| Message ID | 1437401552-30993-1-git-send-email-abrodkin@synopsys.com |
|---|---|
| State | New |
| Headers | show |
On Monday 20 July 2015 07:42 PM, Alexey Brodkin wrote: > From: Claudiu Zissulescu <claziss@synopsys.com> > > Existing version of memset() relies on existence of 64-bit load/stores. > While ARC HS38 may not have those instructions implemented in SoC. > > Proposed implementation checks if "-mno-ll64" option was passed to gcc > (for ARCv2 "-mll64" is set implicitly by default) by checking __LL64__ > definition and if it is not defined uses 32-bit load/stores. > > Signed-off-by: Claudiu Zissulescu <claziss@synopsys.com> > Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com> > Cc: Vineet Gupta <vgupta@synopsys.com> Ping ! Could this patch be applied please ! Thx, -Vineet > --- > libc/string/arc/arcv2/memset.S | 30 ++++++++++++++++++++++++++++++ > 1 file changed, 30 insertions(+) > > diff --git a/libc/string/arc/arcv2/memset.S b/libc/string/arc/arcv2/memset.S > index d076ad1..0918d37 100644 > --- a/libc/string/arc/arcv2/memset.S > +++ b/libc/string/arc/arcv2/memset.S > @@ -52,6 +52,7 @@ ENTRY(memset) > lpnz @.Lset64bytes > ;; LOOP START > PREWRITE(r3, 64) ;Prefetch the next write location > +#ifdef __LL64__ > std.ab r4, [r3, 8] > std.ab r4, [r3, 8] > std.ab r4, [r3, 8] > @@ -60,16 +61,45 @@ ENTRY(memset) > std.ab r4, [r3, 8] > std.ab r4, [r3, 8] > std.ab r4, [r3, 8] > +#else > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > +#endif > .Lset64bytes: > > lsr.f lp_count, r2, 5 ;Last remaining max 124 bytes > lpnz .Lset32bytes > ;; LOOP START > prefetchw [r3, 32] ;Prefetch the next write location > +#ifdef __LL64__ > std.ab r4, [r3, 8] > std.ab r4, [r3, 8] > std.ab r4, [r3, 8] > std.ab r4, [r3, 8] > +#else > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > + st.ab r4, [r3, 4] > +#endif > .Lset32bytes: > > and.f lp_count, r2, 0x1F ;Last remaining 31 bytes >
On Tuesday 20 October 2015 02:33 PM, Vineet Gupta wrote: > On Monday 20 July 2015 07:42 PM, Alexey Brodkin wrote: >> From: Claudiu Zissulescu <claziss@synopsys.com> >> >> Existing version of memset() relies on existence of 64-bit load/stores. >> While ARC HS38 may not have those instructions implemented in SoC. >> >> Proposed implementation checks if "-mno-ll64" option was passed to gcc >> (for ARCv2 "-mll64" is set implicitly by default) by checking __LL64__ >> definition and if it is not defined uses 32-bit load/stores. >> >> Signed-off-by: Claudiu Zissulescu <claziss@synopsys.com> >> Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com> >> Cc: Vineet Gupta <vgupta@synopsys.com> > > > Ping ! Could this patch be applied please ! Ping ! > > Thx, > -Vineet > >> --- >> libc/string/arc/arcv2/memset.S | 30 ++++++++++++++++++++++++++++++ >> 1 file changed, 30 insertions(+) >> >> diff --git a/libc/string/arc/arcv2/memset.S b/libc/string/arc/arcv2/memset.S >> index d076ad1..0918d37 100644 >> --- a/libc/string/arc/arcv2/memset.S >> +++ b/libc/string/arc/arcv2/memset.S >> @@ -52,6 +52,7 @@ ENTRY(memset) >> lpnz @.Lset64bytes >> ;; LOOP START >> PREWRITE(r3, 64) ;Prefetch the next write location >> +#ifdef __LL64__ >> std.ab r4, [r3, 8] >> std.ab r4, [r3, 8] >> std.ab r4, [r3, 8] >> @@ -60,16 +61,45 @@ ENTRY(memset) >> std.ab r4, [r3, 8] >> std.ab r4, [r3, 8] >> std.ab r4, [r3, 8] >> +#else >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> +#endif >> .Lset64bytes: >> >> lsr.f lp_count, r2, 5 ;Last remaining max 124 bytes >> lpnz .Lset32bytes >> ;; LOOP START >> prefetchw [r3, 32] ;Prefetch the next write location >> +#ifdef __LL64__ >> std.ab r4, [r3, 8] >> std.ab r4, [r3, 8] >> std.ab r4, [r3, 8] >> std.ab r4, [r3, 8] >> +#else >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> + st.ab r4, [r3, 4] >> +#endif >> .Lset32bytes: >> >> and.f lp_count, r2, 0x1F ;Last remaining 31 bytes >>
diff --git a/libc/string/arc/arcv2/memset.S b/libc/string/arc/arcv2/memset.S index d076ad1..0918d37 100644 --- a/libc/string/arc/arcv2/memset.S +++ b/libc/string/arc/arcv2/memset.S @@ -52,6 +52,7 @@ ENTRY(memset) lpnz @.Lset64bytes ;; LOOP START PREWRITE(r3, 64) ;Prefetch the next write location +#ifdef __LL64__ std.ab r4, [r3, 8] std.ab r4, [r3, 8] std.ab r4, [r3, 8] @@ -60,16 +61,45 @@ ENTRY(memset) std.ab r4, [r3, 8] std.ab r4, [r3, 8] std.ab r4, [r3, 8] +#else + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] +#endif .Lset64bytes: lsr.f lp_count, r2, 5 ;Last remaining max 124 bytes lpnz .Lset32bytes ;; LOOP START prefetchw [r3, 32] ;Prefetch the next write location +#ifdef __LL64__ std.ab r4, [r3, 8] std.ab r4, [r3, 8] std.ab r4, [r3, 8] std.ab r4, [r3, 8] +#else + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] + st.ab r4, [r3, 4] +#endif .Lset32bytes: and.f lp_count, r2, 0x1F ;Last remaining 31 bytes