Skip to contentSkip to conversation

arm64: perf: Mark expected switch fall-through

PATCH In review Anders Roxell 5 participants 8 messages Jul 26, 2019

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/3 arm64: perf: Mark expected switch fall-through

When fall-through warnings was enabled by default, commit d93512ef0f0e ("Makefile: Globally enable fall-through warning"), the following warnings was starting to show up:

../arch/arm64/kernel/hw_breakpoint.c: In function ‘hw_breakpoint_arch_parse’:
../arch/arm64/kernel/hw_breakpoint.c:540:7: warning: this statement may fall
through [-Wimplicit-fallthrough=]

    if (hw->ctrl.len == ARM_BREAKPOINT_LEN_1)
       ^

../arch/arm64/kernel/hw_breakpoint.c:542:3: note: here
case 2:
^~~~
../arch/arm64/kernel/hw_breakpoint.c:544:7: warning: this statement may fall
through [-Wimplicit-fallthrough=]

    if (hw->ctrl.len == ARM_BREAKPOINT_LEN_2)
       ^

../arch/arm64/kernel/hw_breakpoint.c:546:3: note: here
default:
^~~~~~~

Rework so that the compiler doesn't warn about fall-through. Rework so the code looks like the arm code. Since the comment in the function indicates taht this is supposed to behave the same way as arm32 because it handles 32-bit tasks also.

Cc [email protected] # v3.16+
Fixes 6ee33c2712fc ("ARM: hw_breakpoint: correct and simplify alignment fixup code")
---
 arch/arm64/kernel/hw_breakpoint.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)
arch/arm64/kernel/hw_breakpoint.c +7 -4
index dceb84520948..ea616adf1cf1 100644
@@ -535,14 +535,17 @@ int hw_breakpoint_arch_parse(struct perf_event *bp,
 		case 0:
 			/* Aligned */
 			break;
-		case 1:
-			/* Allow single byte watchpoint. */
-			if (hw->ctrl.len == ARM_BREAKPOINT_LEN_1)
-				break;
 		case 2:
 			/* Allow halfword watchpoints and breakpoints. */
 			if (hw->ctrl.len == ARM_BREAKPOINT_LEN_2)
 				break;
+			/* Fall through */
+		case 1:
+		case 3:
+			/* Allow single byte watchpoint. */
+			if (hw->ctrl.len == ARM_BREAKPOINT_LEN_1)
+				break;
+			/* Fall through */
 		default:
 			return -EINVAL;
 		}
Signature · 1 line
2.20.1
maintainer
#1 lore raw
Quoted 20 lines · Anders Roxell

When fall-through warnings was enabled by default, commit d93512ef0f0e ("Makefile: Globally enable fall-through warning"), the following warnings was starting to show up:

../arch/arm64/kernel/hw_breakpoint.c: In function ‘hw_breakpoint_arch_parse’:
../arch/arm64/kernel/hw_breakpoint.c:540:7: warning: this statement may fall
through [-Wimplicit-fallthrough=]

    if (hw->ctrl.len == ARM_BREAKPOINT_LEN_1)
       ^
../arch/arm64/kernel/hw_breakpoint.c:542:3: note: here
   case 2:
   ^~~~
../arch/arm64/kernel/hw_breakpoint.c:544:7: warning: this statement may fall
 through [-Wimplicit-fallthrough=]
    if (hw->ctrl.len == ARM_BREAKPOINT_LEN_2)
       ^
../arch/arm64/kernel/hw_breakpoint.c:546:3: note: here
   default:
   ^~~~~~~
Quoted 3 lines

Rework so that the compiler doesn't warn about fall-through. Rework so the code looks like the arm code. Since the comment in the function indicates taht this is supposed to behave the same way as arm32 because

Typo: s/taht/that/

Quoted 5 lines

it handles 32-bit tasks also.

Cc [email protected] # v3.16+
Fixes 6ee33c2712fc ("ARM: hw_breakpoint: correct and simplify alignment fixup code")

The patch itself looks fine, but I don't think this needs a CC to stable, nor does it require that fixes tag, as there's no functional problem.

If anything, it fixes:

d93512ef0f0e (" Makefile: Globally enable fall-through warning")

... given the commit message for that patch states:

Now that all the fall-through warnings have been addressed in the kernel, enable the fall-through warning globally.

... and the existence of this patch implies otherwise.

IIUC that patch isn't even in mainline yet, but given this is simple I imagine that Will and Catalin might be happy to pick this up for the next rc.

Thanks,
Mark.

Quoted 33 lines
---
 arch/arm64/kernel/hw_breakpoint.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)
arch/arm64/kernel/hw_breakpoint.c +7 -4
index dceb84520948..ea616adf1cf1 100644
@@ -535,14 +535,17 @@ int hw_breakpoint_arch_parse(struct perf_event *bp,
 		case 0:
 			/* Aligned */
 			break;
-		case 1:
-			/* Allow single byte watchpoint. */
-			if (hw->ctrl.len == ARM_BREAKPOINT_LEN_1)
-				break;
 		case 2:
 			/* Allow halfword watchpoints and breakpoints. */
 			if (hw->ctrl.len == ARM_BREAKPOINT_LEN_2)
 				break;
+			/* Fall through */
+		case 1:
+		case 3:
+			/* Allow single byte watchpoint. */
+			if (hw->ctrl.len == ARM_BREAKPOINT_LEN_1)
+				break;
+			/* Fall through */
 		default:
 			return -EINVAL;
 		}
Signature · 1 line
2.20.1
maintainer Acked-by
#2 lore raw
Quoted 29 lines · Mark Rutland
Quoted 20 lines · Anders Roxell

When fall-through warnings was enabled by default, commit d93512ef0f0e ("Makefile: Globally enable fall-through warning"), the following warnings was starting to show up:

../arch/arm64/kernel/hw_breakpoint.c: In function ‘hw_breakpoint_arch_parse’:
../arch/arm64/kernel/hw_breakpoint.c:540:7: warning: this statement may fall
through [-Wimplicit-fallthrough=]

    if (hw->ctrl.len == ARM_BREAKPOINT_LEN_1)
       ^
../arch/arm64/kernel/hw_breakpoint.c:542:3: note: here
   case 2:
   ^~~~
../arch/arm64/kernel/hw_breakpoint.c:544:7: warning: this statement may fall
 through [-Wimplicit-fallthrough=]
    if (hw->ctrl.len == ARM_BREAKPOINT_LEN_2)
       ^
../arch/arm64/kernel/hw_breakpoint.c:546:3: note: here
   default:
   ^~~~~~~
Quoted 3 lines

Rework so that the compiler doesn't warn about fall-through. Rework so the code looks like the arm code. Since the comment in the function indicates taht this is supposed to behave the same way as arm32 because

Typo: s/taht/that/

Quoted 2 lines

it handles 32-bit tasks also.

Quoted 7 lines
Quoted 3 lines
Cc [email protected] # v3.16+
Fixes 6ee33c2712fc ("ARM: hw_breakpoint: correct and simplify alignment fixup code")

The patch itself looks fine, but I don't think this needs a CC to stable, nor does it require that fixes tag, as there's no functional problem.

Hmm... I now see I spoke too soon, and this is making the 1-byte breakpoint work at a 3-byte offset.

Given that:

... and the fixes and stable tags are appropriate for that portion of the patch.

Sorry for the noise.

Thanks,
Mark.

Quoted 33 lines
Quoted 33 lines
---
 arch/arm64/kernel/hw_breakpoint.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)
arch/arm64/kernel/hw_breakpoint.c +7 -4
index dceb84520948..ea616adf1cf1 100644
@@ -535,14 +535,17 @@ int hw_breakpoint_arch_parse(struct perf_event *bp,
 		case 0:
 			/* Aligned */
 			break;
-		case 1:
-			/* Allow single byte watchpoint. */
-			if (hw->ctrl.len == ARM_BREAKPOINT_LEN_1)
-				break;
 		case 2:
 			/* Allow halfword watchpoints and breakpoints. */
 			if (hw->ctrl.len == ARM_BREAKPOINT_LEN_2)
 				break;
+			/* Fall through */
+		case 1:
+		case 3:
+			/* Allow single byte watchpoint. */
+			if (hw->ctrl.len == ARM_BREAKPOINT_LEN_1)
+				break;
+			/* Fall through */
 		default:
 			return -EINVAL;
 		}
Signature · 1 line
2.20.1
maintainer
#3 lore raw
Quoted 34 lines · Mark Rutland
Quoted 33 lines · Mark Rutland
Quoted 20 lines · Anders Roxell

When fall-through warnings was enabled by default, commit d93512ef0f0e ("Makefile: Globally enable fall-through warning"), the following warnings was starting to show up:

../arch/arm64/kernel/hw_breakpoint.c: In function ‘hw_breakpoint_arch_parse’:
../arch/arm64/kernel/hw_breakpoint.c:540:7: warning: this statement may fall
through [-Wimplicit-fallthrough=]

    if (hw->ctrl.len == ARM_BREAKPOINT_LEN_1)
       ^
../arch/arm64/kernel/hw_breakpoint.c:542:3: note: here
   case 2:
   ^~~~
../arch/arm64/kernel/hw_breakpoint.c:544:7: warning: this statement may fall
 through [-Wimplicit-fallthrough=]
    if (hw->ctrl.len == ARM_BREAKPOINT_LEN_2)
       ^
../arch/arm64/kernel/hw_breakpoint.c:546:3: note: here
   default:
   ^~~~~~~
Quoted 3 lines

Rework so that the compiler doesn't warn about fall-through. Rework so the code looks like the arm code. Since the comment in the function indicates taht this is supposed to behave the same way as arm32 because

Typo: s/taht/that/

Quoted 5 lines

it handles 32-bit tasks also.

Cc [email protected] # v3.16+
Fixes 6ee33c2712fc ("ARM: hw_breakpoint: correct and simplify alignment fixup code")
Quoted 6 lines
Quoted 3 lines

The patch itself looks fine, but I don't think this needs a CC to stable, nor does it require that fixes tag, as there's no functional problem.

Hmm... I now see I spoke too soon, and this is making the 1-byte breakpoint work at a 3-byte offset.

I still don't think it's quite right though, since it forbids a 2-byte watchpoint on a byte-aligned address.

I think the arm64 code matches what we had on 32-bit prior to d968d2b801d8 ("ARM: 7497/1: hw_breakpoint: allow single-byte watchpoints on all addresses"), so we should have one patch bringing us up to speed with that change, and then another annotating the fallthroughs.

Will

#4 lore raw
Quoted 39 lines · Will Deacon
Quoted 38 lines · Mark Rutland
Quoted 36 lines · Mark Rutland
Quoted 20 lines · Anders Roxell

When fall-through warnings was enabled by default, commit d93512ef0f0e ("Makefile: Globally enable fall-through warning"), the following warnings was starting to show up:

../arch/arm64/kernel/hw_breakpoint.c: In function ‘hw_breakpoint_arch_parse’:
../arch/arm64/kernel/hw_breakpoint.c:540:7: warning: this statement may fall
through [-Wimplicit-fallthrough=]

     if (hw->ctrl.len == ARM_BREAKPOINT_LEN_1)
        ^
../arch/arm64/kernel/hw_breakpoint.c:542:3: note: here
    case 2:
    ^~~~
../arch/arm64/kernel/hw_breakpoint.c:544:7: warning: this statement may fall
  through [-Wimplicit-fallthrough=]
     if (hw->ctrl.len == ARM_BREAKPOINT_LEN_2)
        ^
../arch/arm64/kernel/hw_breakpoint.c:546:3: note: here
    default:
    ^~~~~~~
Quoted 3 lines

Rework so that the compiler doesn't warn about fall-through. Rework so the code looks like the arm code. Since the comment in the function indicates taht this is supposed to behave the same way as arm32 because

Typo: s/taht/that/

Quoted 5 lines

it handles 32-bit tasks also.

Cc [email protected] # v3.16+
Fixes 6ee33c2712fc ("ARM: hw_breakpoint: correct and simplify alignment fixup code")

The patch itself looks fine, but I don't think this needs a CC to stable, nor does it require that fixes tag, as there's no functional problem.

Quoted 5 lines
Quoted 2 lines

Hmm... I now see I spoke too soon, and this is making the 1-byte breakpoint work at a 3-byte offset.

I still don't think it's quite right though, since it forbids a 2-byte watchpoint on a byte-aligned address.

Plus, AFAICS, a 1-byte watchpoint on a 2-byte-aligned address.

Not that I know anything about this code, but it does start to look like it might want rewriting without the offending switch statement anyway. At a glance, it looks like the intended semantic might boil down to:

	if (hw->ctrl.len > offset)
		return -EINVAL;

Robin.

Quoted 12 lines

I think the arm64 code matches what we had on 32-bit prior to d968d2b801d8 ("ARM: 7497/1: hw_breakpoint: allow single-byte watchpoints on all addresses"), so we should have one patch bringing us up to speed with that change, and then another annotating the fallthroughs.

Will

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

maintainer
#5 lore raw
Quoted 46 lines · Robin Murphy
Quoted 39 lines · Will Deacon
Quoted 38 lines · Mark Rutland
Quoted 36 lines · Mark Rutland
Quoted 20 lines · Anders Roxell

When fall-through warnings was enabled by default, commit d93512ef0f0e ("Makefile: Globally enable fall-through warning"), the following warnings was starting to show up:

../arch/arm64/kernel/hw_breakpoint.c: In function ‘hw_breakpoint_arch_parse’:
../arch/arm64/kernel/hw_breakpoint.c:540:7: warning: this statement may fall
through [-Wimplicit-fallthrough=]

     if (hw->ctrl.len == ARM_BREAKPOINT_LEN_1)
        ^
../arch/arm64/kernel/hw_breakpoint.c:542:3: note: here
    case 2:
    ^~~~
../arch/arm64/kernel/hw_breakpoint.c:544:7: warning: this statement may fall
  through [-Wimplicit-fallthrough=]
     if (hw->ctrl.len == ARM_BREAKPOINT_LEN_2)
        ^
../arch/arm64/kernel/hw_breakpoint.c:546:3: note: here
    default:
    ^~~~~~~
Quoted 3 lines

Rework so that the compiler doesn't warn about fall-through. Rework so the code looks like the arm code. Since the comment in the function indicates taht this is supposed to behave the same way as arm32 because

Typo: s/taht/that/

Quoted 5 lines

it handles 32-bit tasks also.

Cc [email protected] # v3.16+
Fixes 6ee33c2712fc ("ARM: hw_breakpoint: correct and simplify alignment fixup code")

The patch itself looks fine, but I don't think this needs a CC to stable, nor does it require that fixes tag, as there's no functional problem.

Quoted 5 lines
Quoted 2 lines

Hmm... I now see I spoke too soon, and this is making the 1-byte breakpoint work at a 3-byte offset.

I still don't think it's quite right though, since it forbids a 2-byte watchpoint on a byte-aligned address.

Quoted 8 lines

Plus, AFAICS, a 1-byte watchpoint on a 2-byte-aligned address.

Not that I know anything about this code, but it does start to look like it might want rewriting without the offending switch statement anyway. At a glance, it looks like the intended semantic might boil down to:

	if (hw->ctrl.len > offset)
		return -EINVAL;

Given that it's compat code, I think it's worth staying as close to the arch/arm/ implementation as we can. Also, beware that the ARM_BREAKPOINT_LEN_* definitions are masks because of the BAS fields in the debug architecture.

Will

#6 lore raw
Quoted 41 lines · Will Deacon
Quoted 40 lines · Robin Murphy
Quoted 39 lines · Will Deacon
Quoted 38 lines · Mark Rutland
Quoted 36 lines · Mark Rutland
Quoted 20 lines · Anders Roxell

When fall-through warnings was enabled by default, commit d93512ef0f0e ("Makefile: Globally enable fall-through warning"), the following warnings was starting to show up:

../arch/arm64/kernel/hw_breakpoint.c: In function ‘hw_breakpoint_arch_parse’:
../arch/arm64/kernel/hw_breakpoint.c:540:7: warning: this statement may fall
through [-Wimplicit-fallthrough=]

      if (hw->ctrl.len == ARM_BREAKPOINT_LEN_1)
         ^
../arch/arm64/kernel/hw_breakpoint.c:542:3: note: here
     case 2:
     ^~~~
../arch/arm64/kernel/hw_breakpoint.c:544:7: warning: this statement may fall
   through [-Wimplicit-fallthrough=]
      if (hw->ctrl.len == ARM_BREAKPOINT_LEN_2)
         ^
../arch/arm64/kernel/hw_breakpoint.c:546:3: note: here
     default:
     ^~~~~~~
Quoted 3 lines

Rework so that the compiler doesn't warn about fall-through. Rework so the code looks like the arm code. Since the comment in the function indicates taht this is supposed to behave the same way as arm32 because

Typo: s/taht/that/

Quoted 5 lines

it handles 32-bit tasks also.

Cc [email protected] # v3.16+
Fixes 6ee33c2712fc ("ARM: hw_breakpoint: correct and simplify alignment fixup code")

The patch itself looks fine, but I don't think this needs a CC to stable, nor does it require that fixes tag, as there's no functional problem.

Quoted 7 lines
Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

Hmm... I now see I spoke too soon, and this is making the 1-byte breakpoint work at a 3-byte offset.

I still don't think it's quite right though, since it forbids a 2-byte watchpoint on a byte-aligned address.

Plus, AFAICS, a 1-byte watchpoint on a 2-byte-aligned address.

[and of course, what I missed was that that's the case the fallthrough serves... yuck indeed]

Quoted 9 lines
Quoted 6 lines

Not that I know anything about this code, but it does start to look like it might want rewriting without the offending switch statement anyway. At a glance, it looks like the intended semantic might boil down to:

	if (hw->ctrl.len > offset)
		return -EINVAL;

Given that it's compat code, I think it's worth staying as close to the arch/arm/ implementation as we can.

Right, I also misread the accompanying arch/arm/ patch and got the impression that 32-bit also had a problem such that any fix would happen in parallel - on closer inspection the current arch/arm/ code does actually seem to make sense, even if it is horribly subtle.

Quoted 3 lines

Also, beware that the
ARM_BREAKPOINT_LEN_* definitions are masks because of the BAS fields in the debug architecture.

Fun... Clearly it's a bit too Friday for me to be useful here, so apologies for the confusion :)

Robin.

maintainer
#7 lore raw
In reply to Anders Roxell

Hi,

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag,
fixing commit: 6ee33c2712fc ARM: hw_breakpoint: correct and simplify alignment fixup code.

The bot has tested the following trees: v5.2.2, v5.1.19, v4.19.60, v4.14.134, v4.9.186, v4.4.186.

v5.2.2: Build OK!
v5.1.19: Build OK!
v4.19.60: Build OK!
v4.14.134: Failed to apply! Possible dependencies:
18ff57b22061 ("hw_breakpoint: Factor out __modify_user_hw_breakpoint() function") 195bce73bd10 ("signal/sh: Use force_sig_fault in hw_breakpoint_handler") 705feaf321c3 ("hw_breakpoint: Add perf_event_attr fields check in __modify_user_hw_breakpoint()") 8c449753a681 ("perf/arch/arm64: Implement hw_breakpoint_arch_parse()") 8e983ff9ac02 ("perf/hw_breakpoint: Pass arch breakpoint struct to arch_check_bp_in_kernelspace()") 9a4903dde2c8 ("perf/hw_breakpoint: Split attribute parse and commit") ea6a9d530c17 ("hw_breakpoint: Add modify_bp_slot() function")

v4.9.186: Failed to apply! Possible dependencies:
0ddb8e0b784b ("arm64: Allow hw watchpoint of length 3,5,6 and 7")
18ff57b22061 ("hw_breakpoint: Factor out __modify_user_hw_breakpoint() function") 195bce73bd10 ("signal/sh: Use force_sig_fault in hw_breakpoint_handler") 705feaf321c3 ("hw_breakpoint: Add perf_event_attr fields check in __modify_user_hw_breakpoint()") 8c449753a681 ("perf/arch/arm64: Implement hw_breakpoint_arch_parse()") 8e983ff9ac02 ("perf/hw_breakpoint: Pass arch breakpoint struct to arch_check_bp_in_kernelspace()") 9a4903dde2c8 ("perf/hw_breakpoint: Split attribute parse and commit") b08fb180bb88 ("arm64: Allow hw watchpoint at varied offset from base address") ea6a9d530c17 ("hw_breakpoint: Add modify_bp_slot() function")

v4.4.186: Failed to apply! Possible dependencies:
0ddb8e0b784b ("arm64: Allow hw watchpoint of length 3,5,6 and 7")
18ff57b22061 ("hw_breakpoint: Factor out __modify_user_hw_breakpoint() function") 195bce73bd10 ("signal/sh: Use force_sig_fault in hw_breakpoint_handler") 6ec7026ac01f ("xtensa: use context structure for debug exceptions")
705feaf321c3 ("hw_breakpoint: Add perf_event_attr fields check in __modify_user_hw_breakpoint()") 8c449753a681 ("perf/arch/arm64: Implement hw_breakpoint_arch_parse()") 8e983ff9ac02 ("perf/hw_breakpoint: Pass arch breakpoint struct to arch_check_bp_in_kernelspace()") 9a4903dde2c8 ("perf/hw_breakpoint: Split attribute parse and commit") b08fb180bb88 ("arm64: Allow hw watchpoint at varied offset from base address") c91e02bd9702 ("xtensa: support hardware breakpoints/watchpoints")
ea6a9d530c17 ("hw_breakpoint: Add modify_bp_slot() function")

NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

--
Thanks,
Sasha

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