Skip to contentSkip to conversation

riscv: jump_label: Fixup & Optimization

PATCH In review [email protected] 3 participants 8 messages Jan 26, 2023 – Feb 22, 2023

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

v3 cover riscv: jump_label: Fixup & Optimization

From: Guo Ren <[email protected]>

Patch 1 is the fixup patch should be merged into stable-tree. Patch 2 is a continuous optimization for jump_label patch_text atomicity.

Changes in V3:
- Correct the typo C.JAL -> C.J (Thx Jessica)
- Fixup compile error when CONFIG_RISCV_ISA_C=n
- Rebase on riscv for-next (20230127)

Changes in V2:
https://lore.kernel.org/linux-riscv/[email protected]/

Changes in V1:
https://lore.kernel.org/linux-riscv/[email protected]/

Andy Chiu (1):
riscv: jump_label: Fixup unaligned arch_static_branch function

Guo Ren (1):
riscv: jump_label: Optimize the code size with compressed instruction

 arch/riscv/include/asm/jump_label.h | 14 ++++++++++++--
 arch/riscv/kernel/jump_label.c      | 30 +++++++++++++++++++++++++++--
 2 files changed, 40 insertions(+), 4 deletions(-)
Signature · 1 line
2.36.1
#1 lore raw

PATCH v3 1/2 riscv: jump_label: Fixup unaligned arch_static_branch function

From: Andy Chiu <[email protected]>

Runtime code patching must be done at a naturally aligned address, or we may execute on a partial instruction.

We have encountered problems traced back to static jump functions during the test. We switched the tracer randomly for every 1~5 seconds on a dual-core QEMU setup and found the kernel sucking at a static branch where it jumps to itself.

The reason is that the static branch was 2-byte but not 4-byte aligned. Then, the kernel would patch the instruction, either J or NOP, with two half-word stores if the machine does not have efficient unaligned accesses. Thus, moments exist where half of the NOP mixes with the other half of the J when transitioning the branch. In our particular case, on a little-endian machine, the upper half of the NOP was mixed with the lower part of the J when enabling the branch, resulting in a jump that jumped to itself. Conversely, it would result in a HINT instruction when disabling the branch, but it might not be observable.

ARM64 does not have this problem since all instructions must be 4-byte aligned.

Fixes ebc00dde8a97 ("riscv: Add jump-label implementation")
---
 arch/riscv/include/asm/jump_label.h | 2 ++
 1 file changed, 2 insertions(+)
arch/riscv/include/asm/jump_label.h +2 -0
index 6d58bbb5da46..14a5ea8d8ef0 100644
@@ -18,6 +18,7 @@ static __always_inline bool arch_static_branch(struct static_key * const key,
 					       const bool branch)
 {
 	asm_volatile_goto(
+		"	.align		2			\n\t"
 		"	.option push				\n\t"
 		"	.option norelax				\n\t"
 		"	.option norvc				\n\t"
@@ -39,6 +40,7 @@ static __always_inline bool arch_static_branch_jump(struct static_key * const ke
 						    const bool branch)
 {
 	asm_volatile_goto(
+		"	.align		2			\n\t"
 		"	.option push				\n\t"
 		"	.option norelax				\n\t"
 		"	.option norvc				\n\t"
Signature · 1 line
2.36.1
#6 lore raw

PATCH v3 2/2 riscv: jump_label: Optimize the code size with compressed instruction

In reply to [email protected]

From: Guo Ren <[email protected]>

Reduce the size of the static branch instruction and prevent atomic update problems when CONFIG_RISCV_ISA_C=y. It also reduces the jump range from 1MB to 4KB, but 4KB is enough for the current riscv requirement.

---
 arch/riscv/include/asm/jump_label.h | 16 +++++++++++----
 arch/riscv/kernel/jump_label.c      | 30 +++++++++++++++++++++++++++--
 2 files changed, 40 insertions(+), 6 deletions(-)
arch/riscv/include/asm/jump_label.h +12 -4
index 14a5ea8d8ef0..afc58c31d02b 100644
@@ -12,17 +12,23 @@
 #include <linux/types.h>
 #include <asm/asm.h>
 
+#ifdef CONFIG_RISCV_ISA_C
+#define JUMP_LABEL_NOP_SIZE 2
+#else
 #define JUMP_LABEL_NOP_SIZE 4
+#endif
 
 static __always_inline bool arch_static_branch(struct static_key * const key,
 					       const bool branch)
 {
 	asm_volatile_goto(
-		"	.align		2			\n\t"
 		"	.option push				\n\t"
 		"	.option norelax				\n\t"
-		"	.option norvc				\n\t"
+#ifdef CONFIG_RISCV_ISA_C
+		"1:	c.nop					\n\t"
+#else
 		"1:	nop					\n\t"
+#endif
 		"	.option pop				\n\t"
 		"	.pushsection	__jump_table, \"aw\"	\n\t"
 		"	.align		" RISCV_LGPTR "		\n\t"
@@ -40,11 +46,13 @@ static __always_inline bool arch_static_branch_jump(struct static_key * const ke
 						    const bool branch)
 {
 	asm_volatile_goto(
-		"	.align		2			\n\t"
 		"	.option push				\n\t"
 		"	.option norelax				\n\t"
-		"	.option norvc				\n\t"
+#ifdef CONFIG_RISCV_ISA_C
+		"1:	c.j		%l[label]		\n\t"
+#else
 		"1:	jal		zero, %l[label]		\n\t"
+#endif
 		"	.option pop				\n\t"
 		"	.pushsection	__jump_table, \"aw\"	\n\t"
 		"	.align		" RISCV_LGPTR "		\n\t"
arch/riscv/kernel/jump_label.c +28 -2
index e6694759dbd0..08f42c49e3a0 100644
@@ -11,26 +11,52 @@
 #include <asm/bug.h>
 #include <asm/patch.h>
 
+#ifdef CONFIG_RISCV_ISA_C
+#define RISCV_INSN_NOP 0x0001U
+#define RISCV_INSN_C_J 0xa001U
+#else
 #define RISCV_INSN_NOP 0x00000013U
 #define RISCV_INSN_JAL 0x0000006fU
+#endif
 
 void arch_jump_label_transform(struct jump_entry *entry,
 			       enum jump_label_type type)
 {
 	void *addr = (void *)jump_entry_code(entry);
+#ifdef CONFIG_RISCV_ISA_C
+	u16 insn;
+#else
 	u32 insn;
+#endif
 
 	if (type == JUMP_LABEL_JMP) {
 		long offset = jump_entry_target(entry) - jump_entry_code(entry);
-
-		if (WARN_ON(offset & 1 || offset < -524288 || offset >= 524288))
+		if (WARN_ON(offset & 1 || offset < -2048 || offset >= 2048))
 			return;
 
+#ifdef CONFIG_RISCV_ISA_C
+		/*
+		 * 001 | imm[11|4|9:8|10|6|7|3:1|5] 01 - C.J
+		 */
+		insn = RISCV_INSN_C_J |
+			(((u16)offset & GENMASK(5, 5)) >> (5 - 2)) |
+			(((u16)offset & GENMASK(3, 1)) << (3 - 1)) |
+			(((u16)offset & GENMASK(7, 7)) >> (7 - 6)) |
+			(((u16)offset & GENMASK(6, 6)) << (7 - 6)) |
+			(((u16)offset & GENMASK(10, 10)) >> (10 - 8)) |
+			(((u16)offset & GENMASK(9, 8)) << (9 - 8)) |
+			(((u16)offset & GENMASK(4, 4)) << (11 - 4)) |
+			(((u16)offset & GENMASK(11, 11)) << (12 - 11));
+#else
+		/*
+		 * imm[20|10:1|11|19:12] | rd | 1101111 - JAL
+		 */
 		insn = RISCV_INSN_JAL |
 			(((u32)offset & GENMASK(19, 12)) << (12 - 12)) |
 			(((u32)offset & GENMASK(11, 11)) << (20 - 11)) |
 			(((u32)offset & GENMASK(10,  1)) << (21 -  1)) |
 			(((u32)offset & GENMASK(20, 20)) << (31 - 20));
+#endif
 	} else {
 		insn = RISCV_INSN_NOP;
 	}
Signature · 1 line
2.36.1
Reviewed-by on v3 1/2
#2 lore raw

Re: [PATCH -next V3 1/2] riscv: jump_label: Fixup unaligned arch_static_branch function

In reply to [email protected]
Quoted 20 lines · [email protected] writes:

From: Andy Chiu <[email protected]>

Runtime code patching must be done at a naturally aligned address, or we may execute on a partial instruction.

We have encountered problems traced back to static jump functions during the test. We switched the tracer randomly for every 1~5 seconds on a dual-core QEMU setup and found the kernel sucking at a static branch where it jumps to itself.

The reason is that the static branch was 2-byte but not 4-byte aligned. Then, the kernel would patch the instruction, either J or NOP, with two half-word stores if the machine does not have efficient unaligned accesses. Thus, moments exist where half of the NOP mixes with the other half of the J when transitioning the branch. In our particular case, on a little-endian machine, the upper half of the NOP was mixed with the lower part of the J when enabling the branch, resulting in a jump that jumped to itself. Conversely, it would result in a HINT instruction when disabling the branch, but it might not be observable.

Quoted 2 lines

ARM64 does not have this problem since all instructions must be 4-byte aligned.

Nice catch! And I guess this is an issue for kprobes as well, no? I.e. in general replacing 32b insns with an ebreak. This is only valid for natural aligned 32b insns?

@Guo I don't see the point of doing a series for this, and asking the maintainers to "pick this patch to stable, and the other for next". Isn't that just more work for the maintainers/reviewers?

Björn

author
#3 lore raw

Re: [PATCH -next V3 1/2] riscv: jump_label: Fixup unaligned arch_static_branch function

Quoted 28 lines · Björn Töpel
Quoted 20 lines · [email protected] writes:

From: Andy Chiu <[email protected]>

Runtime code patching must be done at a naturally aligned address, or we may execute on a partial instruction.

We have encountered problems traced back to static jump functions during the test. We switched the tracer randomly for every 1~5 seconds on a dual-core QEMU setup and found the kernel sucking at a static branch where it jumps to itself.

The reason is that the static branch was 2-byte but not 4-byte aligned. Then, the kernel would patch the instruction, either J or NOP, with two half-word stores if the machine does not have efficient unaligned accesses. Thus, moments exist where half of the NOP mixes with the other half of the J when transitioning the branch. In our particular case, on a little-endian machine, the upper half of the NOP was mixed with the lower part of the J when enabling the branch, resulting in a jump that jumped to itself. Conversely, it would result in a HINT instruction when disabling the branch, but it might not be observable.

Quoted 2 lines

ARM64 does not have this problem since all instructions must be 4-byte aligned.

Quoted 7 lines

Nice catch! And I guess this is an issue for kprobes as well, no? I.e. in general replacing 32b insns with an ebreak. This is only valid for natural aligned 32b insns?

@Guo I don't see the point of doing a series for this, and asking the maintainers to "pick this patch to stable, and the other for next". Isn't that just more work for the maintainers/reviewers?

If these two patches are separated, they are all fixup that issue and competed with each other. Making my patch an optimization patch must depend on it. That's why I put them in one series.

Quoted 3 lines

Björn

Signature · 2 lines
Best Regards
 Guo Ren
#4 lore raw

Re: [PATCH -next V3 1/2] riscv: jump_label: Fixup unaligned arch_static_branch function

Trimming Cc.

Quoted 6 lines · Guo Ren
Quoted 3 lines

@Guo I don't see the point of doing a series for this, and asking the maintainers to "pick this patch to stable, and the other for next". Isn't that just more work for the maintainers/reviewers?

If these two patches are separated, they are all fixup that issue and competed with each other. Making my patch an optimization patch must depend on it. That's why I put them in one series.

They are not depedent at all, and not "fixup". The first is a fix, and should go into -fixes ASAP. The other patch is completely stand-alone, and an optimization (maybe). If that go in, it's for -next.

Having them as separate patches, makes it easier for reviewers/maintainers. Now, with your approach there's more work (cognitive, and manual) for others.

Björn

_______________________________________________
linux-riscv mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-riscv

author
#5 lore raw

Re: [PATCH -next V3 1/2] riscv: jump_label: Fixup unaligned arch_static_branch function

Quoted 12 lines · Björn Töpel

Trimming Cc.

Quoted 6 lines · Guo Ren
Quoted 3 lines

@Guo I don't see the point of doing a series for this, and asking the maintainers to "pick this patch to stable, and the other for next". Isn't that just more work for the maintainers/reviewers?

If these two patches are separated, they are all fixup that issue and competed with each other. Making my patch an optimization patch must depend on it. That's why I put them in one series.

Quoted 7 lines

They are not depedent at all, and not "fixup". The first is a fix, and should go into -fixes ASAP. The other patch is completely stand-alone, and an optimization (maybe). If that go in, it's for -next.

Having them as separate patches, makes it easier for reviewers/maintainers. Now, with your approach there's more work (cognitive, and manual) for others.

Okay, I would separate them.

Quoted 3 lines

Björn

Signature · 7 lines
Best Regards
 Guo Ren

_______________________________________________
linux-riscv mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-riscv
#7 lore raw
In reply to [email protected]
This series was applied to riscv/linux.git (for-next) by Palmer Dabbelt <[email protected]>:

Hello:

This series was applied to riscv/linux.git (for-next) by Palmer Dabbelt <[email protected]>:

Quoted 6 lines · On Thu, 26 Jan 2023 12:06:05 -0500 you wrote:

From: Guo Ren <[email protected]>

Patch 1 is the fixup patch should be merged into stable-tree. Patch 2 is a continuous optimization for jump_label patch_text atomicity.

Quoted 6 lines

Changes in V3:
- Correct the typo C.JAL -> C.J (Thx Jessica)
- Fixup compile error when CONFIG_RISCV_ISA_C=n
- Rebase on riscv for-next (20230127)

[...]

Here is the summary with links:
- [-next,V3,1/2] riscv: jump_label: Fixup unaligned arch_static_branch function

    https://git.kernel.org/riscv/c/9ddfc3cd8060

- [-next,V3,2/2] riscv: jump_label: Optimize the code size with compressed instruction (no matching commit)

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html