[PATCH 0/2] s390/ftrace: implement hotpatching

[PATCH 0/2] s390/ftrace: implement hotpatching

[Ilya Leoshkevich]

Hello,

This series implements the stop_machine()-less ftrace code patching
for s390. It's trickier than that of e.g. Intel, because the only
on-the-fly code patching allowed by s390 is changing a mask of a
conditional branch instruction. Patch 1 contains a new common code hook
that we need for that, patch 2 contains the actual implementation as
well as explanation of all the small details.

We would like to take this series through the s390 tree, and we need a
review on the patch 1 for that. Could you please take a look?

Best regards,
Ilya

Ilya Leoshkevich (2):
  ftrace: Introduce ftrace_need_init_nop()
  s390/ftrace: implement hotpatching

 arch/s390/include/asm/ftrace.h     |  46 +------
 arch/s390/include/asm/ftrace.lds.h |  21 +++
 arch/s390/include/asm/module.h     |   8 ++
 arch/s390/kernel/ftrace.c          | 207 ++++++++++++++++++++++++++---
 arch/s390/kernel/ftrace.h          |  24 ++++
 arch/s390/kernel/module.c          |  45 +++++++
 arch/s390/kernel/vmlinux.lds.S     |   2 +
 include/linux/ftrace.h             |  16 +++
 kernel/trace/ftrace.c              |   4 +-
 9 files changed, 314 insertions(+), 59 deletions(-)
 create mode 100644 arch/s390/include/asm/ftrace.lds.h
 create mode 100644 arch/s390/kernel/ftrace.h

-- 
2.31.1

[PATCH 1/2] ftrace: Introduce ftrace_need_init_nop()

[Ilya Leoshkevich]

Implementing live patching on s390 requires each function's prologue to
contain a very special kind of nop, which gcc and clang don't generate.
However, the current code assumes that if CC_USING_NOP_MCOUNT is
defined, then whatever the compiler generates is good enough.

Move the CC_USING_NOP_MCOUNT check into the new ftrace_need_init_nop()
macro, that the architectures can override.

An alternative solution is to disable using -mnop-mcount in the
Makefile, however, this makes the build logic (even) more complicated
and forces the arch-specific code to deal with the useless __fentry__
symbol.

Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
---
 include/linux/ftrace.h | 16 ++++++++++++++++
 kernel/trace/ftrace.c  |  4 ++--
 2 files changed, 18 insertions(+), 2 deletions(-)

diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h
index a69f363b61bf..f1ea8988a045 100644
--- a/include/linux/ftrace.h
+++ b/include/linux/ftrace.h
@@ -643,6 +643,22 @@ static inline int ftrace_disable_ftrace_graph_caller(void) { return 0; }
 extern int ftrace_make_nop(struct module *mod,
 			   struct dyn_ftrace *rec, unsigned long addr);
 
+/**
+ * ftrace_need_init_nop - return whether nop call sites should be initialized
+ *
+ * Normally the compiler's -mnop-mcount generates suitable nops, so we don't
+ * need to call ftrace_init_nop() if the code is built with that flag.
+ * Architectures where this is not always the case may define their own
+ * condition.
+ *
+ * Return must be:
+ *  0       if ftrace_init_nop() should be called
+ *  Nonzero if ftrace_init_nop() should not be called
+ */
+
+#ifndef ftrace_need_init_nop
+#define ftrace_need_init_nop() (!__is_defined(CC_USING_NOP_MCOUNT))
+#endif
 
 /**
  * ftrace_init_nop - initialize a nop call site
diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
index 7b180f61e6d3..7efbc8aaf7f6 100644
--- a/kernel/trace/ftrace.c
+++ b/kernel/trace/ftrace.c
@@ -3100,6 +3100,7 @@ ops_references_rec(struct ftrace_ops *ops, struct dyn_ftrace *rec)
 
 static int ftrace_update_code(struct module *mod, struct ftrace_page *new_pgs)
 {
+	bool init_nop = ftrace_need_init_nop();
 	struct ftrace_page *pg;
 	struct dyn_ftrace *p;
 	u64 start, stop;
@@ -3138,8 +3139,7 @@ static int ftrace_update_code(struct module *mod, struct ftrace_page *new_pgs)
 			 * Do the initial record conversion from mcount jump
 			 * to the NOP instructions.
 			 */
-			if (!__is_defined(CC_USING_NOP_MCOUNT) &&
-			    !ftrace_nop_initialize(mod, p))
+			if (init_nop && !ftrace_nop_initialize(mod, p))
 				break;
 
 			update_cnt++;
-- 
2.31.1

Re: [PATCH 1/2] ftrace: Introduce ftrace_need_init_nop()

[Steven Rostedt]

On Wed, 28 Jul 2021 23:25:45 +0200
Ilya Leoshkevich <iii@linux.ibm.com> wrote:

> Implementing live patching on s390 requires each function's prologue to
> contain a very special kind of nop, which gcc and clang don't generate.
> However, the current code assumes that if CC_USING_NOP_MCOUNT is
> defined, then whatever the compiler generates is good enough.
> 
> Move the CC_USING_NOP_MCOUNT check into the new ftrace_need_init_nop()
> macro, that the architectures can override.
> 
> An alternative solution is to disable using -mnop-mcount in the
> Makefile, however, this makes the build logic (even) more complicated
> and forces the arch-specific code to deal with the useless __fentry__
> symbol.
> 
> Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>

I'm fine with this patch, and also you taking it in via the s390 tree.

Reviewed-by: Steven Rostedt (VMware) <rostedt@goodmis.org>

-- Steve


> ---
>  include/linux/ftrace.h | 16 ++++++++++++++++
>  kernel/trace/ftrace.c  |  4 ++--
>  2 files changed, 18 insertions(+), 2 deletions(-)
> 
> diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h
> index a69f363b61bf..f1ea8988a045 100644
> --- a/include/linux/ftrace.h
> +++ b/include/linux/ftrace.h
> @@ -643,6 +643,22 @@ static inline int ftrace_disable_ftrace_graph_caller(void) { return 0; }
>  extern int ftrace_make_nop(struct module *mod,
>  			   struct dyn_ftrace *rec, unsigned long addr);
>  
> +/**
> + * ftrace_need_init_nop - return whether nop call sites should be initialized
> + *
> + * Normally the compiler's -mnop-mcount generates suitable nops, so we don't
> + * need to call ftrace_init_nop() if the code is built with that flag.
> + * Architectures where this is not always the case may define their own
> + * condition.
> + *
> + * Return must be:
> + *  0       if ftrace_init_nop() should be called
> + *  Nonzero if ftrace_init_nop() should not be called
> + */
> +
> +#ifndef ftrace_need_init_nop
> +#define ftrace_need_init_nop() (!__is_defined(CC_USING_NOP_MCOUNT))
> +#endif
>  
>  /**
>   * ftrace_init_nop - initialize a nop call site
> diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
> index 7b180f61e6d3..7efbc8aaf7f6 100644
> --- a/kernel/trace/ftrace.c
> +++ b/kernel/trace/ftrace.c
> @@ -3100,6 +3100,7 @@ ops_references_rec(struct ftrace_ops *ops, struct dyn_ftrace *rec)
>  
>  static int ftrace_update_code(struct module *mod, struct ftrace_page *new_pgs)
>  {
> +	bool init_nop = ftrace_need_init_nop();
>  	struct ftrace_page *pg;
>  	struct dyn_ftrace *p;
>  	u64 start, stop;
> @@ -3138,8 +3139,7 @@ static int ftrace_update_code(struct module *mod, struct ftrace_page *new_pgs)
>  			 * Do the initial record conversion from mcount jump
>  			 * to the NOP instructions.
>  			 */
> -			if (!__is_defined(CC_USING_NOP_MCOUNT) &&
> -			    !ftrace_nop_initialize(mod, p))
> +			if (init_nop && !ftrace_nop_initialize(mod, p))
>  				break;
>  
>  			update_cnt++;

Re: [PATCH 1/2] ftrace: Introduce ftrace_need_init_nop()

[Heiko Carstens]

On Fri, Jul 30, 2021 at 10:07:41PM -0400, Steven Rostedt wrote:
> On Wed, 28 Jul 2021 23:25:45 +0200
> Ilya Leoshkevich <iii@linux.ibm.com> wrote:
> 
> > Implementing live patching on s390 requires each function's prologue to
> > contain a very special kind of nop, which gcc and clang don't generate.
> > However, the current code assumes that if CC_USING_NOP_MCOUNT is
> > defined, then whatever the compiler generates is good enough.
> > 
> > Move the CC_USING_NOP_MCOUNT check into the new ftrace_need_init_nop()
> > macro, that the architectures can override.
> > 
> > An alternative solution is to disable using -mnop-mcount in the
> > Makefile, however, this makes the build logic (even) more complicated
> > and forces the arch-specific code to deal with the useless __fentry__
> > symbol.
> > 
> > Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
> 
> I'm fine with this patch, and also you taking it in via the s390 tree.
> 
> Reviewed-by: Steven Rostedt (VMware) <rostedt@goodmis.org>

Thanks a lot!

[PATCH 2/2] s390/ftrace: implement hotpatching

[Ilya Leoshkevich]

s390 allows hotpatching the mask of a conditional jump instruction.
Make use of this feature in order to avoid the expensive stop_machine()
call.

The new trampolines are split in 3 stages:

- A first stage is a 6-byte relative conditional long branch located at
  each function's entry point. Its offset always points to the second
  stage for the corresponding function, and its mask is either all 0s
  (ftrace off) or all 1s (ftrace on). The code for flipping the mask is
  borrowed from ftrace_{enable,disable}_ftrace_graph_caller. After
  flipping, ftrace_arch_code_modify_post_process() syncs with all the
  other CPUs by sending SIGPs.

- Second stages for vmlinux are stored in a separate part of the .text
  section reserved by the linker script, and in dynamically allocated
  memory for modules. This prevents the icache pollution. The total
  size of second stages is about 1.5% of that of the kernel image.

  Putting second stages in the .bss section is possible and decreases
  the size of the non-compressed vmlinux, but splits the kernel 1:1
  mapping, which is a bad tradeoff.

  Each second stage contains a call to the third stage, a pointer to
  the part of the intercepted function right after the first stage, and
  a pointer to an interceptor function (e.g. ftrace_caller).

  Second stages are 8-byte aligned for the future direct calls
  implementation.

- There are only two copies of the third stage: in the .text section
  for vmlinux and in dynamically allocated memory for modules. It can be
  an expoline, which is relatively large, so inlining it into each
  second stage is prohibitively expensive.

As a result of this organization, phoronix-test-suite with ftrace off
does not show any performance degradation.

At least a couple things are missing right now, but can be implemented
separately later:

- Expoline version of the third stage is supported only for z10+.
  Implementing it for earlier machines is possible, but is relatively
  complicated due to the lack of available registers.

- On z15+ trampolines may use a BIC instruction, which loads an address
  from memory and jumps to it.

Suggested-by: Sven Schnelle <svens@linux.ibm.com>
Suggested-by: Vasily Gorbik <gor@linux.ibm.com>
Co-developed-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
---
 arch/s390/include/asm/ftrace.h     |  46 +------
 arch/s390/include/asm/ftrace.lds.h |  21 +++
 arch/s390/include/asm/module.h     |   8 ++
 arch/s390/kernel/ftrace.c          | 207 ++++++++++++++++++++++++++---
 arch/s390/kernel/ftrace.h          |  24 ++++
 arch/s390/kernel/module.c          |  45 +++++++
 arch/s390/kernel/vmlinux.lds.S     |   2 +
 7 files changed, 296 insertions(+), 57 deletions(-)
 create mode 100644 arch/s390/include/asm/ftrace.lds.h
 create mode 100644 arch/s390/kernel/ftrace.h

diff --git a/arch/s390/include/asm/ftrace.h b/arch/s390/include/asm/ftrace.h
index 345cbe982a8b..e8b460f39c58 100644
--- a/arch/s390/include/asm/ftrace.h
+++ b/arch/s390/include/asm/ftrace.h
@@ -18,7 +18,6 @@
 void ftrace_caller(void);
 
 extern char ftrace_graph_caller_end;
-extern unsigned long ftrace_plt;
 extern void *ftrace_func;
 
 struct dyn_arch_ftrace { };
@@ -31,10 +30,11 @@ struct dyn_arch_ftrace { };
 
 struct module;
 struct dyn_ftrace;
-/*
- * Either -mhotpatch or -mnop-mcount is used - no explicit init is required
- */
-static inline int ftrace_init_nop(struct module *mod, struct dyn_ftrace *rec) { return 0; }
+
+bool ftrace_need_init_nop(void);
+#define ftrace_need_init_nop ftrace_need_init_nop
+
+int ftrace_init_nop(struct module *mod, struct dyn_ftrace *rec);
 #define ftrace_init_nop ftrace_init_nop
 
 static inline unsigned long ftrace_call_adjust(unsigned long addr)
@@ -42,42 +42,6 @@ static inline unsigned long ftrace_call_adjust(unsigned long addr)
 	return addr;
 }
 
-struct ftrace_insn {
-	u16 opc;
-	s32 disp;
-} __packed;
-
-static inline void ftrace_generate_nop_insn(struct ftrace_insn *insn)
-{
-#ifdef CONFIG_FUNCTION_TRACER
-	/* brcl 0,0 */
-	insn->opc = 0xc004;
-	insn->disp = 0;
-#endif
-}
-
-static inline int is_ftrace_nop(struct ftrace_insn *insn)
-{
-#ifdef CONFIG_FUNCTION_TRACER
-	if (insn->disp == 0)
-		return 1;
-#endif
-	return 0;
-}
-
-static inline void ftrace_generate_call_insn(struct ftrace_insn *insn,
-					     unsigned long ip)
-{
-#ifdef CONFIG_FUNCTION_TRACER
-	unsigned long target;
-
-	/* brasl r0,ftrace_caller */
-	target = is_module_addr((void *) ip) ? ftrace_plt : FTRACE_ADDR;
-	insn->opc = 0xc005;
-	insn->disp = (target - ip) / 2;
-#endif
-}
-
 /*
  * Even though the system call numbers are identical for s390/s390x a
  * different system call table is used for compat tasks. This may lead
diff --git a/arch/s390/include/asm/ftrace.lds.h b/arch/s390/include/asm/ftrace.lds.h
new file mode 100644
index 000000000000..90b3d86227ca
--- /dev/null
+++ b/arch/s390/include/asm/ftrace.lds.h
@@ -0,0 +1,21 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef DIV_ROUND_UP
+#define DIV_ROUND_UP(n, d) (((n) + (d) - 1) / (d))
+#endif
+
+#define SIZEOF_MCOUNT_LOC_ENTRY 8
+#define SIZEOF_FTRACE_HOTPATCH_TRAMPOLINE 24
+#define FTRACE_HOTPATCH_TRAMPOLINES_SIZE(n)                                    \
+	DIV_ROUND_UP(SIZEOF_FTRACE_HOTPATCH_TRAMPOLINE * (n),                  \
+		     SIZEOF_MCOUNT_LOC_ENTRY)
+
+#ifdef CONFIG_FUNCTION_TRACER
+#define FTRACE_HOTPATCH_TRAMPOLINES_TEXT                                       \
+	. = ALIGN(8);                                                          \
+	__ftrace_hotpatch_trampolines_start = .;                               \
+	. = . + FTRACE_HOTPATCH_TRAMPOLINES_SIZE(__stop_mcount_loc -           \
+						 __start_mcount_loc);          \
+	__ftrace_hotpatch_trampolines_end = .;
+#else
+#define FTRACE_HOTPATCH_TRAMPOLINES_TEXT
+#endif
diff --git a/arch/s390/include/asm/module.h b/arch/s390/include/asm/module.h
index 71e57b8af757..9f1eea15872c 100644
--- a/arch/s390/include/asm/module.h
+++ b/arch/s390/include/asm/module.h
@@ -28,6 +28,14 @@ struct mod_arch_specific {
 	int nsyms;
 	/* Additional symbol information (got and plt offsets). */
 	struct mod_arch_syminfo *syminfo;
+#ifdef CONFIG_FUNCTION_TRACER
+	/* Start of memory reserved for ftrace hotpatch trampolines. */
+	struct ftrace_hotpatch_trampoline *trampolines_start;
+	/* End of memory reserved for ftrace hotpatch trampolines. */
+	struct ftrace_hotpatch_trampoline *trampolines_end;
+	/* Next unused ftrace hotpatch trampoline slot. */
+	struct ftrace_hotpatch_trampoline *next_trampoline;
+#endif /* CONFIG_FUNCTION_TRACER */
 };
 
 #endif /* _ASM_S390_MODULE_H */
diff --git a/arch/s390/kernel/ftrace.c b/arch/s390/kernel/ftrace.c
index 2d8f595d9196..5d3fa255533a 100644
--- a/arch/s390/kernel/ftrace.c
+++ b/arch/s390/kernel/ftrace.c
@@ -18,8 +18,11 @@
 #include <trace/syscall.h>
 #include <asm/asm-offsets.h>
 #include <asm/cacheflush.h>
+#include <asm/ftrace.lds.h>
+#include <asm/nospec-branch.h>
 #include <asm/set_memory.h>
 #include "entry.h"
+#include "ftrace.h"
 
 /*
  * To generate function prologue either gcc's hotpatch feature (since gcc 4.8)
@@ -41,7 +44,115 @@
  */
 
 void *ftrace_func __read_mostly = ftrace_stub;
-unsigned long ftrace_plt;
+struct ftrace_insn {
+	u16 opc;
+	s32 disp;
+} __packed;
+
+asm(
+	"ftrace_shared_hotpatch_trampoline_br:\n"
+	"	lmg	%r0,%r1,2(%r1)\n"
+	"	br	%r1\n"
+	"ftrace_shared_hotpatch_trampoline_br_end:\n"
+);
+
+#ifdef CONFIG_EXPOLINE
+asm(
+	"ftrace_shared_hotpatch_trampoline_exrl:\n"
+	"	lmg	%r0,%r1,2(%r1)\n"
+	"	.insn	ril,0xc60000000000,%r0,0f\n" /* exrl */
+	"	j	.\n"
+	"0:	br	%r1\n"
+	"ftrace_shared_hotpatch_trampoline_exrl_end:\n"
+);
+#endif /* CONFIG_EXPOLINE */
+
+#ifdef CONFIG_MODULES
+static char *ftrace_plt;
+
+asm(
+	"	.data\n"
+	"ftrace_plt_template:\n"
+	"	basr	%r1,%r0\n"
+	"	lg	%r1,0f-.(%r1)\n"
+	"	br	%r1\n"
+	"0:	.quad	ftrace_caller\n"
+	"ftrace_plt_template_end:\n"
+	"	.previous\n"
+);
+#endif /* CONFIG_MODULES */
+
+static const char *ftrace_shared_hotpatch_trampoline(const char **end)
+{
+#ifdef CONFIG_EXPOLINE
+	if (!nospec_disable) {
+		if (test_facility(18)) {
+			if (end)
+				*end = ftrace_shared_hotpatch_trampoline_exrl_end;
+			return ftrace_shared_hotpatch_trampoline_exrl;
+		}
+		return NULL;
+	}
+#endif /* CONFIG_EXPOLINE */
+	if (end)
+		*end = ftrace_shared_hotpatch_trampoline_br_end;
+	return ftrace_shared_hotpatch_trampoline_br;
+}
+
+bool ftrace_need_init_nop(void)
+{
+	return ftrace_shared_hotpatch_trampoline(NULL);
+}
+
+int ftrace_init_nop(struct module *mod, struct dyn_ftrace *rec)
+{
+	static struct ftrace_hotpatch_trampoline *next_vmlinux_trampoline =
+		__ftrace_hotpatch_trampolines_start;
+	static const char orig[6] = { 0xc0, 0x04, 0x00, 0x00, 0x00, 0x00 };
+	static struct ftrace_hotpatch_trampoline *trampoline;
+	struct ftrace_hotpatch_trampoline **next_trampoline;
+	struct ftrace_hotpatch_trampoline *trampolines_end;
+	struct ftrace_hotpatch_trampoline tmp;
+	struct ftrace_insn *insn;
+	const char *shared;
+	s32 disp;
+
+	BUILD_BUG_ON(sizeof(struct ftrace_hotpatch_trampoline) !=
+		     SIZEOF_FTRACE_HOTPATCH_TRAMPOLINE);
+
+	next_trampoline = &next_vmlinux_trampoline;
+	trampolines_end = __ftrace_hotpatch_trampolines_end;
+	shared = ftrace_shared_hotpatch_trampoline(NULL);
+#ifdef CONFIG_MODULES
+	if (mod) {
+		next_trampoline = &mod->arch.next_trampoline;
+		trampolines_end = mod->arch.trampolines_end;
+		shared = ftrace_plt;
+	}
+#endif
+
+	if (WARN_ON_ONCE(*next_trampoline >= trampolines_end))
+		return -ENOMEM;
+	trampoline = (*next_trampoline)++;
+
+	/* Check for the compiler-generated fentry nop (brcl 0, .). */
+	if (WARN_ON_ONCE(memcmp((const void *)rec->ip, &orig, sizeof(orig))))
+		return -EINVAL;
+
+	/* Generate the trampoline. */
+	tmp.brasl_opc = 0xc015; /* brasl %r1, shared */
+	tmp.brasl_disp = (shared - (const char *)&trampoline->brasl_opc) / 2;
+	tmp.interceptor = FTRACE_ADDR;
+	tmp.rest_of_intercepted_function = rec->ip + sizeof(struct ftrace_insn);
+	s390_kernel_write(trampoline, &tmp, sizeof(tmp));
+
+	/* Generate a jump to the trampoline. */
+	disp = ((char *)trampoline - (char *)rec->ip) / 2;
+	insn = (struct ftrace_insn *)rec->ip;
+	s390_kernel_write(&insn->disp, &disp, sizeof(disp));
+
+	return 0;
+}
 
 int ftrace_modify_call(struct dyn_ftrace *rec, unsigned long old_addr,
 		       unsigned long addr)
@@ -49,11 +160,45 @@ int ftrace_modify_call(struct dyn_ftrace *rec, unsigned long old_addr,
 	return 0;
 }
 
+static void ftrace_generate_nop_insn(struct ftrace_insn *insn)
+{
+	/* brcl 0,0 */
+	insn->opc = 0xc004;
+	insn->disp = 0;
+}
+
+static void ftrace_generate_call_insn(struct ftrace_insn *insn,
+				      unsigned long ip)
+{
+	unsigned long target;
+
+	/* brasl r0,ftrace_caller */
+	target = FTRACE_ADDR;
+#ifdef CONFIG_MODULES
+	if (is_module_addr((void *)ip))
+		target = (unsigned long)ftrace_plt;
+#endif /* CONFIG_MODULES */
+	insn->opc = 0xc005;
+	insn->disp = (target - ip) / 2;
+}
+
+static void brcl_disable(void *brcl)
+{
+	u8 op = 0x04; /* set mask field to zero */
+
+	s390_kernel_write((char *)brcl + 1, &op, sizeof(op));
+}
+
 int ftrace_make_nop(struct module *mod, struct dyn_ftrace *rec,
 		    unsigned long addr)
 {
 	struct ftrace_insn orig, new, old;
 
+	if (ftrace_shared_hotpatch_trampoline(NULL)) {
+		brcl_disable((void *)rec->ip);
+		return 0;
+	}
+
 	if (copy_from_kernel_nofault(&old, (void *) rec->ip, sizeof(old)))
 		return -EFAULT;
 	/* Replace ftrace call with a nop. */
@@ -67,10 +212,22 @@ int ftrace_make_nop(struct module *mod, struct dyn_ftrace *rec,
 	return 0;
 }
 
+static void brcl_enable(void *brcl)
+{
+	u8 op = 0xf4; /* set mask field to all ones */
+
+	s390_kernel_write((char *)brcl + 1, &op, sizeof(op));
+}
+
 int ftrace_make_call(struct dyn_ftrace *rec, unsigned long addr)
 {
 	struct ftrace_insn orig, new, old;
 
+	if (ftrace_shared_hotpatch_trampoline(NULL)) {
+		brcl_enable((void *)rec->ip);
+		return 0;
+	}
+
 	if (copy_from_kernel_nofault(&old, (void *) rec->ip, sizeof(old)))
 		return -EFAULT;
 	/* Replace nop with an ftrace call. */
@@ -95,22 +252,44 @@ int __init ftrace_dyn_arch_init(void)
 	return 0;
 }
 
+void arch_ftrace_update_code(int command)
+{
+	if (ftrace_shared_hotpatch_trampoline(NULL))
+		ftrace_modify_all_code(command);
+	else
+		ftrace_run_stop_machine(command);
+}
+
+static void __ftrace_sync(void *dummy)
+{
+}
+
+int ftrace_arch_code_modify_post_process(void)
+{
+	if (ftrace_shared_hotpatch_trampoline(NULL)) {
+		/* Send SIGP to the other CPUs, so they see the new code. */
+		smp_call_function(__ftrace_sync, NULL, 1);
+	}
+	return 0;
+}
+
 #ifdef CONFIG_MODULES
 
 static int __init ftrace_plt_init(void)
 {
-	unsigned int *ip;
+	const char *start, *end;
 
-	ftrace_plt = (unsigned long) module_alloc(PAGE_SIZE);
+	ftrace_plt = module_alloc(PAGE_SIZE);
 	if (!ftrace_plt)
 		panic("cannot allocate ftrace plt\n");
-	ip = (unsigned int *) ftrace_plt;
-	ip[0] = 0x0d10e310; /* basr 1,0; lg 1,10(1); br 1 */
-	ip[1] = 0x100a0004;
-	ip[2] = 0x07f10000;
-	ip[3] = FTRACE_ADDR >> 32;
-	ip[4] = FTRACE_ADDR & 0xffffffff;
-	set_memory_ro(ftrace_plt, 1);
+
+	start = ftrace_shared_hotpatch_trampoline(&end);
+	if (!start) {
+		start = ftrace_plt_template;
+		end = ftrace_plt_template_end;
+	}
+	memcpy(ftrace_plt, start, end - start);
+	set_memory_ro((unsigned long)ftrace_plt, 1);
 	return 0;
 }
 device_initcall(ftrace_plt_init);
@@ -147,17 +326,13 @@ NOKPROBE_SYMBOL(prepare_ftrace_return);
  */
 int ftrace_enable_ftrace_graph_caller(void)
 {
-	u8 op = 0x04; /* set mask field to zero */
-
-	s390_kernel_write(__va(ftrace_graph_caller)+1, &op, sizeof(op));
+	brcl_disable(__va(ftrace_graph_caller));
 	return 0;
 }
 
 int ftrace_disable_ftrace_graph_caller(void)
 {
-	u8 op = 0xf4; /* set mask field to all ones */
-
-	s390_kernel_write(__va(ftrace_graph_caller)+1, &op, sizeof(op));
+	brcl_enable(__va(ftrace_graph_caller));
 	return 0;
 }
 
diff --git a/arch/s390/kernel/ftrace.h b/arch/s390/kernel/ftrace.h
new file mode 100644
index 000000000000..7f75a9616406
--- /dev/null
+++ b/arch/s390/kernel/ftrace.h
@@ -0,0 +1,24 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _FTRACE_H
+#define _FTRACE_H
+
+#include <asm/types.h>
+
+struct ftrace_hotpatch_trampoline {
+	u16 brasl_opc;
+	s32 brasl_disp;
+	s16: 16;
+	u64 rest_of_intercepted_function;
+	u64 interceptor;
+} __packed;
+
+extern struct ftrace_hotpatch_trampoline __ftrace_hotpatch_trampolines_start[];
+extern struct ftrace_hotpatch_trampoline __ftrace_hotpatch_trampolines_end[];
+extern const char ftrace_shared_hotpatch_trampoline_br[];
+extern const char ftrace_shared_hotpatch_trampoline_br_end[];
+extern const char ftrace_shared_hotpatch_trampoline_exrl[];
+extern const char ftrace_shared_hotpatch_trampoline_exrl_end[];
+extern const char ftrace_plt_template[];
+extern const char ftrace_plt_template_end[];
+
+#endif /* _FTRACE_H */
diff --git a/arch/s390/kernel/module.c b/arch/s390/kernel/module.c
index 4055f1c49814..b01ba460b7ca 100644
--- a/arch/s390/kernel/module.c
+++ b/arch/s390/kernel/module.c
@@ -14,6 +14,7 @@
 #include <linux/elf.h>
 #include <linux/vmalloc.h>
 #include <linux/fs.h>
+#include <linux/ftrace.h>
 #include <linux/string.h>
 #include <linux/kernel.h>
 #include <linux/kasan.h>
@@ -23,6 +24,8 @@
 #include <asm/alternative.h>
 #include <asm/nospec-branch.h>
 #include <asm/facility.h>
+#include <asm/ftrace.lds.h>
+#include <asm/set_memory.h>
 
 #if 0
 #define DEBUGP printk
@@ -48,6 +51,13 @@ void *module_alloc(unsigned long size)
 	return p;
 }
 
+#ifdef CONFIG_FUNCTION_TRACER
+void module_arch_cleanup(struct module *mod)
+{
+	module_memfree(mod->arch.trampolines_start);
+}
+#endif
+
 void module_arch_freeing_init(struct module *mod)
 {
 	if (is_livepatch_module(mod) &&
@@ -466,6 +476,30 @@ int apply_relocate_add(Elf_Shdr *sechdrs, const char *strtab,
 				    write);
 }
 
+#ifdef CONFIG_FUNCTION_TRACER
+static int module_alloc_ftrace_hotpatch_trampolines(struct module *me,
+						    const Elf_Shdr *s)
+{
+	char *start, *end;
+	int numpages;
+	size_t size;
+
+	size = FTRACE_HOTPATCH_TRAMPOLINES_SIZE(s->sh_size);
+	numpages = DIV_ROUND_UP(size, PAGE_SIZE);
+	start = module_alloc(numpages * PAGE_SIZE);
+	if (!start)
+		return -ENOMEM;
+	set_memory_ro((unsigned long)start, numpages);
+	end = start + size;
+
+	me->arch.trampolines_start = (struct ftrace_hotpatch_trampoline *)start;
+	me->arch.trampolines_end = (struct ftrace_hotpatch_trampoline *)end;
+	me->arch.next_trampoline = me->arch.trampolines_start;
+
+	return 0;
+}
+#endif /* CONFIG_FUNCTION_TRACER */
+
 int module_finalize(const Elf_Ehdr *hdr,
 		    const Elf_Shdr *sechdrs,
 		    struct module *me)
@@ -473,6 +507,9 @@ int module_finalize(const Elf_Ehdr *hdr,
 	const Elf_Shdr *s;
 	char *secstrings, *secname;
 	void *aseg;
+#ifdef CONFIG_FUNCTION_TRACER
+	int ret;
+#endif
 
 	if (IS_ENABLED(CONFIG_EXPOLINE) &&
 	    !nospec_disable && me->arch.plt_size) {
@@ -507,6 +544,14 @@ int module_finalize(const Elf_Ehdr *hdr,
 		if (IS_ENABLED(CONFIG_EXPOLINE) &&
 		    (str_has_prefix(secname, ".s390_return")))
 			nospec_revert(aseg, aseg + s->sh_size);
+
+#ifdef CONFIG_FUNCTION_TRACER
+		if (!strcmp(FTRACE_CALLSITE_SECTION, secname)) {
+			ret = module_alloc_ftrace_hotpatch_trampolines(me, s);
+			if (ret < 0)
+				return ret;
+		}
+#endif /* CONFIG_FUNCTION_TRACER */
 	}
 
 	jump_label_apply_nops(me);
diff --git a/arch/s390/kernel/vmlinux.lds.S b/arch/s390/kernel/vmlinux.lds.S
index b18288d26ca8..b6caa810af3a 100644
--- a/arch/s390/kernel/vmlinux.lds.S
+++ b/arch/s390/kernel/vmlinux.lds.S
@@ -5,6 +5,7 @@
 
 #include <asm/thread_info.h>
 #include <asm/page.h>
+#include <asm/ftrace.lds.h>
 
 /*
  * Put .bss..swapper_pg_dir as the first thing in .bss. This will
@@ -46,6 +47,7 @@ SECTIONS
 		KPROBES_TEXT
 		IRQENTRY_TEXT
 		SOFTIRQENTRY_TEXT
+		FTRACE_HOTPATCH_TRAMPOLINES_TEXT
 		*(.text.*_indirect_*)
 		*(.fixup)
 		*(.gnu.warning)
-- 
2.31.1

Re: [PATCH 0/2] s390/ftrace: implement hotpatching

[Heiko Carstens]

On Wed, Jul 28, 2021 at 11:25:44PM +0200, Ilya Leoshkevich wrote:
> Hello,
> 
> This series implements the stop_machine()-less ftrace code patching
> for s390. It's trickier than that of e.g. Intel, because the only
> on-the-fly code patching allowed by s390 is changing a mask of a
> conditional branch instruction. Patch 1 contains a new common code hook
> that we need for that, patch 2 contains the actual implementation as
> well as explanation of all the small details.
> 
> We would like to take this series through the s390 tree, and we need a
> review on the patch 1 for that. Could you please take a look?
> 
> Best regards,
> Ilya
> 
> Ilya Leoshkevich (2):
>   ftrace: Introduce ftrace_need_init_nop()
>   s390/ftrace: implement hotpatching
> 
>  arch/s390/include/asm/ftrace.h     |  46 +------
>  arch/s390/include/asm/ftrace.lds.h |  21 +++
>  arch/s390/include/asm/module.h     |   8 ++
>  arch/s390/kernel/ftrace.c          | 207 ++++++++++++++++++++++++++---
>  arch/s390/kernel/ftrace.h          |  24 ++++
>  arch/s390/kernel/module.c          |  45 +++++++
>  arch/s390/kernel/vmlinux.lds.S     |   2 +
>  include/linux/ftrace.h             |  16 +++
>  kernel/trace/ftrace.c              |   4 +-
>  9 files changed, 314 insertions(+), 59 deletions(-)
>  create mode 100644 arch/s390/include/asm/ftrace.lds.h
>  create mode 100644 arch/s390/kernel/ftrace.h

Series applied, thank you!