lst.de> - 2016-01-25 18:10 +0100
    [PATCH v6 8/9] Implement kernel live patching for ppc64le (ABIv2) Torsten Duwe <duwe@lst.de> - 2016-01-25 18:10 +0100
      Re: [PATCH v6 8/9] Implement kernel live patching for ppc64le  (ABIv2) Miroslav Benes <mbenes@suse.cz> - 2016-01-26 12:00 +0100
        Re: [PATCH v6 8/9] Implement kernel live patching for ppc64le (ABIv2) Petr Mladek <pmladek@suse.com> - 2016-01-26 13:50 +0100
          Re: [PATCH v6 8/9] Implement kernel live patching for ppc64le  (ABIv2) Torsten Duwe <duwe@lst.de> - 2016-01-26 15:00 +0100
        Re: [PATCH v6 8/9] Implement kernel live patching for ppc64le  (ABIv2) Torsten Duwe <duwe@lst.de> - 2016-01-26 15:10 +0100
          Re: [PATCH v6 8/9] Implement kernel live patching for ppc64le  (ABIv2) Miroslav Benes <mbenes@suse.cz> - 2016-01-26 15:20 +0100
            Re: Implement kernel live patching for ppc64le (ABIv2) Jessica Yu <jeyu@redhat.com> - 2016-01-27 03:00 +0100
    [PATCH v6 2/9] ppc64le FTRACE_WITH_REGS implementation Torsten Duwe <duwe@lst.de> - 2016-01-25 18:10 +0100
    [PATCH v6 7/9] ppc64 ftrace: disable profiling for some files Torsten Duwe <duwe@lst.de> - 2016-01-25 18:20 +0100
    Re: [PATCH v6 0/9] ftrace with regs + live patching for ppc64 LE  (ABI v2) Balbir Singh <bsingharora@gmail.com> - 2016-01-27 12:00 +0100
      Re: [PATCH v6 0/9] ftrace with regs + live patching for ppc64 LE  (ABI v2) Torsten Duwe <duwe@lst.de> - 2016-01-27 13:20 +0100

#1317059 — [PATCH v6 0/9] ftrace with regs + live patching for ppc64 LE (ABI v2)

Changes since v5:
  * extra "std r0,LRSAVE(r1)" for gcc-6
    This makes the code compiler-agnostic.
  * Follow Petr Mladek's suggestion to avoid
    redefinition of HAVE_LIVEPATCH

Changes since v4:
  * change comment style in entry_64.S to C89
    (nobody is using assembler syntax comments there).
  * the bool function restore_r2 shouldn't return 2,
    that's a little confusing.
  * Test whether the compiler supports -mprofile-kernel
    and only then define CC_USING_MPROFILE_KERNEL
  * also make the return value of klp_check_compiler_support
    depend on that.

Major changes since v3:
  * the graph tracer works now.
    It turned out the stack frame it tried to manipulate does not
    exist at that point.
  * changes only needed in order to support -mprofile-kernel are now
    in a separate patch, prepended.
  * Kconfig cleanup so this is only selectable on ppc64le.


Torsten Duwe (9):
  ppc64 (le): prepare for -mprofile-kernel
  ppc64le FTRACE_WITH_REGS implementation
  ppc use ftrace_modify_all_code default
  ppc64 ftrace_with_regs configuration variables
  ppc64 ftrace_with_regs: spare early boot and low level
  ppc64 ftrace: disable profiling for some functions
  ppc64 ftrace: disable profiling for some files
  Implement kernel live patching for ppc64le (ABIv2)
  Enable LIVEPATCH to be configured on ppc64le and add livepatch.o if it
    is selected.

 arch/powerpc/Kconfig                 |   7 ++
 arch/powerpc/Makefile                |  10 +++
 arch/powerpc/include/asm/ftrace.h    |   5 ++
 arch/powerpc/include/asm/livepatch.h |  45 ++++++++++
 arch/powerpc/kernel/Makefile         |  13 +--
 arch/powerpc/kernel/entry_64.S       | 166 ++++++++++++++++++++++++++++++++++-
 arch/powerpc/kernel/ftrace.c         |  88 ++++++++++++++-----
 arch/powerpc/kernel/livepatch.c      |  38 ++++++++
 arch/powerpc/kernel/module_64.c      |  39 +++++++-
 arch/powerpc/kernel/process.c        |   2 +-
 arch/powerpc/lib/Makefile            |   4 +-
 arch/powerpc/mm/fault.c              |   2 +-
 arch/powerpc/mm/hash_utils_64.c      |  18 ++--
 arch/powerpc/mm/hugetlbpage-hash64.c |   2 +-
 arch/powerpc/mm/hugetlbpage.c        |   4 +-
 arch/powerpc/mm/mem.c                |   2 +-
 arch/powerpc/mm/pgtable_64.c         |   2 +-
 arch/powerpc/mm/slb.c                |   6 +-
 arch/powerpc/mm/slice.c              |   8 +-
 kernel/trace/Kconfig                 |   5 ++
 20 files changed, 412 insertions(+), 54 deletions(-)
 create mode 100644 arch/powerpc/include/asm/livepatch.h
 create mode 100644 arch/powerpc/kernel/livepatch.c

-- 
1.8.5.6

[toc] | [next] | [standalone]

#1317060 — [PATCH v6 5/9] ppc64 ftrace_with_regs: spare early boot and low level

Using -mprofile-kernel on early boot code not only confuses the
checker but is also useless, as the infrastructure is not yet in
place. Proceed like with -pg (remove it from CFLAGS), equally with
time.o and ftrace itself.

  * arch/powerpc/kernel/Makefile:
    - remove -mprofile-kernel from low level and boot code objects'
      CFLAGS for FUNCTION_TRACER configurations.

Signed-off-by: Torsten Duwe <duwe@suse.de>
---
 arch/powerpc/kernel/Makefile | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/arch/powerpc/kernel/Makefile b/arch/powerpc/kernel/Makefile
index ba33693..0f417d5 100644
--- a/arch/powerpc/kernel/Makefile
+++ b/arch/powerpc/kernel/Makefile
@@ -16,14 +16,14 @@ endif
 
 ifdef CONFIG_FUNCTION_TRACER
 # Do not trace early boot code
-CFLAGS_REMOVE_cputable.o = -pg -mno-sched-epilog
-CFLAGS_REMOVE_prom_init.o = -pg -mno-sched-epilog
-CFLAGS_REMOVE_btext.o = -pg -mno-sched-epilog
-CFLAGS_REMOVE_prom.o = -pg -mno-sched-epilog
+CFLAGS_REMOVE_cputable.o = -pg -mno-sched-epilog -mprofile-kernel
+CFLAGS_REMOVE_prom_init.o = -pg -mno-sched-epilog -mprofile-kernel
+CFLAGS_REMOVE_btext.o = -pg -mno-sched-epilog -mprofile-kernel
+CFLAGS_REMOVE_prom.o = -pg -mno-sched-epilog -mprofile-kernel
 # do not trace tracer code
-CFLAGS_REMOVE_ftrace.o = -pg -mno-sched-epilog
+CFLAGS_REMOVE_ftrace.o = -pg -mno-sched-epilog -mprofile-kernel
 # timers used by tracing
-CFLAGS_REMOVE_time.o = -pg -mno-sched-epilog
+CFLAGS_REMOVE_time.o = -pg -mno-sched-epilog -mprofile-kernel
 endif
 
 obj-y				:= cputable.o ptrace.o syscalls.o \
-- 
1.8.5.6

[toc] | [prev] | [next] | [standalone]

#1317061 — [PATCH v6 8/9] Implement kernel live patching for ppc64le (ABIv2)

  * create the appropriate files+functions
    arch/powerpc/include/asm/livepatch.h
        klp_check_compiler_support,
        klp_arch_set_pc
    arch/powerpc/kernel/livepatch.c with a stub for
        klp_write_module_reloc
    This is architecture-independent work in progress.
  * introduce a fixup in arch/powerpc/kernel/entry_64.S
    for local calls that are becoming global due to live patching.
    And of course do the main KLP thing: return to a maybe different
    address, possibly altered by the live patching ftrace op.

Signed-off-by: Torsten Duwe <duwe@suse.de>
---
 arch/powerpc/include/asm/livepatch.h | 45 +++++++++++++++++++++++++++++++
 arch/powerpc/kernel/entry_64.S       | 51 +++++++++++++++++++++++++++++++++---
 arch/powerpc/kernel/livepatch.c      | 38 +++++++++++++++++++++++++++
 3 files changed, 130 insertions(+), 4 deletions(-)
 create mode 100644 arch/powerpc/include/asm/livepatch.h
 create mode 100644 arch/powerpc/kernel/livepatch.c

diff --git a/arch/powerpc/include/asm/livepatch.h b/arch/powerpc/include/asm/livepatch.h
new file mode 100644
index 0000000..44e8a2d
--- /dev/null
+++ b/arch/powerpc/include/asm/livepatch.h
@@ -0,0 +1,45 @@
+/*
+ * livepatch.h - powerpc-specific Kernel Live Patching Core
+ *
+ * Copyright (C) 2015 SUSE
+ *
+ * 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.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, see <http://www.gnu.org/licenses/>.
+ */
+#ifndef _ASM_POWERPC64_LIVEPATCH_H
+#define _ASM_POWERPC64_LIVEPATCH_H
+
+#include <linux/module.h>
+#include <linux/ftrace.h>
+
+#ifdef CONFIG_LIVEPATCH
+static inline int klp_check_compiler_support(void)
+{
+#if !defined(_CALL_ELF) || _CALL_ELF != 2 || !defined(CC_USING_MPROFILE_KERNEL)
+	return 1;
+#endif
+	return 0;
+}
+
+extern int klp_write_module_reloc(struct module *mod, unsigned long type,
+				   unsigned long loc, unsigned long value);
+
+static inline void klp_arch_set_pc(struct pt_regs *regs, unsigned long ip)
+{
+	regs->nip = ip;
+}
+#else
+#error Live patching support is disabled; check CONFIG_LIVEPATCH
+#endif
+
+#endif /* _ASM_POWERPC64_LIVEPATCH_H */
diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S
index 9e98aa1..f6e3ee7 100644
--- a/arch/powerpc/kernel/entry_64.S
+++ b/arch/powerpc/kernel/entry_64.S
@@ -1267,6 +1267,9 @@ _GLOBAL(ftrace_caller)
 	mflr    r3
 	std     r3, _NIP(r1)
 	std	r3, 16(r1)
+#ifdef CONFIG_LIVEPATCH
+	mr	r14,r3		/* remember old NIP */
+#endif
 	subi    r3, r3, MCOUNT_INSN_SIZE
 	mfmsr   r4
 	std     r4, _MSR(r1)
@@ -1283,7 +1286,10 @@ ftrace_call:
 	nop
 
 	ld	r3, _NIP(r1)
-	mtlr	r3
+	mtctr	r3		/* prepare to jump there */
+#ifdef CONFIG_LIVEPATCH
+	cmpd	r14,r3		/* has NIP been altered? */
+#endif
 
 	REST_8GPRS(0,r1)
 	REST_8GPRS(8,r1)
@@ -1296,6 +1302,27 @@ ftrace_call:
 	mtlr	r12
 	mr	r2,r0		/* restore callee's TOC */
 
+#ifdef CONFIG_LIVEPATCH
+	beq+	4f		/* likely(old_NIP == new_NIP) */
+
+	/* For a local call, restore this TOC after calling the patch function.
+	 * For a global call, it does not matter what we restore here,
+	 * since the global caller does its own restore right afterwards,
+	 * anyway. Just insert a KLP_return_helper frame in any case,
+	 * so a patch function can always count on the changed stack offsets.
+	 */
+	stdu	r1,-32(r1)	/* open new mini stack frame */
+	std	r0,24(r1)	/* save TOC now, unconditionally. */
+	bl	5f
+5:	mflr	r12
+	addi	r12,r12,(KLP_return_helper+4-.)@l
+	std	r12,LRSAVE(r1)
+	mtlr	r12
+	mfctr	r12		/* allow for TOC calculation in newfunc */
+	bctr
+4:
+#endif
+
 #ifdef CONFIG_FUNCTION_GRAPH_TRACER
 	stdu	r1, -112(r1)
 .globl ftrace_graph_call
@@ -1305,15 +1332,31 @@ _GLOBAL(ftrace_graph_stub)
 	addi	r1, r1, 112
 #endif
 
-	mflr	r0		/* move this LR to CTR */
-	mtctr	r0
-
 	ld	r0,LRSAVE(r1)	/* restore callee's lr at _mcount site */
 	mtlr	r0
 	bctr			/* jump after _mcount site */
 #endif /* CC_USING_MPROFILE_KERNEL */
 _GLOBAL(ftrace_stub)
 	blr
+
+#ifdef CONFIG_LIVEPATCH
+/* Helper function for local calls that are becoming global
+   due to live patching.
+   We can't simply patch the NOP after the original call,
+   because, depending on the consistency model, some kernel
+   threads may still have called the original, local function
+   *without* saving their TOC in the respective stack frame slot,
+   so the decision is made per-thread during function return by
+   maybe inserting a KLP_return_helper frame or not.
+*/
+KLP_return_helper:
+	ld	r2,24(r1)	/* restore TOC (saved by ftrace_caller) */
+	addi r1, r1, 32		/* destroy mini stack frame */
+	ld	r0,LRSAVE(r1)	/* get the real return address */
+	mtlr	r0
+	blr
+#endif
+
 #else
 _GLOBAL_TOC(_mcount)
 	/* Taken from output of objdump from lib64/glibc */
diff --git a/arch/powerpc/kernel/livepatch.c b/arch/powerpc/kernel/livepatch.c
new file mode 100644
index 0000000..564eafa
--- /dev/null
+++ b/arch/powerpc/kernel/livepatch.c
@@ -0,0 +1,38 @@
+/*
+ * livepatch.c - powerpc-specific Kernel Live Patching Core
+ *
+ * Copyright (C) 2015 SUSE
+ *
+ * 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.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, see <http://www.gnu.org/licenses/>.
+ */
+#include <linux/module.h>
+#include <asm/livepatch.h>
+
+/**
+ * klp_write_module_reloc() - write a relocation in a module
+ * @mod:       module in which the section to be modified is found
+ * @type:      ELF relocation type (see asm/elf.h)
+ * @loc:       address that the relocation should be written to
+ * @value:     relocation value (sym address + addend)
+ *
+ * This function writes a relocation to the specified location for
+ * a particular module.
+ */
+int klp_write_module_reloc(struct module *mod, unsigned long type,
+			    unsigned long loc, unsigned long value)
+{
+	/* This requires infrastructure changes; we need the loadinfos. */
+	pr_err("lpc_write_module_reloc not yet supported\n");
+	return -ENOSYS;
+}
-- 
1.8.5.6

[toc] | [prev] | [next] | [standalone]

#1317734 — Re: [PATCH v6 8/9] Implement kernel live patching for ppc64le (ABIv2)

[ added Petr to CC list ]

On Mon, 25 Jan 2016, Torsten Duwe wrote:

>   * create the appropriate files+functions
>     arch/powerpc/include/asm/livepatch.h
>         klp_check_compiler_support,
>         klp_arch_set_pc
>     arch/powerpc/kernel/livepatch.c with a stub for
>         klp_write_module_reloc
>     This is architecture-independent work in progress.
>   * introduce a fixup in arch/powerpc/kernel/entry_64.S
>     for local calls that are becoming global due to live patching.
>     And of course do the main KLP thing: return to a maybe different
>     address, possibly altered by the live patching ftrace op.
> 
> Signed-off-by: Torsten Duwe <duwe@suse.de>

Hi,

I have a few questions...

We still need Petr's patch from [1] to make livepatch work, right? Could 
you, please, add it to this patch set to make it self-sufficient?

Second, what is the situation with mcount prologue between gcc < 6 and 
gcc-6? Are there only 12 bytes in gcc-6 prologue? If yes, we need to 
change Petr's patch to make it more general and to be able to cope with 
different prologues. This is unfortunate. Either way, please mention it 
somewhere in a changelog.

I haven't reviewed the patch properly yet, but there is a comment below.

[1] http://lkml.kernel.org/g/20151203160004.GE8047@pathway.suse.cz

> +/**
> + * klp_write_module_reloc() - write a relocation in a module
> + * @mod:       module in which the section to be modified is found
> + * @type:      ELF relocation type (see asm/elf.h)
> + * @loc:       address that the relocation should be written to
> + * @value:     relocation value (sym address + addend)
> + *
> + * This function writes a relocation to the specified location for
> + * a particular module.
> + */
> +int klp_write_module_reloc(struct module *mod, unsigned long type,
> +			    unsigned long loc, unsigned long value)
> +{
> +	/* This requires infrastructure changes; we need the loadinfos. */
> +	pr_err("lpc_write_module_reloc not yet supported\n");

This is a nit, but there is no lpc_write_module_reloc. It should be 
klp_write_module_reloc.

Thanks,
Miroslav

[toc] | [prev] | [next] | [standalone]

#1317845 — Re: [PATCH v6 8/9] Implement kernel live patching for ppc64le (ABIv2)

On Tue 2016-01-26 11:50:25, Miroslav Benes wrote:
> 
> [ added Petr to CC list ]
> 
> On Mon, 25 Jan 2016, Torsten Duwe wrote:
> 
> >   * create the appropriate files+functions
> >     arch/powerpc/include/asm/livepatch.h
> >         klp_check_compiler_support,
> >         klp_arch_set_pc
> >     arch/powerpc/kernel/livepatch.c with a stub for
> >         klp_write_module_reloc
> >     This is architecture-independent work in progress.
> >   * introduce a fixup in arch/powerpc/kernel/entry_64.S
> >     for local calls that are becoming global due to live patching.
> >     And of course do the main KLP thing: return to a maybe different
> >     address, possibly altered by the live patching ftrace op.
> > 
> > Signed-off-by: Torsten Duwe <duwe@suse.de>
> 
> Hi,
> 
> I have a few questions...
> 
> We still need Petr's patch from [1] to make livepatch work, right? Could 
> you, please, add it to this patch set to make it self-sufficient?
> 
> Second, what is the situation with mcount prologue between gcc < 6 and 
> gcc-6? Are there only 12 bytes in gcc-6 prologue? If yes, we need to 
> change Petr's patch to make it more general and to be able to cope with 
> different prologues. This is unfortunate. Either way, please mention it 
> somewhere in a changelog.

I am going to update the extra patch. There is an idea to detect the
offset during build by scrips/recordmcount. This tool looks for the
ftrace locations. The offset should always be a constant that depends
on the used architecture, compiler, and compiler flags.

The tool is called post build. We might need to pass the constant
as a symbol added to the binary. The tool already adds some symbols.


Best Regards,
Petr

[toc] | [prev] | [next] | [standalone]

#1317928 — Re: [PATCH v6 8/9] Implement kernel live patching for ppc64le (ABIv2)

On Tue, Jan 26, 2016 at 01:48:53PM +0100, Petr Mladek wrote:
> On Tue 2016-01-26 11:50:25, Miroslav Benes wrote:
> > 
> > We still need Petr's patch from [1] to make livepatch work, right? Could 
> > you, please, add it to this patch set to make it self-sufficient?

It's Petr's patch, I don't want to decide how to best tackle this, see below.
I think Michael is already aware that it is needed, too.

> > Second, what is the situation with mcount prologue between gcc < 6 and 
> > gcc-6? Are there only 12 bytes in gcc-6 prologue? If yes, we need to 

Precisely, it's commit e95d0248daced44 (in http://repo.or.cz/official-gcc.git)
or svn trunk change 222352 "No need for -mprofile-kernel to save LR to stack."
It's efficient, I like it.

> I am going to update the extra patch. There is an idea to detect the
> offset during build by scrips/recordmcount. This tool looks for the
> ftrace locations. The offset should always be a constant that depends
> on the used architecture, compiler, and compiler flags.

My first idea was to check for compiler version defines, but some vendors
are rumoured to patch their compilers ;-)

> The tool is called post build. We might need to pass the constant
> as a symbol added to the binary. The tool already adds some symbols.

That's even better.
Thanks!
	Torsten

[toc] | [prev] | [next] | [standalone]

#1317941 — Re: [PATCH v6 8/9] Implement kernel live patching for ppc64le (ABIv2)

On Tue, Jan 26, 2016 at 11:50:25AM +0100, Miroslav Benes wrote:
> > + */
> > +int klp_write_module_reloc(struct module *mod, unsigned long type,
> > +			    unsigned long loc, unsigned long value)
> > +{
> > +	/* This requires infrastructure changes; we need the loadinfos. */
> > +	pr_err("lpc_write_module_reloc not yet supported\n");
> 
> This is a nit, but there is no lpc_write_module_reloc. It should be 
> klp_write_module_reloc.

Indeed. Michael, feel free to fix this on the fly or not. It needs to
disappear anyway and be replaced with functionality.

	Torsten

[toc] | [prev] | [next] | [standalone]

#1317956 — Re: [PATCH v6 8/9] Implement kernel live patching for ppc64le (ABIv2)

[ Jessica added to CC list so she is aware that there are plans to 
implement livepatch on ppc64le ]

On Tue, 26 Jan 2016, Torsten Duwe wrote:

> On Tue, Jan 26, 2016 at 11:50:25AM +0100, Miroslav Benes wrote:
> > > + */
> > > +int klp_write_module_reloc(struct module *mod, unsigned long type,
> > > +			    unsigned long loc, unsigned long value)
> > > +{
> > > +	/* This requires infrastructure changes; we need the loadinfos. */
> > > +	pr_err("lpc_write_module_reloc not yet supported\n");
> > 
> > This is a nit, but there is no lpc_write_module_reloc. It should be 
> > klp_write_module_reloc.
> 
> Indeed. Michael, feel free to fix this on the fly or not. It needs to
> disappear anyway and be replaced with functionality.

Or not at all thanks to Jessica's effort [1].

Miroslav

[1] http://lkml.kernel.org/g/1452281304-28618-1-git-send-email-jeyu@redhat.com

[toc] | [prev] | [next] | [standalone]

#1318543 — Re: Implement kernel live patching for ppc64le (ABIv2)

+++ Miroslav Benes [26/01/16 15:14 +0100]:
>
>[ Jessica added to CC list so she is aware that there are plans to
>implement livepatch on ppc64le ]
>
>On Tue, 26 Jan 2016, Torsten Duwe wrote:
>
>> On Tue, Jan 26, 2016 at 11:50:25AM +0100, Miroslav Benes wrote:
>> > > + */
>> > > +int klp_write_module_reloc(struct module *mod, unsigned long type,
>> > > +			    unsigned long loc, unsigned long value)
>> > > +{
>> > > +	/* This requires infrastructure changes; we need the loadinfos. */
>> > > +	pr_err("lpc_write_module_reloc not yet supported\n");
>> >
>> > This is a nit, but there is no lpc_write_module_reloc. It should be
>> > klp_write_module_reloc.
>>
>> Indeed. Michael, feel free to fix this on the fly or not. It needs to
>> disappear anyway and be replaced with functionality.
>
>Or not at all thanks to Jessica's effort [1].
>
>Miroslav
>
>[1] http://lkml.kernel.org/g/1452281304-28618-1-git-send-email-jeyu@redhat.com

Miroslav, thanks for the CC. Indeed, if things go well, there may be
no need to implement klp_write_module_reloc() anymore in the near
future. :-)

Jessica

[toc] | [prev] | [next] | [standalone]

#1317062 — [PATCH v6 2/9] ppc64le FTRACE_WITH_REGS implementation

Implement FTRACE_WITH_REGS for powerpc64, on ELF ABI v2.
Initial work started by Vojtech Pavlik, used with permission.

  * arch/powerpc/kernel/entry_64.S:
    - Implement an effective ftrace_caller that works from
      within the kernel binary as well as from modules.
  * arch/powerpc/kernel/ftrace.c:
    - be prepared to deal with ppc64 ELF ABI v2, especially
      calls to _mcount that result from gcc -mprofile-kernel
    - a little more error verbosity
  * arch/powerpc/kernel/module_64.c:
    - do not save the TOC pointer on the trampoline when the
      destination is ftrace_caller. This trampoline jump happens from
      a function prologue before a new stack frame is set up, so bad
      things may happen otherwise...
    - relax is_module_trampoline() to recognise the modified
      trampoline.

Signed-off-by: Torsten Duwe <duwe@suse.de>
---
 arch/powerpc/include/asm/ftrace.h |  5 +++
 arch/powerpc/kernel/entry_64.S    | 78 +++++++++++++++++++++++++++++++++++++++
 arch/powerpc/kernel/ftrace.c      | 60 +++++++++++++++++++++++++++---
 arch/powerpc/kernel/module_64.c   | 25 ++++++++++++-
 4 files changed, 161 insertions(+), 7 deletions(-)

diff --git a/arch/powerpc/include/asm/ftrace.h b/arch/powerpc/include/asm/ftrace.h
index ef89b14..50ca758 100644
--- a/arch/powerpc/include/asm/ftrace.h
+++ b/arch/powerpc/include/asm/ftrace.h
@@ -46,6 +46,8 @@
 extern void _mcount(void);
 
 #ifdef CONFIG_DYNAMIC_FTRACE
+# define FTRACE_ADDR ((unsigned long)ftrace_caller)
+# define FTRACE_REGS_ADDR FTRACE_ADDR
 static inline unsigned long ftrace_call_adjust(unsigned long addr)
 {
        /* reloction of mcount call site is the same as the address */
@@ -58,6 +60,9 @@ struct dyn_arch_ftrace {
 #endif /*  CONFIG_DYNAMIC_FTRACE */
 #endif /* __ASSEMBLY__ */
 
+#ifdef CONFIG_DYNAMIC_FTRACE_WITH_REGS
+#define ARCH_SUPPORTS_FTRACE_OPS 1
+#endif
 #endif
 
 #if defined(CONFIG_FTRACE_SYSCALLS) && defined(CONFIG_PPC64) && !defined(__ASSEMBLY__)
diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S
index e7cd043..9e98aa1 100644
--- a/arch/powerpc/kernel/entry_64.S
+++ b/arch/powerpc/kernel/entry_64.S
@@ -1213,6 +1213,7 @@ _GLOBAL(_mcount)
 	mtlr	r0
 	bctr
 
+#ifndef CC_USING_MPROFILE_KERNEL
 _GLOBAL_TOC(ftrace_caller)
 	/* Taken from output of objdump from lib64/glibc */
 	mflr	r3
@@ -1234,6 +1235,83 @@ _GLOBAL(ftrace_graph_stub)
 	ld	r0, 128(r1)
 	mtlr	r0
 	addi	r1, r1, 112
+#else
+_GLOBAL(ftrace_caller)
+	std	r0,LRSAVE(r1)
+#if defined(_CALL_ELF) && _CALL_ELF == 2
+	mflr	r0
+	bl	2f
+2:	mflr	r12
+	mtlr	r0
+	mr      r0,r2   /* save callee's TOC */
+	addis	r2,r12,(.TOC.-ftrace_caller-12)@ha
+	addi    r2,r2,(.TOC.-ftrace_caller-12)@l
+#else
+	mr	r0,r2
+#endif
+	ld	r12,LRSAVE(r1)	/* get caller's address */
+
+	stdu	r1,-SWITCH_FRAME_SIZE(r1)
+
+	std     r12, _LINK(r1)
+	SAVE_8GPRS(0,r1)
+	std	r0, 24(r1)	/* save TOC */
+	SAVE_8GPRS(8,r1)
+	SAVE_8GPRS(16,r1)
+	SAVE_8GPRS(24,r1)
+
+	addis	r3,r2,function_trace_op@toc@ha
+	addi	r3,r3,function_trace_op@toc@l
+	ld	r5,0(r3)
+
+	mflr    r3
+	std     r3, _NIP(r1)
+	std	r3, 16(r1)
+	subi    r3, r3, MCOUNT_INSN_SIZE
+	mfmsr   r4
+	std     r4, _MSR(r1)
+	mfctr   r4
+	std     r4, _CTR(r1)
+	mfxer   r4
+	std     r4, _XER(r1)
+	mr	r4, r12
+	addi    r6, r1 ,STACK_FRAME_OVERHEAD
+
+.globl ftrace_call
+ftrace_call:
+	bl	ftrace_stub
+	nop
+
+	ld	r3, _NIP(r1)
+	mtlr	r3
+
+	REST_8GPRS(0,r1)
+	REST_8GPRS(8,r1)
+	REST_8GPRS(16,r1)
+	REST_8GPRS(24,r1)
+
+	addi r1, r1, SWITCH_FRAME_SIZE
+
+	ld	r12, LRSAVE(r1)  /* get caller's address */
+	mtlr	r12
+	mr	r2,r0		/* restore callee's TOC */
+
+#ifdef CONFIG_FUNCTION_GRAPH_TRACER
+	stdu	r1, -112(r1)
+.globl ftrace_graph_call
+ftrace_graph_call:
+	b	ftrace_graph_stub
+_GLOBAL(ftrace_graph_stub)
+	addi	r1, r1, 112
+#endif
+
+	mflr	r0		/* move this LR to CTR */
+	mtctr	r0
+
+	ld	r0,LRSAVE(r1)	/* restore callee's lr at _mcount site */
+	mtlr	r0
+	bctr			/* jump after _mcount site */
+#endif /* CC_USING_MPROFILE_KERNEL */
 _GLOBAL(ftrace_stub)
 	blr
 #else
diff --git a/arch/powerpc/kernel/ftrace.c b/arch/powerpc/kernel/ftrace.c
index 080c525..310137f 100644
--- a/arch/powerpc/kernel/ftrace.c
+++ b/arch/powerpc/kernel/ftrace.c
@@ -61,8 +61,11 @@ ftrace_modify_code(unsigned long ip, unsigned int old, unsigned int new)
 		return -EFAULT;
 
 	/* Make sure it is what we expect it to be */
-	if (replaced != old)
+	if (replaced != old) {
+		pr_err("%p: replaced (%#x) != old (%#x)",
+		(void *)ip, replaced, old);
 		return -EINVAL;
+	}
 
 	/* replace the text with the new text */
 	if (patch_instruction((unsigned int *)ip, new))
@@ -106,14 +109,16 @@ static int
 __ftrace_make_nop(struct module *mod,
 		  struct dyn_ftrace *rec, unsigned long addr)
 {
-	unsigned int op;
+	unsigned int op, op0, op1, pop;
 	unsigned long entry, ptr;
 	unsigned long ip = rec->ip;
 	void *tramp;
 
 	/* read where this goes */
-	if (probe_kernel_read(&op, (void *)ip, sizeof(int)))
+	if (probe_kernel_read(&op, (void *)ip, sizeof(int))) {
+		pr_err("Fetching opcode failed.\n");
 		return -EFAULT;
+	}
 
 	/* Make sure that that this is still a 24bit jump */
 	if (!is_bl_op(op)) {
@@ -158,10 +163,46 @@ __ftrace_make_nop(struct module *mod,
 	 *
 	 * Use a b +8 to jump over the load.
 	 */
-	op = 0x48000008;	/* b +8 */
 
-	if (patch_instruction((unsigned int *)ip, op))
+	pop = 0x48000008;	/* b +8 */
+
+	/*
+	 * Check what is in the next instruction. We can see ld r2,40(r1), but
+	 * on first pass after boot we will see mflr r0.
+	 */
+	if (probe_kernel_read(&op, (void *)(ip+4), MCOUNT_INSN_SIZE)) {
+		pr_err("Fetching op failed.\n");
+		return -EFAULT;
+	}
+
+	if (op != 0xe8410028) { /* ld r2,STACK_OFFSET(r1) */
+
+		if (probe_kernel_read(&op0, (void *)(ip-8), MCOUNT_INSN_SIZE)) {
+			pr_err("Fetching op0 failed.\n");
+			return -EFAULT;
+		}
+
+		if (probe_kernel_read(&op1, (void *)(ip-4), MCOUNT_INSN_SIZE)) {
+			pr_err("Fetching op1 failed.\n");
+			return -EFAULT;
+		}
+
+		/* mflr r0 ; std r0,LRSAVE(r1) */
+		if (op0 != 0x7c0802a6 && op1 != 0xf8010010) {
+			pr_err("Unexpected instructions around bl\n"
+				"when enabling dynamic ftrace!\t"
+				"(%08x,%08x,bl,%08x)\n", op0, op1, op);
+			return -EINVAL;
+		}
+
+		/* When using -mkernel_profile there is no load to jump over */
+		pop = PPC_INST_NOP;
+	}
+
+	if (patch_instruction((unsigned int *)ip, pop)) {
+		pr_err("Patching NOP failed.\n");
 		return -EPERM;
+	}
 
 	return 0;
 }
@@ -287,6 +328,13 @@ int ftrace_make_nop(struct module *mod,
 
 #ifdef CONFIG_MODULES
 #ifdef CONFIG_PPC64
+#ifdef CONFIG_DYNAMIC_FTRACE_WITH_REGS
+int ftrace_modify_call(struct dyn_ftrace *rec, unsigned long old_addr,
+			unsigned long addr)
+{
+	return ftrace_make_call(rec, addr);
+}
+#endif
 static int
 __ftrace_make_call(struct dyn_ftrace *rec, unsigned long addr)
 {
@@ -338,7 +386,7 @@ __ftrace_make_call(struct dyn_ftrace *rec, unsigned long addr)
 
 	return 0;
 }
-#else
+#else  /* !CONFIG_PPC64: */
 static int
 __ftrace_make_call(struct dyn_ftrace *rec, unsigned long addr)
 {
diff --git a/arch/powerpc/kernel/module_64.c b/arch/powerpc/kernel/module_64.c
index 30f6be1..a8facb6 100644
--- a/arch/powerpc/kernel/module_64.c
+++ b/arch/powerpc/kernel/module_64.c
@@ -138,12 +138,25 @@ static u32 ppc64_stub_insns[] = {
 	0x4e800420			/* bctr */
 };
 
+#ifdef CC_USING_MPROFILE_KERNEL
+/* In case of _mcount calls or dynamic ftracing, Do not save the
+ * current callee's TOC (in R2) again into the original caller's stack
+ * frame during this trampoline hop. The stack frame already holds
+ * that of the original caller.  _mcount and ftrace_caller will take
+ * care of this TOC value themselves.
+ */
+#define SQUASH_TOC_SAVE_INSN(trampoline_addr) \
+	(((struct ppc64_stub_entry *)(trampoline_addr))->jump[2] = PPC_INST_NOP)
+#else
+#define SQUASH_TOC_SAVE_INSN(trampoline_addr)
+#endif
+
 #ifdef CONFIG_DYNAMIC_FTRACE
 
 static u32 ppc64_stub_mask[] = {
 	0xffff0000,
 	0xffff0000,
-	0xffffffff,
+	0x00000000,
 	0xffffffff,
 #if !defined(_CALL_ELF) || _CALL_ELF != 2
 	0xffffffff,
@@ -170,6 +183,9 @@ bool is_module_trampoline(u32 *p)
 		if ((insna & mask) != (insnb & mask))
 			return false;
 	}
+	if (insns[2] != ppc64_stub_insns[2] &&
+	    insns[2] != PPC_INST_NOP)
+		return false;
 
 	return true;
 }
@@ -619,6 +635,9 @@ int apply_relocate_add(Elf64_Shdr *sechdrs,
 					return -ENOENT;
 				if (!restore_r2((u32 *)location + 1, me))
 					return -ENOEXEC;
+				/* Squash the TOC saver for profiler calls */
+				if (!strcmp("_mcount", strtab+sym->st_name))
+					SQUASH_TOC_SAVE_INSN(value);
 			} else
 				value += local_entry_offset(sym);
 
@@ -679,6 +698,10 @@ int apply_relocate_add(Elf64_Shdr *sechdrs,
 	me->arch.tramp = stub_for_addr(sechdrs,
 				       (unsigned long)ftrace_caller,
 				       me);
+	/* ftrace_caller will take care of the TOC;
+	 * do not clobber original caller's value.
+	 */
+	SQUASH_TOC_SAVE_INSN(me->arch.tramp);
 #endif
 
 	return 0;
-- 
1.8.5.6

[toc] | [prev] | [next] | [standalone]

#1317073 — [PATCH v6 7/9] ppc64 ftrace: disable profiling for some files

This patch complements the "notrace" attribute for selected functions.
It adds -mprofile-kernel to the cc flags to be stripped from the command
line for code-patching.o and feature-fixups.o, in addition to "-pg"

Signed-off-by: Torsten Duwe <duwe@suse.de>
---
 arch/powerpc/lib/Makefile | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/powerpc/lib/Makefile b/arch/powerpc/lib/Makefile
index a47e142..98e22b2 100644
--- a/arch/powerpc/lib/Makefile
+++ b/arch/powerpc/lib/Makefile
@@ -6,8 +6,8 @@ subdir-ccflags-$(CONFIG_PPC_WERROR) := -Werror
 
 ccflags-$(CONFIG_PPC64)	:= $(NO_MINIMAL_TOC)
 
-CFLAGS_REMOVE_code-patching.o = -pg
-CFLAGS_REMOVE_feature-fixups.o = -pg
+CFLAGS_REMOVE_code-patching.o = -pg -mprofile-kernel
+CFLAGS_REMOVE_feature-fixups.o = -pg -mprofile-kernel
 
 obj-y += string.o alloc.o crtsavres.o ppc_ksyms.o code-patching.o \
 	 feature-fixups.o
-- 
1.8.5.6

[toc] | [prev] | [next] | [standalone]

#1318854 — Re: [PATCH v6 0/9] ftrace with regs + live patching for ppc64 LE (ABI v2)

On Mon, 25 Jan 2016 16:38:48 +0100
Torsten Duwe <duwe@lst.de> wrote:

> Changes since v5:
>   * extra "std r0,LRSAVE(r1)" for gcc-6
>     This makes the code compiler-agnostic.
>   * Follow Petr Mladek's suggestion to avoid
>     redefinition of HAVE_LIVEPATCH

I looked at the patches - well mostly patches 1 and 2, some quick questions

1. I know -mprofile-kernel is a big optimization win, do we need it or can
we incrementally add it?
2. Some of the hardcoded checks for opcode are hard to review, I know they've
been there in similar forms for a while. May be as an iterative step we should
give the numbers some meaning and use proper helpers for it.

I am going to give the patches a spin

Balbir Singh.

[toc] | [prev] | [next] | [standalone]

#1318911 — Re: [PATCH v6 0/9] ftrace with regs + live patching for ppc64 LE (ABI v2)

On Wed, Jan 27, 2016 at 09:51:12PM +1100, Balbir Singh wrote:
> On Mon, 25 Jan 2016 16:38:48 +0100
> Torsten Duwe <duwe@lst.de> wrote:
> 
> > Changes since v5:
> >   * extra "std r0,LRSAVE(r1)" for gcc-6
> >     This makes the code compiler-agnostic.
> >   * Follow Petr Mladek's suggestion to avoid
> >     redefinition of HAVE_LIVEPATCH
> 
> I looked at the patches - well mostly patches 1 and 2, some quick questions
> 
> 1. I know -mprofile-kernel is a big optimization win, do we need it or can
> we incrementally add it?

There's a reason why these are first ;-)
The following ones assume -mprofile-kernel is used.
The disadvantage is all relevant registers need to be saved before calling
further C code in between functions. On the Pro side, no stack frame has been
created at that point. These are assumptions made all over the ftrace-with-regs
and live patching code here.

> 2. Some of the hardcoded checks for opcode are hard to review, I know they've
> been there in similar forms for a while. May be as an iterative step we should
> give the numbers some meaning and use proper helpers for it.

Yes, Michael has already criticised that. No further literal hex constants, I promise.

> I am going to give the patches a spin

Thanks! Make sure you use a compiler that can disable -mprofile-kernel with "notrace".

	Torsten

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web