Re: V2 [PATCH] Fix __clock_gettime64 for sysdeps/generic/hp-timing.h

[Lukasz Majewski <lukma@denx.de>]

[-- Attachment #1: Type: text/plain, Size: 16270 bytes --]

On Sun, 24 May 2020 16:34:44 -0700
"H.J. Lu" <hjl.tools@gmail.com> wrote:

> On Sun, May 24, 2020 at 9:13 AM H.J. Lu <hjl.tools@gmail.com> wrote:
> >
> > On Sun, May 24, 2020 at 8:30 AM Lukasz Majewski <lukma@denx.de>
> > wrote:  
> > >
> > > On Sun, 24 May 2020 04:39:47 -0700
> > > "H.J. Lu" <hjl.tools@gmail.com> wrote:
> > >  
> > > > On Sun, May 24, 2020 at 2:03 AM Lukasz Majewski <lukma@denx.de>
> > > > wrote:  
> > > > >
> > > > > On Sat, 23 May 2020 13:32:05 -0700
> > > > > "H.J. Lu" <hjl.tools@gmail.com> wrote:
> > > > >  
> > > > > > On Thu, May 21, 2020 at 4:11 AM H.J. Lu
> > > > > > <hjl.tools@gmail.com> wrote:  
> > > > > > >
> > > > > > > On Thu, May 21, 2020 at 3:24 AM Lukasz Majewski
> > > > > > > <lukma@denx.de> wrote:  
> > > > > > > >
> > > > > > > > Hi "H.J. Lu",
> > > > > > > >  
> > > > > > > > > On Tue, May 5, 2020 at 11:21 AM Lukasz Majewski
> > > > > > > > > <lukma@denx.de> wrote:  
> > > > > > > > > >
> > > > > > > > > > The __clock_gettime internal function is not
> > > > > > > > > > supporting 64 bit time on architectures with
> > > > > > > > > > __WORDSIZE == 32 and __TIMESIZE != 64 (like e.g.
> > > > > > > > > > ARM 32 bit).
> > > > > > > > > >
> > > > > > > > > > The __clock_gettime64 function shall be used
> > > > > > > > > > instead in the glibc itself as it supports 64 bit
> > > > > > > > > > time on those systems. This patch does not bring
> > > > > > > > > > any changes to systems with __WORDSIZE == 64 as for
> > > > > > > > > > them the __clock_gettime64 is aliased to
> > > > > > > > > > __clock_gettime (in ./include/time.h).
> > > > > > > > > >
> > > > > > > > > > ---
> > > > > > > > > > Changes for v3:
> > > > > > > > > > - Fix the commit message (add missing "function"
> > > > > > > > > > after __clock_gettime64)
> > > > > > > > > > - Refactor __nisfind_server() function to support
> > > > > > > > > > 64 bit time with struct timespec instead of struct
> > > > > > > > > > timeval
> > > > > > > > > > - Change expires to __time64_t
> > > > > > > > > >
> > > > > > > > > > Changes for v2:
> > > > > > > > > > - Use only TIMESPEC_TO_TIMEVAL instead of
> > > > > > > > > > valid_timespec64_to_timeval in logout.c and
> > > > > > > > > > logwtmp.c as it is generic enough to also support
> > > > > > > > > > struct __timespec64 conversion to struct timeval ---
> > > > > > > > > >  benchtests/bench-timing.h                        |
> > > > > > > > > >  2 +- include/random-bits.h
> > > > > > > > > >    |  4 ++-- login/logout.c
> > > > > > > > > >           | 4 ++-- login/logwtmp.c
> > > > > > > > > >                 | 5 +++-- nis/nis_call.c
> > > > > > > > > > | 16 +++++++--------- sysdeps/generic/hp-timing.h
> > > > > > > > > >     | 4 ++-- sysdeps/generic/memusage.h
> > > > > > > > > > |  4 ++--
> > > > > > > > > > sysdeps/unix/sysv/linux/alpha/osf_gettimeofday.c |
> > > > > > > > > > 4 ++-- sysdeps/unix/sysv/linux/clock.c
> > > > > > > > > >     |  7 ++----- 9 files changed, 23 insertions(+),
> > > > > > > > > > 27 deletions(-)  
> > > > > > > > >
> > > > > > > > > Does this patch work with "make bench" when
> > > > > > > > > sysdeps/generic/hp-timing.h is used, like on arm?  I
> > > > > > > > > got In file included from ./bench-timing.h:23,
> > > > > > > > >                  from ./bench-skeleton.c:25,
> > > > > > > > >                  from
> > > > > > > > > /export/build/gnu/tools-build/glibc-gitlab/build-x86_64-linux/benchtests/bench-rint.c:45:
> > > > > > > > > ./bench-skeleton.c: In function ‘main’:
> > > > > > > > > ../sysdeps/generic/hp-timing.h:37:23: error: storage
> > > > > > > > > size of ‘tv’ isn’t known 37 |   struct __timespec64
> > > > > > > > > tv;      \ |                       ^~
> > > > > > > > > ./bench-timing.h:35:25: note: in expansion of macro
> > > > > > > > > ‘HP_TIMING_NOW’ 35 | #define TIMING_NOW(var)
> > > > > > > > > HP_TIMING_NOW (var) |
> > > > > > > > > ^~~~~~~~~~~~~ ./bench-skeleton.c:82:8: note: in
> > > > > > > > > expansion of macro ‘TIMING_NOW’ 82 |
> > > > > > > > > TIMING_NOW (start); |        ^~~~~~~~~~
> > > > > > > > > ../sysdeps/generic/hp-timing.h:38:3: error: implicit
> > > > > > > > > declaration of function ‘__clock_gettime64’; did you
> > > > > > > > > mean ‘clock_gettime’?
> > > > > > > > > [-Werror=implicit-function-declaration] 38 |
> > > > > > > > > __clock_gettime64 (CLOCK_MONOTONIC, &tv);   \ |
> > > > > > > > > ^~~~~~~~~~~~~~~~~ ./bench-timing.h:35:25: note: in
> > > > > > > > > expansion of macro ‘HP_TIMING_NOW’ 35 | #define
> > > > > > > > > TIMING_NOW(var) HP_TIMING_NOW (var) |
> > > > > > > > >         ^~~~~~~~~~~~~ ./bench-skeleton.c:82:8: note:
> > > > > > > > > in expansion of macro ‘TIMING_NOW’ 82 |
> > > > > > > > > TIMING_NOW (start); |        ^~~~~~~~~~
> > > > > > > > > ../sysdeps/generic/hp-timing.h:37:23: error: unused
> > > > > > > > > variable ‘tv’ [-Werror=unused-variable]
> > > > > > > > >    37 |   struct __timespec64 tv;      \
> > > > > > > > >       |                       ^~
> > > > > > > > > ./bench-timing.h:35:25: note: in expansion of macro
> > > > > > > > > ‘HP_TIMING_NOW’ 35 | #define TIMING_NOW(var)
> > > > > > > > > HP_TIMING_NOW (var) |
> > > > > > > > > ^~~~~~~~~~~~~ ./bench-skeleton.c:82:8: note: in
> > > > > > > > > expansion of macro ‘TIMING_NOW’ 82 |
> > > > > > > > > TIMING_NOW (start); |        ^~~~~~~~~~
> > > > > > > > > ../sysdeps/generic/hp-timing.h:37:23: error: storage
> > > > > > > > > size of ‘tv’ isn’t known 37 |   struct __timespec64
> > > > > > > > > tv;      \ |                       ^~
> > > > > > > > > ./bench-timing.h:35:25: note: in expansion of macro
> > > > > > > > > ‘HP_TIMING_NOW’ 35 | #define TIMING_NOW(var)
> > > > > > > > > HP_TIMING_NOW (var) |
> > > > > > > > > ^~~~~~~~~~~~~ ./bench-skeleton.c:86:8: note: in
> > > > > > > > > expansion of macro ‘TIMING_NOW’ 86 |
> > > > > > > > > TIMING_NOW (end); |        ^~~~~~~~~~
> > > > > > > > > ../sysdeps/generic/hp-timing.h:37:23: error: unused
> > > > > > > > > variable ‘tv’ [-Werror=unused-variable]
> > > > > > > > >    37 |   struct __timespec64 tv;      \
> > > > > > > > >       |                       ^~
> > > > > > > > > ./bench-timing.h:35:25: note: in expansion of macro
> > > > > > > > > ‘HP_TIMING_NOW’ 35 | #define TIMING_NOW(var)
> > > > > > > > > HP_TIMING_NOW (var) |
> > > > > > > > > ^~~~~~~~~~~~~ ./bench-skeleton.c:86:8: note: in
> > > > > > > > > expansion of macro ‘TIMING_NOW’ 86 |
> > > > > > > > > TIMING_NOW (end); |        ^~~~~~~~~~
> > > > > > > > > ../sysdeps/generic/hp-timing.h:37:23: error: storage
> > > > > > > > > size of ‘tv’ isn’t known 37 |   struct __timespec64
> > > > > > > > > tv;      \ |                       ^~
> > > > > > > > > ./bench-timing.h:35:25: note: in expansion of macro
> > > > > > > > > ‘HP_TIMING_NOW’ 35 | #define TIMING_NOW(var)
> > > > > > > > > HP_TIMING_NOW (var) |
> > > > > > > > > ^~~~~~~~~~~~~ ./bench-skeleton.c:90:8: note: in
> > > > > > > > > expansion of macro ‘TIMING_NOW’ 90 |
> > > > > > > > > TIMING_NOW (start); |        ^~~~~~~~~~
> > > > > > > > > ../sysdeps/generic/hp-timing.h:37:23: error: unused
> > > > > > > > > variable ‘tv’ [-Werror=unused-variable]
> > > > > > > > >    37 |   struct __timespec64 tv;      \
> > > > > > > > >       |                       ^~
> > > > > > > > > ./bench-timing.h:35:25: note: in expansion of macro
> > > > > > > > > ‘HP_TIMING_NOW’ 35 | #define TIMING_NOW(var)
> > > > > > > > > HP_TIMING_NOW (var) |
> > > > > > > > > ^~~~~~~~~~~~~ ./bench-skeleton.c:90:8: note: in
> > > > > > > > > expansion of macro ‘TIMING_NOW’ 90 |
> > > > > > > > > TIMING_NOW (start); |        ^~~~~~~~~~
> > > > > > > > > ../sysdeps/generic/hp-timing.h:37:23: error: storage
> > > > > > > > > size of ‘tv’ isn’t known 37 |   struct __timespec64
> > > > > > > > > tv;      \ |                       ^~
> > > > > > > > > ./bench-timing.h:35:25: note: in expansion of macro
> > > > > > > > > ‘HP_TIMING_NOW’ 35 | #define TIMING_NOW(var)
> > > > > > > > > HP_TIMING_NOW (var) |
> > > > > > > > > ^~~~~~~~~~~~~ ./bench-skeleton.c:94:8: note: in
> > > > > > > > > expansion of macro ‘TIMING_NOW’ 94 |
> > > > > > > > > TIMING_NOW (end); |        ^~~~~~~~~~
> > > > > > > > > ../sysdeps/generic/hp-timing.h:37:23: error: unused
> > > > > > > > > variable ‘tv’ [-Werror=unused-variable]
> > > > > > > > >    37 |   struct __timespec64 tv;      \
> > > > > > > > >       |                       ^~
> > > > > > > > > ./bench-timing.h:35:25: note: in expansion of macro
> > > > > > > > > ‘HP_TIMING_NOW’ 35 | #define TIMING_NOW(var)
> > > > > > > > > HP_TIMING_NOW (var) |
> > > > > > > > > ^~~~~~~~~~~~~ ./bench-skeleton.c:94:8: note: in
> > > > > > > > > expansion of macro ‘TIMING_NOW’ 94 |
> > > > > > > > > TIMING_NOW (end); |        ^~~~~~~~~~
> > > > > > > > > ../sysdeps/generic/hp-timing.h:37:23: error: storage
> > > > > > > > > size of ‘tv’ isn’t known 37 |   struct __timespec64
> > > > > > > > > tv;      \ |                       ^~
> > > > > > > > > ./bench-timing.h:35:25: note: in expansion of macro
> > > > > > > > > ‘HP_TIMING_NOW’ 35 | #define TIMING_NOW(var)
> > > > > > > > > HP_TIMING_NOW (var) |
> > > > > > > > > ^~~~~~~~~~~~~ ./bench-skeleton.c:103:3: note: in
> > > > > > > > > expansion of macro ‘TIMING_NOW’ 103 |   TIMING_NOW
> > > > > > > > > (start); |   ^~~~~~~~~~
> > > > > > > > > ../sysdeps/generic/hp-timing.h:37:23: error: unused
> > > > > > > > > variable ‘tv’ [-Werror=unused-variable]
> > > > > > > > >    37 |   struct __timespec64 tv;      \
> > > > > > > > >       |                       ^~
> > > > > > > > > ./bench-timing.h:35:25: note: in expansion of macro
> > > > > > > > > ‘HP_TIMING_NOW’ 35 | #define TIMING_NOW(var)
> > > > > > > > > HP_TIMING_NOW (var) |
> > > > > > > > > ^~~~~~~~~~~~~ ./bench-skeleton.c:103:3: note: in
> > > > > > > > > expansion of macro ‘TIMING_NOW’ 103 |   TIMING_NOW
> > > > > > > > > (start); |   ^~~~~~~~~~
> > > > > > > > > ../sysdeps/generic/hp-timing.h:37:23: error: storage
> > > > > > > > > size of ‘tv’ isn’t known 37 |   struct __timespec64
> > > > > > > > > tv;      \ |                       ^~
> > > > > > > > > ./bench-timing.h:35:25: note: in expansion of macro
> > > > > > > > > ‘HP_TIMING_NOW’ 35 | #define TIMING_NOW(var)
> > > > > > > > > HP_TIMING_NOW (var) |
> > > > > > > > > ^~~~~~~~~~~~~ ./bench-skeleton.c:106:3: note: in
> > > > > > > > > expansion of macro ‘TIMING_NOW’ 106 |   TIMING_NOW
> > > > > > > > > (end); |   ^~~~~~~~~~
> > > > > > > > > ../sysdeps/generic/hp-timing.h:37:23: error: unused
> > > > > > > > > variable ‘tv’ [-Werror=unused-variable]
> > > > > > > > >    37 |   struct __timespec64 tv;      \
> > > > > > > > >       |                       ^~
> > > > > > > > > ./bench-timing.h:35:25: note: in expansion of macro
> > > > > > > > > ‘HP_TIMING_NOW’ 35 | #define TIMING_NOW(var)
> > > > > > > > > HP_TIMING_NOW (var) |
> > > > > > > > > ^~~~~~~~~~~~~ ./bench-skeleton.c:106:3: note: in
> > > > > > > > > expansion of macro ‘TIMING_NOW’ 106 |   TIMING_NOW
> > > > > > > > > (end); |   ^~~~~~~~~~
> > > > > > > > > cc1: all warnings being treated as errors  
> > > > > > > >
> > > > > > > > How shall I reproduce this error?  
> > > > > > >
> > > > > > > Apply the enclosed patch and do
> > > > > > >
> > > > > > > $ make USE_CLOCK_GETTIME=1 bench
> > > > > > >  
> > > > > > > > Was it supposed to be catch with build-many-glibc.py ?
> > > > > > > >  
> > > > > > >
> > > > > > > No, "make bench" isn't tested by build-many-glibc.py.
> > > > > > >  
> > > > > >
> > > > > > Here is the patch to fix __clock_gettime64 for
> > > > > > sysdeps/generic/hp-timing.h. OK for master?
> > > > > >  
> > > > >
> > > > > Thanks for preparing the fix.
> > > > >
> > > > > If I may ask - I do have some question about the following
> > > > > code: +#else
> > > > > +struct __timespec64;
> > > > > +extern int __clock_gettime64 (clockid_t clock_id, struct
> > > > > __timespec64 *tp); +#endif
> > > > >
> > > > > Why do we need the "struct __timespec64;" declaration there as
> > > > > in ./sysdeps/generic/hp-timing.h you explicitly added:
> > > > > #include <struct___timespec64.h> ?
> > > > >  
> > > >
> > > > include/time.h is also included by other files which may not
> > > > include <struct___timespec64.h>.
> > > >  
> > >
> > > In which files you observe this situation?
> > >
> > > The <struct___timespec64.h> was introduced to avoid including
> > > time.h when only new, y2038 safe struct __timespec64 is necessary.
> > >  
> >
> > Without it, I got
> >
> > cc -m32 getprtent_r.c -c -std=gnu11 -fgnu89-inline  -O2 -g
> > -march=i686 -Wall -Wwrite-strings -Wundef -Werror
> > -fmerge-all-constants -frounding-math -fno-stack-protector
> > -Wstrict-prototypes -Wold-style-definition -fmath-errno   -fPIC
> > -Wa,-mtune=i686 -fexceptions  -ftls-model=initial-exec
> > -I../include
> > -I/export/build/gnu/tools-build/glibc-32bit-gitlab/build-i686-linux/inet
> > -I/export/build/gnu/tools-build/glibc-32bit-gitlab/build-i686-linux
> > -I../sysdeps/unix/sysv/linux/i386/i686  -I../sysdeps/i386/i686/nptl
> > -I../sysdeps/unix/sysv/linux/i386
> > -I../sysdeps/unix/sysv/linux/x86/include
> > -I../sysdeps/unix/sysv/linux/x86  -I../sysdeps/x86/nptl
> > -I../sysdeps/i386/nptl  -I../sysdeps/unix/sysv/linux/include
> > -I../sysdeps/unix/sysv/linux  -I../sysdeps/nptl
> > -I../sysdeps/pthread -I../sysdeps/gnu  -I../sysdeps/unix/inet
> > -I../sysdeps/unix/sysv -I../sysdeps/unix/i386  -I../sysdeps/unix
> > -I../sysdeps/posix -I../sysdeps/i386/i686/fpu/multiarch
> > -I../sysdeps/i386/i686/fpu -I../sysdeps/i386/i686/multiarch
> > -I../sysdeps/i386/i686 -I../sysdeps/i386/fpu  -I../sysdeps/x86/fpu
> > -I../sysdeps/i386 -I../sysdeps/x86  -I../sysdeps/wordsize-32
> > -I../sysdeps/ieee754/float128  -I../sysdeps/ieee754/ldbl-96/include
> > -I../sysdeps/ieee754/ldbl-96  -I../sysdeps/ieee754/dbl-64
> > -I../sysdeps/ieee754/flt-32  -I../sysdeps/ieee754
> > -I../sysdeps/generic  -I.. -I../libio -I.   -D_LIBC_REENTRANT
> > -include
> > /export/build/gnu/tools-build/glibc-32bit-gitlab/build-i686-linux/libc-modules.h
> > -DMODULE_NAME=libc -include ../include/libc-symbols.h  -DPIC
> > -DSHARED -DTOP_NAMESPACE=glibc -o
> > /export/build/gnu/tools-build/glibc-32bit-gitlab/build-i686-linux/inet/getprtent_r.os
> > -MD -MP -MF
> > /export/build/gnu/tools-build/glibc-32bit-gitlab/build-i686-linux/inet/getprtent_r.os.dt
> > -MT
> > /export/build/gnu/tools-build/glibc-32bit-gitlab/build-i686-linux/inet/getprtent_r.os
> > In file included from ../sysdeps/nptl/lowlevellock.h:24, from
> > ../sysdeps/unix/sysv/linux/x86/lowlevellock.h:23, from
> > ../nptl/descr.h:29, from ../sysdeps/i386/nptl/tls.h:120, from
> > ../sysdeps/i386/i686/nptl/tls.h:33, from ../include/link.h:51, from
> > ../include/dlfcn.h:4, from ../nss/nsswitch.h:28, from
> > ../include/nsswitch.h:1, from ../inet/netgroup.h:22,
> >                  from ../include/netdb.h:203,
> >                  from getprtent_r.c:18:
> > ../include/time.h:9:58: error: ‘struct __timespec64’ declared inside
> > parameter list will not be visible outside of this definition or
> > declaration [-Werror]
> >     9 | extern int __clock_gettime64 (clockid_t clock_id, struct
> > __timespec64 *tp);
> >       |
> > ^~~~~~~~~~~~ cc1: all warnings being treated as errors
> >
> > This may happen for all 32-bit targets.  
> 
> We can't change include/time.h.   Here is the updated patch
> to put __clock_gettime64 in sysdeps/generic/hp-timing.h if
> _ISOMAC is defined.
> 

It looks fine for me, but I think that some other (i.e. Joseph, Florian)
glibc developers shall also review it.


Best regards,

Lukasz Majewski

--

DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email: lukma@denx.de

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]