Skip to contentSkip to conversation

RISC-V: fix compile error from deduplicated __ALTERNATIVE_CFG_2

PATCH In review Heiko Stuebner 3 participants 5 messages Jan 5, 2023 – Jan 6, 2023 RISC-V

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 RISC-V: fix compile error from deduplicated __ALTERNATIVE_CFG_2

From: Heiko Stuebner <[email protected]>

On the non-assembler-side wrapping alternative-macros inside other macros to prevent duplication of code works, as the end result will just be a string that gets fed to the asm instruction.

In real assembler code, wrapping .macro blocks inside other .macro blocks brings more restrictions on usage it seems and the optimization done by commit 2ba8c7dc71c0 ("riscv: Don't duplicate __ALTERNATIVE_CFG in __ALTERNATIVE_CFG_2") results in a compile error like:

../arch/riscv/lib/strcmp.S: Assembler messages:
../arch/riscv/lib/strcmp.S:15: Error: too many positional arguments ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "886:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "887:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "886:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "887:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "886:" ../arch/riscv/lib/strcmp.S:15: Error: attempt to move .org backwards

Wrapping the variables containing assembler code in quotes solves this issue, compilation and the code in question still works and objdump also shows sane decompiled results of the affected code.

Fixes 2ba8c7dc71c0 ("riscv: Don't duplicate __ALTERNATIVE_CFG in __ALTERNATIVE_CFG_2")

---
changes in v2:
- don't revert the affected cleanup but use quotes around the parts

Tested on qemu + Allwinner D1 + specially created test-cases using ALTERNATIVE_2 in full assembler .S files

 arch/riscv/include/asm/alternative-macros.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/riscv/include/asm/alternative-macros.h +1 -1
index 7226e2462584..2c0f4c887289 100644
@@ -46,7 +46,7 @@
 
 .macro ALTERNATIVE_CFG_2 old_c, new_c_1, vendor_id_1, errata_id_1, enable_1,	\
 				new_c_2, vendor_id_2, errata_id_2, enable_2
-	ALTERNATIVE_CFG \old_c, \new_c_1, \vendor_id_1, \errata_id_1, \enable_1
+	ALTERNATIVE_CFG "\old_c", "\new_c_1", \vendor_id_1, \errata_id_1, \enable_1
 	ALT_NEW_CONTENT \vendor_id_2, \errata_id_2, \enable_2, \new_c_2
 .endm
 
Signature · 7 lines
2.35.1


_______________________________________________
linux-riscv mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-riscv
maintainer Reviewed-by on v2 1/1
#1 lore raw
Quoted 41 lines · On Thu, 05 Jan 2023 11:26:10 PST (-0800), [email protected] wrote:

From: Heiko Stuebner <[email protected]>

On the non-assembler-side wrapping alternative-macros inside other macros to prevent duplication of code works, as the end result will just be a string that gets fed to the asm instruction.

In real assembler code, wrapping .macro blocks inside other .macro blocks brings more restrictions on usage it seems and the optimization done by commit 2ba8c7dc71c0 ("riscv: Don't duplicate __ALTERNATIVE_CFG in __ALTERNATIVE_CFG_2") results in a compile error like:

../arch/riscv/lib/strcmp.S: Assembler messages:
../arch/riscv/lib/strcmp.S:15: Error: too many positional arguments ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "886:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "887:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "886:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "887:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "886:" ../arch/riscv/lib/strcmp.S:15: Error: attempt to move .org backwards

Wrapping the variables containing assembler code in quotes solves this issue, compilation and the code in question still works and objdump also shows sane decompiled results of the affected code.

Fixes 2ba8c7dc71c0 ("riscv: Don't duplicate __ALTERNATIVE_CFG in __ALTERNATIVE_CFG_2")

---
changes in v2:
- don't revert the affected cleanup but use quotes around the parts

Tested on qemu + Allwinner D1 + specially created test-cases using ALTERNATIVE_2 in full assembler .S files

 arch/riscv/include/asm/alternative-macros.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/riscv/include/asm/alternative-macros.h +0 -0
index 7226e2462584..2c0f4c887289 100644

arch/riscv/include/asm/alternative-macros.h line 51

Quoted 7 lines
arch/riscv/include/asm/alternative-macros.h +1 -1
@@ -46,7 +46,7 @@
 
 .macro ALTERNATIVE_CFG_2 old_c, new_c_1, vendor_id_1, errata_id_1, enable_1,	\
 				new_c_2, vendor_id_2, errata_id_2, enable_2
-	ALTERNATIVE_CFG \old_c, \new_c_1, \vendor_id_1, \errata_id_1, \enable_1
+	ALTERNATIVE_CFG "\old_c", "\new_c_1", \vendor_id_1, \errata_id_1, \enable_1
 	ALT_NEW_CONTENT \vendor_id_2, \errata_id_2, \enable_2, \new_c_2
 .endm

I'm going to hold off on this one for this week, though -- at least to give it some time on the lists, but it appears this hasn't landed at lore/patchwork yet and thus might be stuck. It is in the mailman archives...

http://lists.infradead.org/pipermail/linux-riscv/2023-January/024682.html

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

#2 lore raw

Am Donnerstag, 5. Januar 2023, 21:47:47 CET schrieb Palmer Dabbelt:

Quoted 52 lines · On Thu, 05 Jan 2023 11:26:10 PST (-0800), [email protected] wrote:
Quoted 41 lines

From: Heiko Stuebner <[email protected]>

On the non-assembler-side wrapping alternative-macros inside other macros to prevent duplication of code works, as the end result will just be a string that gets fed to the asm instruction.

In real assembler code, wrapping .macro blocks inside other .macro blocks brings more restrictions on usage it seems and the optimization done by commit 2ba8c7dc71c0 ("riscv: Don't duplicate __ALTERNATIVE_CFG in __ALTERNATIVE_CFG_2") results in a compile error like:

../arch/riscv/lib/strcmp.S: Assembler messages:
../arch/riscv/lib/strcmp.S:15: Error: too many positional arguments ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "886:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "887:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "886:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "887:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "886:" ../arch/riscv/lib/strcmp.S:15: Error: attempt to move .org backwards

Wrapping the variables containing assembler code in quotes solves this issue, compilation and the code in question still works and objdump also shows sane decompiled results of the affected code.

Fixes 2ba8c7dc71c0 ("riscv: Don't duplicate __ALTERNATIVE_CFG in __ALTERNATIVE_CFG_2")

---
changes in v2:
- don't revert the affected cleanup but use quotes around the parts

Tested on qemu + Allwinner D1 + specially created test-cases using ALTERNATIVE_2 in full assembler .S files

 arch/riscv/include/asm/alternative-macros.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/riscv/include/asm/alternative-macros.h +0 -0
index 7226e2462584..2c0f4c887289 100644

arch/riscv/include/asm/alternative-macros.h line 51

Quoted 7 lines
arch/riscv/include/asm/alternative-macros.h +1 -1
@@ -46,7 +46,7 @@
 
 .macro ALTERNATIVE_CFG_2 old_c, new_c_1, vendor_id_1, errata_id_1, enable_1,	\
 				new_c_2, vendor_id_2, errata_id_2, enable_2
-	ALTERNATIVE_CFG \old_c, \new_c_1, \vendor_id_1, \errata_id_1, \enable_1
+	ALTERNATIVE_CFG "\old_c", "\new_c_1", \vendor_id_1, \errata_id_1, \enable_1
 	ALT_NEW_CONTENT \vendor_id_2, \errata_id_2, \enable_2, \new_c_2
 .endm
Quoted 4 lines

I'm going to hold off on this one for this week, though -- at least to give it some time on the lists, but it appears this hasn't landed at lore/patchwork yet and thus might be stuck. It is in the mailman archives...

No worries - there isn't any real-word breakage happening with 6.2-rc So we're in no particular rush to fix somebodies broken boardfarm :-)

Heiko

Quoted 2 lines

http://lists.infradead.org/pipermail/linux-riscv/2023-January/024682.html

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

Reviewed-by on v2 1/1
#3 lore raw
In reply to Heiko Stuebner
Quoted 52 lines · Heiko Stuebner

From: Heiko Stuebner <[email protected]>

On the non-assembler-side wrapping alternative-macros inside other macros to prevent duplication of code works, as the end result will just be a string that gets fed to the asm instruction.

In real assembler code, wrapping .macro blocks inside other .macro blocks brings more restrictions on usage it seems and the optimization done by commit 2ba8c7dc71c0 ("riscv: Don't duplicate __ALTERNATIVE_CFG in __ALTERNATIVE_CFG_2") results in a compile error like:

../arch/riscv/lib/strcmp.S: Assembler messages:
../arch/riscv/lib/strcmp.S:15: Error: too many positional arguments ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "886:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "887:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "886:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "887:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "886:" ../arch/riscv/lib/strcmp.S:15: Error: attempt to move .org backwards

Wrapping the variables containing assembler code in quotes solves this issue, compilation and the code in question still works and objdump also shows sane decompiled results of the affected code.

Fixes 2ba8c7dc71c0 ("riscv: Don't duplicate __ALTERNATIVE_CFG in __ALTERNATIVE_CFG_2")

---
changes in v2:
- don't revert the affected cleanup but use quotes around the parts

Tested on qemu + Allwinner D1 + specially created test-cases using ALTERNATIVE_2 in full assembler .S files

 arch/riscv/include/asm/alternative-macros.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/riscv/include/asm/alternative-macros.h +1 -1
index 7226e2462584..2c0f4c887289 100644
@@ -46,7 +46,7 @@
 
 .macro ALTERNATIVE_CFG_2 old_c, new_c_1, vendor_id_1, errata_id_1, enable_1,	\
 				new_c_2, vendor_id_2, errata_id_2, enable_2
-	ALTERNATIVE_CFG \old_c, \new_c_1, \vendor_id_1, \errata_id_1, \enable_1
+	ALTERNATIVE_CFG "\old_c", "\new_c_1", \vendor_id_1, \errata_id_1, \enable_1
 	ALT_NEW_CONTENT \vendor_id_2, \errata_id_2, \enable_2, \new_c_2
 .endm
 
Signature · 1 line
2.35.1

I was sort of hoping we'd change the vararg "hack" at the same time, with a separate patch, but I guess we can focus on getting this breakage fixed first and worry about the vararg later.

Thanks,
drew

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

#4 lore raw

Am Freitag, 6. Januar 2023, 07:54:49 CET schrieb Andrew Jones:

Quoted 43 lines · Heiko Stuebner
Quoted 42 lines

From: Heiko Stuebner <[email protected]>

On the non-assembler-side wrapping alternative-macros inside other macros to prevent duplication of code works, as the end result will just be a string that gets fed to the asm instruction.

In real assembler code, wrapping .macro blocks inside other .macro blocks brings more restrictions on usage it seems and the optimization done by commit 2ba8c7dc71c0 ("riscv: Don't duplicate __ALTERNATIVE_CFG in __ALTERNATIVE_CFG_2") results in a compile error like:

../arch/riscv/lib/strcmp.S: Assembler messages:
../arch/riscv/lib/strcmp.S:15: Error: too many positional arguments ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "886:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "887:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "886:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "887:" ../arch/riscv/lib/strcmp.S:15: Error: backward ref to unknown label "886:" ../arch/riscv/lib/strcmp.S:15: Error: attempt to move .org backwards

Wrapping the variables containing assembler code in quotes solves this issue, compilation and the code in question still works and objdump also shows sane decompiled results of the affected code.

Fixes 2ba8c7dc71c0 ("riscv: Don't duplicate __ALTERNATIVE_CFG in __ALTERNATIVE_CFG_2")

---
changes in v2:
- don't revert the affected cleanup but use quotes around the parts

Tested on qemu + Allwinner D1 + specially created test-cases using ALTERNATIVE_2 in full assembler .S files

 arch/riscv/include/asm/alternative-macros.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/riscv/include/asm/alternative-macros.h +0 -0
index 7226e2462584..2c0f4c887289 100644
@@ -46,7 +46,7 @@
 
Quoted 12 lines

arch/riscv/include/asm/alternative-macros.h line 52

Quoted 8 lines
arch/riscv/include/asm/alternative-macros.h +1 -1
 .macro ALTERNATIVE_CFG_2 old_c, new_c_1, vendor_id_1, errata_id_1, enable_1,	\
 				new_c_2, vendor_id_2, errata_id_2, enable_2
-	ALTERNATIVE_CFG \old_c, \new_c_1, \vendor_id_1, \errata_id_1, \enable_1
+	ALTERNATIVE_CFG "\old_c", "\new_c_1", \vendor_id_1, \errata_id_1, \enable_1
 	ALT_NEW_CONTENT \vendor_id_2, \errata_id_2, \enable_2, \new_c_2
 .endm
 

I was sort of hoping we'd change the vararg "hack" at the same time, with a separate patch, but I guess we can focus on getting this breakage fixed first and worry about the vararg later.

I was debating with myself about that :-) . But then did go with the minimal variant, as I expect this to go into 6.2-rc and thus not wanting to introduce too much changes here.

Quoted 1 line

thanks :-)

Heiko

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