[PATCH 0/2] ARC fixes/updates

[PATCH 0/2] ARC fixes/updates

[Vineet Gupta via Libc-alpha]

Hi,

A couple of updates for ARC.

Thx,
-Vineet

Vineet Gupta (2):
  ARC: fp: (micro)optimize FPU_STATUS read by eliding FWE bit clearing
  ARC: elf: make type safe

 sysdeps/arc/dl-machine.h  | 23 ++++++++++-------------
 sysdeps/arc/fpu_control.h | 15 +++++++--------
 2 files changed, 17 insertions(+), 21 deletions(-)

-- 
2.25.1

[PATCH 1/2] ARC: fp: (micro)optimize FPU_STATUS read by eliding FWE bit clearing

[Vineet Gupta via Libc-alpha]

Any FPU_STATUS write needs setting the FWE bit (31) whcih just provides
a "control signal" to enable explicit write (vs. the side-effect of FPU
instructions).  However this bit is RAZ and write-only, thus effectively
never stored in FPU_STATUS register. Thus when reading the register
there is no need to clear it. This shaves off a BCLR instruction from
the fe*exceptino family of functions and while no big deal still makes
sense to do.

This came up when debugging a race in math/test-fenv-tls [1]

[1]: https://github.com/foss-for-synopsys-dwc-arc-processors/linux/issues/54

Signed-off-by: Vineet Gupta <vgupta@synopsys.com>
---
 sysdeps/arc/fpu_control.h | 15 +++++++--------
 1 file changed, 7 insertions(+), 8 deletions(-)

diff --git a/sysdeps/arc/fpu_control.h b/sysdeps/arc/fpu_control.h
index c7d101e7838f..ae4348321c16 100644
--- a/sysdeps/arc/fpu_control.h
+++ b/sysdeps/arc/fpu_control.h
@@ -81,21 +81,20 @@ typedef unsigned int fpu_control_t;
 #  define _FPU_SETCW(cw) __asm__ volatile ("sr %0, [0x300]" : : "r" (cw))
 
 /*  Macros for accessing the hardware status word.
-    FWE bit is special as it controls if actual status bits could be wrritten
-    explicitly (other than FPU instructions). We handle it here to keep the
-    callers agnostic of it:
-      - clear it out when reporting status bits
-      - always set it when changing status bits.  */
+    Writing to FPU_STATUS requires a "control" bit FWE to be able to set the
+    exception flags directly (as opposed to side-effects of FP instructions).
+    That is done in the macro here to keeps callers agnostic of this detail.
+    And given FWE is write-only and RAZ, no need to "clear" it in _FPU_GETS
+    macro.  */
 #  define _FPU_GETS(cw)				\
     __asm__ volatile ("lr   %0, [0x301]	\r\n" 	\
-                      "bclr %0, %0, 31	\r\n" 	\
                       : "=r" (cw))
 
 #  define _FPU_SETS(cw)				\
     do {					\
-      unsigned int __tmp = 0x80000000 | (cw);	\
+      unsigned int __fwe = 0x80000000 | (cw);	\
       __asm__ volatile ("sr  %0, [0x301] \r\n" 	\
-                        : : "r" (__tmp));	\
+                        : : "r" (__fwe));	\
     } while (0)
 
 /* Default control word set at startup.  */
-- 
2.25.1

[PATCH 2/2] ARC: elf: make type safe

[Vineet Gupta via Libc-alpha]

Signed-off-by: Vineet Gupta <vgupta@synopsys.com>
---
 sysdeps/arc/dl-machine.h | 23 ++++++++++-------------
 1 file changed, 10 insertions(+), 13 deletions(-)

diff --git a/sysdeps/arc/dl-machine.h b/sysdeps/arc/dl-machine.h
index 60b49af0d5ca..e6ce7f0ff6d9 100644
--- a/sysdeps/arc/dl-machine.h
+++ b/sysdeps/arc/dl-machine.h
@@ -86,7 +86,7 @@
 
 /* Return nonzero iff ELF header is compatible with the running host.  */
 static inline int
-elf_machine_matches_host (const Elf32_Ehdr *ehdr)
+elf_machine_matches_host (const ElfW(Ehdr) *ehdr)
 {
   return (ehdr->e_machine == EM_ARCV2		 /* ARC HS.  */
 	  || ehdr->e_machine == EM_ARC_COMPACT); /* ARC 700.  */
@@ -124,7 +124,7 @@ static inline int
 __attribute__ ((always_inline))
 elf_machine_runtime_setup (struct link_map *l, int lazy, int profile)
 {
-  extern void _dl_runtime_resolve (Elf32_Word);
+  extern void _dl_runtime_resolve (void);
 
   if (l->l_info[DT_JMPREL] && lazy)
     {
@@ -202,7 +202,7 @@ __start:								\n\
 /* A reloc type used for ld.so cmdline arg lookups to reject PLT entries.  */
 #define ELF_MACHINE_JMP_SLOT  R_ARC_JUMP_SLOT
 
-/* ARC uses Elf32_Rela relocations.  */
+/* ARC uses Rela relocations.  */
 #define ELF_MACHINE_NO_REL 1
 #define ELF_MACHINE_NO_RELA 0
 
@@ -211,19 +211,14 @@ __start:								\n\
 static inline ElfW(Addr)
 elf_machine_fixup_plt (struct link_map *map, lookup_t t,
 		       const ElfW(Sym) *refsym, const ElfW(Sym) *sym,
-		       const Elf32_Rela *reloc,
+		       const ElfW(Rela) *reloc,
 		       ElfW(Addr) *reloc_addr, ElfW(Addr) value)
 {
   return *reloc_addr = value;
 }
 
 /* Return the final value of a plt relocation.  */
-static inline ElfW(Addr)
-elf_machine_plt_value (struct link_map *map, const Elf32_Rela *reloc,
-                       ElfW(Addr) value)
-{
-  return value;
-}
+#define elf_machine_plt_value(map, reloc, value) (value)
 
 /* Names of the architecture-specific auditing callback functions.  */
 #define ARCH_LA_PLTENTER arc_gnu_pltenter
@@ -239,8 +234,9 @@ elf_machine_rela (struct link_map *map, const ElfW(Rela) *reloc,
                   const ElfW(Sym) *sym, const struct r_found_version *version,
                   void *const reloc_addr_arg, int skip_ifunc)
 {
+  ElfW(Addr) r_info = reloc->r_info;
+  const unsigned long int r_type = ELFW (R_TYPE) (r_info);
   ElfW(Addr) *const reloc_addr = reloc_addr_arg;
-  const unsigned int r_type = ELF32_R_TYPE (reloc->r_info);
 
   if (__glibc_unlikely (r_type == R_ARC_RELATIVE))
     *reloc_addr += map->l_addr;
@@ -334,11 +330,12 @@ elf_machine_lazy_rel (struct link_map *map, ElfW(Addr) l_addr,
                       const ElfW(Rela) *reloc, int skip_ifunc)
 {
   ElfW(Addr) *const reloc_addr = (void *) (l_addr + reloc->r_offset);
+  const unsigned int r_type = ELFW (R_TYPE) (reloc->r_info);
 
-  if (ELF32_R_TYPE (reloc->r_info) == R_ARC_JUMP_SLOT)
+  if (r_type == R_ARC_JUMP_SLOT)
     *reloc_addr += l_addr;
   else
-    _dl_reloc_bad_type (map, ELF32_R_TYPE (reloc->r_info), 1);
+    _dl_reloc_bad_type (map, r_type, 1);
 }
 
 #endif /* RESOLVE_MAP */
-- 
2.25.1

Re: [PATCH 1/2] ARC: fp: (micro)optimize FPU_STATUS read by eliding FWE bit clearing

[Vineet Gupta via Libc-alpha]

On 7/20/21 1:57 PM, Vineet Gupta via Libc-alpha wrote:
> Any FPU_STATUS write needs setting the FWE bit (31) whcih just provides
> a "control signal" to enable explicit write (vs. the side-effect of FPU
> instructions).  However this bit is RAZ and write-only, thus effectively
> never stored in FPU_STATUS register. Thus when reading the register
> there is no need to clear it. This shaves off a BCLR instruction from
> the fe*exceptino family of functions and while no big deal still makes
> sense to do.
> 
> This came up when debugging a race in math/test-fenv-tls [1]
> 
> [1]: https://github.com/foss-for-synopsys-dwc-arc-processors/linux/issues/54
> 
> Signed-off-by: Vineet Gupta <vgupta@synopsys.com>

Committed !

> ---
>   sysdeps/arc/fpu_control.h | 15 +++++++--------
>   1 file changed, 7 insertions(+), 8 deletions(-)
> 
> diff --git a/sysdeps/arc/fpu_control.h b/sysdeps/arc/fpu_control.h
> index c7d101e7838f..ae4348321c16 100644
> --- a/sysdeps/arc/fpu_control.h
> +++ b/sysdeps/arc/fpu_control.h
> @@ -81,21 +81,20 @@ typedef unsigned int fpu_control_t;
>   #  define _FPU_SETCW(cw) __asm__ volatile ("sr %0, [0x300]" : : "r" (cw))
>   
>   /*  Macros for accessing the hardware status word.
> -    FWE bit is special as it controls if actual status bits could be wrritten
> -    explicitly (other than FPU instructions). We handle it here to keep the
> -    callers agnostic of it:
> -      - clear it out when reporting status bits
> -      - always set it when changing status bits.  */
> +    Writing to FPU_STATUS requires a "control" bit FWE to be able to set the
> +    exception flags directly (as opposed to side-effects of FP instructions).
> +    That is done in the macro here to keeps callers agnostic of this detail.
> +    And given FWE is write-only and RAZ, no need to "clear" it in _FPU_GETS
> +    macro.  */
>   #  define _FPU_GETS(cw)				\
>       __asm__ volatile ("lr   %0, [0x301]	\r\n" 	\
> -                      "bclr %0, %0, 31	\r\n" 	\
>                         : "=r" (cw))
>   
>   #  define _FPU_SETS(cw)				\
>       do {					\
> -      unsigned int __tmp = 0x80000000 | (cw);	\
> +      unsigned int __fwe = 0x80000000 | (cw);	\
>         __asm__ volatile ("sr  %0, [0x301] \r\n" 	\
> -                        : : "r" (__tmp));	\
> +                        : : "r" (__fwe));	\
>       } while (0)
>   
>   /* Default control word set at startup.  */
> 

Re: [PATCH 2/2] ARC: elf: make type safe

[Vineet Gupta via Libc-alpha]

On 7/20/21 1:58 PM, Vineet Gupta via Libc-alpha wrote:
> Signed-off-by: Vineet Gupta <vgupta@synopsys.com>

Committed !

> ---
>   sysdeps/arc/dl-machine.h | 23 ++++++++++-------------
>   1 file changed, 10 insertions(+), 13 deletions(-)
> 
> diff --git a/sysdeps/arc/dl-machine.h b/sysdeps/arc/dl-machine.h
> index 60b49af0d5ca..e6ce7f0ff6d9 100644
> --- a/sysdeps/arc/dl-machine.h
> +++ b/sysdeps/arc/dl-machine.h
> @@ -86,7 +86,7 @@
>   
>   /* Return nonzero iff ELF header is compatible with the running host.  */
>   static inline int
> -elf_machine_matches_host (const Elf32_Ehdr *ehdr)
> +elf_machine_matches_host (const ElfW(Ehdr) *ehdr)
>   {
>     return (ehdr->e_machine == EM_ARCV2		 /* ARC HS.  */
>   	  || ehdr->e_machine == EM_ARC_COMPACT); /* ARC 700.  */
> @@ -124,7 +124,7 @@ static inline int
>   __attribute__ ((always_inline))
>   elf_machine_runtime_setup (struct link_map *l, int lazy, int profile)
>   {
> -  extern void _dl_runtime_resolve (Elf32_Word);
> +  extern void _dl_runtime_resolve (void);
>   
>     if (l->l_info[DT_JMPREL] && lazy)
>       {
> @@ -202,7 +202,7 @@ __start:								\n\
>   /* A reloc type used for ld.so cmdline arg lookups to reject PLT entries.  */
>   #define ELF_MACHINE_JMP_SLOT  R_ARC_JUMP_SLOT
>   
> -/* ARC uses Elf32_Rela relocations.  */
> +/* ARC uses Rela relocations.  */
>   #define ELF_MACHINE_NO_REL 1
>   #define ELF_MACHINE_NO_RELA 0
>   
> @@ -211,19 +211,14 @@ __start:								\n\
>   static inline ElfW(Addr)
>   elf_machine_fixup_plt (struct link_map *map, lookup_t t,
>   		       const ElfW(Sym) *refsym, const ElfW(Sym) *sym,
> -		       const Elf32_Rela *reloc,
> +		       const ElfW(Rela) *reloc,
>   		       ElfW(Addr) *reloc_addr, ElfW(Addr) value)
>   {
>     return *reloc_addr = value;
>   }
>   
>   /* Return the final value of a plt relocation.  */
> -static inline ElfW(Addr)
> -elf_machine_plt_value (struct link_map *map, const Elf32_Rela *reloc,
> -                       ElfW(Addr) value)
> -{
> -  return value;
> -}
> +#define elf_machine_plt_value(map, reloc, value) (value)
>   
>   /* Names of the architecture-specific auditing callback functions.  */
>   #define ARCH_LA_PLTENTER arc_gnu_pltenter
> @@ -239,8 +234,9 @@ elf_machine_rela (struct link_map *map, const ElfW(Rela) *reloc,
>                     const ElfW(Sym) *sym, const struct r_found_version *version,
>                     void *const reloc_addr_arg, int skip_ifunc)
>   {
> +  ElfW(Addr) r_info = reloc->r_info;
> +  const unsigned long int r_type = ELFW (R_TYPE) (r_info);
>     ElfW(Addr) *const reloc_addr = reloc_addr_arg;
> -  const unsigned int r_type = ELF32_R_TYPE (reloc->r_info);
>   
>     if (__glibc_unlikely (r_type == R_ARC_RELATIVE))
>       *reloc_addr += map->l_addr;
> @@ -334,11 +330,12 @@ elf_machine_lazy_rel (struct link_map *map, ElfW(Addr) l_addr,
>                         const ElfW(Rela) *reloc, int skip_ifunc)
>   {
>     ElfW(Addr) *const reloc_addr = (void *) (l_addr + reloc->r_offset);
> +  const unsigned int r_type = ELFW (R_TYPE) (reloc->r_info);
>   
> -  if (ELF32_R_TYPE (reloc->r_info) == R_ARC_JUMP_SLOT)
> +  if (r_type == R_ARC_JUMP_SLOT)
>       *reloc_addr += l_addr;
>     else
> -    _dl_reloc_bad_type (map, ELF32_R_TYPE (reloc->r_info), 1);
> +    _dl_reloc_bad_type (map, r_type, 1);
>   }
>   
>   #endif /* RESOLVE_MAP */
>