From: YunQiang Su <syq@debian.org>
To: Adhemerval Zanella Netto <adhemerval.zanella@linaro.org>
Cc: YunQiang Su <yunqiang.su@cipunited.com>,
libc-alpha@sourceware.org, aurelien@aurel32.net,
jiaxun.yang@flygoat.com, macro@orcam.me.uk
Subject: Re: [PATCH v2] Use in_int32_t_range to check need 64bit syscall
Date: Tue, 8 Nov 2022 10:39:00 +0800 [thread overview]
Message-ID: <CAKcpw6V27brbNV1WBFEQ+3Ksv3O9cW8ajqbKMPqSxyrxbGc1gw@mail.gmail.com> (raw)
In-Reply-To: <CAKcpw6VK3EO9kLbjnPD1wgU2z2GFTA_BT8xm64yuwzFdw-=WNQ@mail.gmail.com>
YunQiang Su <syq@debian.org> 于2022年11月8日周二 10:26写道：
>
> Adhemerval Zanella Netto <adhemerval.zanella@linaro.org> 于2022年11月8日周二 03:49写道：
> >
> >
> >
> > On 03/11/22 22:47, YunQiang Su wrote:
> > > The current version, we use in_time_t_range to detect overflow
> > > of a time. Whenever overflow we try to use 64bit version syscall.
> > >
> > > While the function name is quite confused. So let's use a new
> > > function called in_int32_t_range to do so.
> > >
> > > We still keep in_time_t_range, and use it to detect overflow of
> > > the return value from syscalls.
> >
> > The commit message sounds confusing, maybe:
> >
> > "Currently glibc uses in_time_t_range to detects time_t overflow,
> > and if it occurs fallbacks to 64 bit syscall version.
> >
> > The function name is confusing because internally time_t might be
> > either 32 bits or 64 bits (depending on __TIMESIZE).
> >
> > This patch refactors the in_time_t_range by replacing it with
> > in_int32_t_range for the case to check if the 64 bit time_t syscall
> > should be used.
> >
> > The in_time_t range is used to detect overflow of the
> > syscall return value".
> >
>
> Yes. Your description is much better.
>
> > > ---
> > > include/time.h | 10 +++++++++-
> > > nptl/futex-internal.c | 4 ++--
> > > sysdeps/unix/sysv/linux/clock_adjtime.c | 2 +-
> > > sysdeps/unix/sysv/linux/clock_nanosleep.c | 2 +-
> > > sysdeps/unix/sysv/linux/clock_settime.c | 2 +-
> > > sysdeps/unix/sysv/linux/mq_timedreceive.c | 2 +-
> > > sysdeps/unix/sysv/linux/mq_timedsend.c | 2 +-
> > > sysdeps/unix/sysv/linux/ppoll.c | 2 +-
> > > sysdeps/unix/sysv/linux/pselect.c | 2 +-
> > > sysdeps/unix/sysv/linux/recvmmsg.c | 2 +-
> > > sysdeps/unix/sysv/linux/select.c | 2 +-
> > > sysdeps/unix/sysv/linux/semtimedop.c | 2 +-
> > > sysdeps/unix/sysv/linux/setitimer.c | 4 ++--
> > > sysdeps/unix/sysv/linux/setsockopt.c | 2 +-
> > > sysdeps/unix/sysv/linux/sigtimedwait.c | 2 +-
> > > sysdeps/unix/sysv/linux/timer_settime.c | 4 ++--
> > > sysdeps/unix/sysv/linux/timerfd_settime.c | 4 ++--
> > > sysdeps/unix/sysv/linux/utimensat.c | 4 ++--
> > > 18 files changed, 31 insertions(+), 23 deletions(-)
> > >
> > > diff --git a/include/time.h b/include/time.h
> > > index 20abea69d4..f599eeed4e 100644
> > > --- a/include/time.h
> > > +++ b/include/time.h
> > > @@ -347,12 +347,20 @@ libc_hidden_proto (__time64)
> > > /* Check whether T fits in int32_t, assume all usages are for
> > > sizeof(time_t) == 32. */
> > > static inline bool
> > > -in_time_t_range (__time64_t t)
> > > +in_int32_t_range (__time64_t t)
> > > {
> > > int32_t s = t;
> > > return s == t;
> > > }
> > >
> > > +/* Check whether T fits in time_t. */
> > > +static inline bool
> > > +in_time_t_range (__time64_t t)
> > > +{
> > > + time_t s = t;
> > > + return s == t;
> > > +}
> > > +
> >
> > I think it might be better to avoid define this function for __TIMESIZE == 64,
> > the function always returns true and for 64 bit time_t its occurence might
> > be a misuse.
> >
>
> The reason it exists on 64bit platforms is due to the fact that we may
> need to share
> code on both 32bit and 64bit platforms.
>
> So I think that we should keep it.
>
> For performance, I think that we can make in_time_t_range return true directly.
>
Ohh, the compilers are smart enough to take care of it.
So we don't need it.
> > The rest of the path looks ok, the remaining in_time_t_range usage are only
> > for check of syscall return value (besides the generic usage on mktime and
> > gmtime).
> >
> > > /* Convert a known valid struct timeval into a struct __timespec64. */
> > > static inline struct __timespec64
> > > valid_timeval_to_timespec64 (const struct timeval tv)
> > > diff --git a/nptl/futex-internal.c b/nptl/futex-internal.c
> > > index 7ec228e8fb..6b8ac2304d 100644
> > > --- a/nptl/futex-internal.c
> > > +++ b/nptl/futex-internal.c
> > > @@ -87,7 +87,7 @@ __futex_abstimed_wait_common (unsigned int* futex_word,
> > > err = __futex_abstimed_wait_common64 (futex_word, expected, op, abstime,
> > > private, cancel);
> > > #else
> > > - bool need_time64 = abstime != NULL && !in_time_t_range (abstime->tv_sec);
> > > + bool need_time64 = abstime != NULL && !in_int32_t_range (abstime->tv_sec);
> > > if (need_time64)
> > > {
> > > err = __futex_abstimed_wait_common64 (futex_word, expected, op, abstime,
> > > @@ -162,7 +162,7 @@ __futex_lock_pi64 (int *futex_word, clockid_t clockid,
> > > # ifdef __ASSUME_TIME64_SYSCALLS
> > > err = INTERNAL_SYSCALL_CALL (futex_time64, futex_word, op_pi, 0, abstime);
> > > # else
> > > - bool need_time64 = abstime != NULL && !in_time_t_range (abstime->tv_sec);
> > > + bool need_time64 = abstime != NULL && !in_int32_t_range (abstime->tv_sec);
> > > if (need_time64)
> > > err = INTERNAL_SYSCALL_CALL (futex_time64, futex_word, op_pi, 0, abstime);
> > > else
> > > diff --git a/sysdeps/unix/sysv/linux/clock_adjtime.c b/sysdeps/unix/sysv/linux/clock_adjtime.c
> > > index 5ded82f506..63ea68c3af 100644
> > > --- a/sysdeps/unix/sysv/linux/clock_adjtime.c
> > > +++ b/sysdeps/unix/sysv/linux/clock_adjtime.c
> > > @@ -35,7 +35,7 @@ __clock_adjtime64 (const clockid_t clock_id, struct __timex64 *tx64)
> > > return r;
> > >
> > > if (tx64->modes & ADJ_SETOFFSET
> > > - && ! in_time_t_range (tx64->time.tv_sec))
> > > + && ! in_int32_t_range (tx64->time.tv_sec))
> > > {
> > > __set_errno (EOVERFLOW);
> > > return -1;
> > > diff --git a/sysdeps/unix/sysv/linux/clock_nanosleep.c b/sysdeps/unix/sysv/linux/clock_nanosleep.c
> > > index e610fd4e8d..9031060486 100644
> > > --- a/sysdeps/unix/sysv/linux/clock_nanosleep.c
> > > +++ b/sysdeps/unix/sysv/linux/clock_nanosleep.c
> > > @@ -48,7 +48,7 @@ __clock_nanosleep_time64 (clockid_t clock_id, int flags,
> > > r = INTERNAL_SYSCALL_CANCEL (clock_nanosleep_time64, clock_id, flags, req,
> > > rem);
> > > #else
> > > - if (!in_time_t_range (req->tv_sec))
> > > + if (!in_int32_t_range (req->tv_sec))
> > > {
> > > r = INTERNAL_SYSCALL_CANCEL (clock_nanosleep_time64, clock_id, flags,
> > > req, rem);
> > > diff --git a/sysdeps/unix/sysv/linux/clock_settime.c b/sysdeps/unix/sysv/linux/clock_settime.c
> > > index 2a32e2eb13..db9698ba0d 100644
> > > --- a/sysdeps/unix/sysv/linux/clock_settime.c
> > > +++ b/sysdeps/unix/sysv/linux/clock_settime.c
> > > @@ -41,7 +41,7 @@ __clock_settime64 (clockid_t clock_id, const struct __timespec64 *tp)
> > > if (ret == 0 || errno != ENOSYS)
> > > return ret;
> > >
> > > - if (! in_time_t_range (tp->tv_sec))
> > > + if (! in_int32_t_range (tp->tv_sec))
> > > {
> > > __set_errno (EOVERFLOW);
> > > return -1;
> > > diff --git a/sysdeps/unix/sysv/linux/mq_timedreceive.c b/sysdeps/unix/sysv/linux/mq_timedreceive.c
> > > index 5bf1e0a83b..6c719d4b70 100644
> > > --- a/sysdeps/unix/sysv/linux/mq_timedreceive.c
> > > +++ b/sysdeps/unix/sysv/linux/mq_timedreceive.c
> > > @@ -36,7 +36,7 @@ ___mq_timedreceive_time64 (mqd_t mqdes, char *__restrict msg_ptr, size_t msg_len
> > > msg_prio, abs_timeout);
> > > #else
> > > bool need_time64 = abs_timeout != NULL
> > > - && !in_time_t_range (abs_timeout->tv_sec);
> > > + && !in_int32_t_range (abs_timeout->tv_sec);
> > > if (need_time64)
> > > {
> > > int r = SYSCALL_CANCEL (mq_timedreceive_time64, mqdes, msg_ptr, msg_len,
> > > diff --git a/sysdeps/unix/sysv/linux/mq_timedsend.c b/sysdeps/unix/sysv/linux/mq_timedsend.c
> > > index 3ca5050976..eb8353feef 100644
> > > --- a/sysdeps/unix/sysv/linux/mq_timedsend.c
> > > +++ b/sysdeps/unix/sysv/linux/mq_timedsend.c
> > > @@ -36,7 +36,7 @@ ___mq_timedsend_time64 (mqd_t mqdes, const char *msg_ptr, size_t msg_len,
> > > msg_prio, abs_timeout);
> > > #else
> > > bool need_time64 = abs_timeout != NULL
> > > - && !in_time_t_range (abs_timeout->tv_sec);
> > > + && !in_int32_t_range (abs_timeout->tv_sec);
> > > if (need_time64)
> > > {
> > > int r = SYSCALL_CANCEL (mq_timedsend_time64, mqdes, msg_ptr, msg_len,
> > > diff --git a/sysdeps/unix/sysv/linux/ppoll.c b/sysdeps/unix/sysv/linux/ppoll.c
> > > index 1105e29b00..6653e1bc55 100644
> > > --- a/sysdeps/unix/sysv/linux/ppoll.c
> > > +++ b/sysdeps/unix/sysv/linux/ppoll.c
> > > @@ -43,7 +43,7 @@ __ppoll64 (struct pollfd *fds, nfds_t nfds, const struct __timespec64 *timeout,
> > > __NSIG_BYTES);
> > > #else
> > > int ret;
> > > - bool need_time64 = timeout != NULL && !in_time_t_range (timeout->tv_sec);
> > > + bool need_time64 = timeout != NULL && !in_int32_t_range (timeout->tv_sec);
> > > if (need_time64)
> > > {
> > > ret = SYSCALL_CANCEL (ppoll_time64, fds, nfds, timeout, sigmask,
> > > diff --git a/sysdeps/unix/sysv/linux/pselect.c b/sysdeps/unix/sysv/linux/pselect.c
> > > index eba1c111f8..23fd0dfad6 100644
> > > --- a/sysdeps/unix/sysv/linux/pselect.c
> > > +++ b/sysdeps/unix/sysv/linux/pselect.c
> > > @@ -56,7 +56,7 @@ __pselect64 (int nfds, fd_set *readfds, fd_set *writefds, fd_set *exceptfds,
> > > return pselect64_syscall (nfds, readfds, writefds, exceptfds, timeout,
> > > sigmask);
> > > #else
> > > - bool need_time64 = timeout != NULL && !in_time_t_range (timeout->tv_sec);
> > > + bool need_time64 = timeout != NULL && !in_int32_t_range (timeout->tv_sec);
> > > if (need_time64)
> > > {
> > > int r = pselect64_syscall (nfds, readfds, writefds, exceptfds, timeout,
> > > diff --git a/sysdeps/unix/sysv/linux/recvmmsg.c b/sysdeps/unix/sysv/linux/recvmmsg.c
> > > index d9664d61c4..1ba5a4f25c 100644
> > > --- a/sysdeps/unix/sysv/linux/recvmmsg.c
> > > +++ b/sysdeps/unix/sysv/linux/recvmmsg.c
> > > @@ -35,7 +35,7 @@ recvmmsg_syscall (int fd, struct mmsghdr *vmessages, unsigned int vlen,
> > > struct timespec ts32, *pts32 = NULL;
> > > if (timeout != NULL)
> > > {
> > > - if (! in_time_t_range (timeout->tv_sec))
> > > + if (! in_int32_t_range (timeout->tv_sec))
> > > {
> > > __set_errno (EINVAL);
> > > return -1;
> > > diff --git a/sysdeps/unix/sysv/linux/select.c b/sysdeps/unix/sysv/linux/select.c
> > > index a3f0a2eba7..fd147dfdd4 100644
> > > --- a/sysdeps/unix/sysv/linux/select.c
> > > +++ b/sysdeps/unix/sysv/linux/select.c
> > > @@ -72,7 +72,7 @@ __select64 (int nfds, fd_set *readfds, fd_set *writefds, fd_set *exceptfds,
> > > TIMESPEC_TO_TIMEVAL (timeout, pts64);
> > > return r;
> > > #else
> > > - bool need_time64 = timeout != NULL && !in_time_t_range (timeout->tv_sec);
> > > + bool need_time64 = timeout != NULL && !in_int32_t_range (timeout->tv_sec);
> > > if (need_time64)
> > > {
> > > int r = SYSCALL_CANCEL (pselect6_time64, nfds, readfds, writefds,
> > > diff --git a/sysdeps/unix/sysv/linux/semtimedop.c b/sysdeps/unix/sysv/linux/semtimedop.c
> > > index 38a401bb6f..0cdc0fad4b 100644
> > > --- a/sysdeps/unix/sysv/linux/semtimedop.c
> > > +++ b/sysdeps/unix/sysv/linux/semtimedop.c
> > > @@ -42,7 +42,7 @@ __semtimedop64 (int semid, struct sembuf *sops, size_t nsops,
> > > #ifdef __ASSUME_TIME64_SYSCALLS
> > > return semtimedop_syscall (semid, sops, nsops, timeout);
> > > #else
> > > - bool need_time64 = timeout != NULL && !in_time_t_range (timeout->tv_sec);
> > > + bool need_time64 = timeout != NULL && !in_int32_t_range (timeout->tv_sec);
> > > if (need_time64)
> > > {
> > > int r = semtimedop_syscall (semid, sops, nsops, timeout);
> > > diff --git a/sysdeps/unix/sysv/linux/setitimer.c b/sysdeps/unix/sysv/linux/setitimer.c
> > > index 6d5183aa63..55e6003e0f 100644
> > > --- a/sysdeps/unix/sysv/linux/setitimer.c
> > > +++ b/sysdeps/unix/sysv/linux/setitimer.c
> > > @@ -32,8 +32,8 @@ __setitimer64 (__itimer_which_t which,
> > > #else
> > > struct __itimerval32 new_value_32;
> > >
> > > - if (! in_time_t_range (new_value->it_interval.tv_sec)
> > > - || ! in_time_t_range (new_value->it_value.tv_sec))
> > > + if (! in_int32_t_range (new_value->it_interval.tv_sec)
> > > + || ! in_int32_t_range (new_value->it_value.tv_sec))
> > > {
> > > __set_errno (EOVERFLOW);
> > > return -1;
> > > diff --git a/sysdeps/unix/sysv/linux/setsockopt.c b/sysdeps/unix/sysv/linux/setsockopt.c
> > > index 4292f411f3..6506ca03ab 100644
> > > --- a/sysdeps/unix/sysv/linux/setsockopt.c
> > > +++ b/sysdeps/unix/sysv/linux/setsockopt.c
> > > @@ -54,7 +54,7 @@ setsockopt32 (int fd, int level, int optname, const void *optval,
> > > }
> > >
> > > struct __timeval64 *tv64 = (struct __timeval64 *) optval;
> > > - if (! in_time_t_range (tv64->tv_sec))
> > > + if (! in_int32_t_range (tv64->tv_sec))
> > > {
> > > __set_errno (EOVERFLOW);
> > > break;
> > > diff --git a/sysdeps/unix/sysv/linux/sigtimedwait.c b/sysdeps/unix/sysv/linux/sigtimedwait.c
> > > index 069fa1d5e4..1993e3f85a 100644
> > > --- a/sysdeps/unix/sysv/linux/sigtimedwait.c
> > > +++ b/sysdeps/unix/sysv/linux/sigtimedwait.c
> > > @@ -31,7 +31,7 @@ __sigtimedwait64 (const sigset_t *set, siginfo_t *info,
> > > result = SYSCALL_CANCEL (rt_sigtimedwait_time64, set, info, timeout,
> > > __NSIG_BYTES);
> > > #else
> > > - bool need_time64 = timeout != NULL && !in_time_t_range (timeout->tv_sec);
> > > + bool need_time64 = timeout != NULL && !in_int32_t_range (timeout->tv_sec);
> > > if (need_time64)
> > > {
> > > result = SYSCALL_CANCEL (rt_sigtimedwait_time64, set, info, timeout,
> > > diff --git a/sysdeps/unix/sysv/linux/timer_settime.c b/sysdeps/unix/sysv/linux/timer_settime.c
> > > index 518fed9c59..056e0aff59 100644
> > > --- a/sysdeps/unix/sysv/linux/timer_settime.c
> > > +++ b/sysdeps/unix/sysv/linux/timer_settime.c
> > > @@ -46,8 +46,8 @@ ___timer_settime64 (timer_t timerid, int flags,
> > > # endif
> > > struct itimerspec its32, oits32;
> > >
> > > - if (! in_time_t_range ((value->it_value).tv_sec)
> > > - || ! in_time_t_range ((value->it_interval).tv_sec))
> > > + if (! in_int32_t_range ((value->it_value).tv_sec)
> > > + || ! in_int32_t_range ((value->it_interval).tv_sec))
> > > {
> > > __set_errno (EOVERFLOW);
> > > return -1;
> > > diff --git a/sysdeps/unix/sysv/linux/timerfd_settime.c b/sysdeps/unix/sysv/linux/timerfd_settime.c
> > > index 59282bf1ad..c5f3f14c3b 100644
> > > --- a/sysdeps/unix/sysv/linux/timerfd_settime.c
> > > +++ b/sysdeps/unix/sysv/linux/timerfd_settime.c
> > > @@ -33,8 +33,8 @@ __timerfd_settime64 (int fd, int flags, const struct __itimerspec64 *value,
> > > #ifdef __ASSUME_TIME64_SYSCALLS
> > > return INLINE_SYSCALL_CALL (timerfd_settime64, fd, flags, value, ovalue);
> > > #else
> > > - bool need_time64 = !in_time_t_range (value->it_value.tv_sec)
> > > - || !in_time_t_range (value->it_interval.tv_sec);
> > > + bool need_time64 = !in_int32_t_range (value->it_value.tv_sec)
> > > + || !in_int32_t_range (value->it_interval.tv_sec);
> > > if (need_time64)
> > > {
> > > int r = INLINE_SYSCALL_CALL (timerfd_settime64, fd, flags, value,
> > > diff --git a/sysdeps/unix/sysv/linux/utimensat.c b/sysdeps/unix/sysv/linux/utimensat.c
> > > index 5662ecc9aa..54a5e47618 100644
> > > --- a/sysdeps/unix/sysv/linux/utimensat.c
> > > +++ b/sysdeps/unix/sysv/linux/utimensat.c
> > > @@ -41,9 +41,9 @@ __utimensat64_helper (int fd, const char *file,
> > >
> > > bool need_time64 = tsp64 != NULL
> > > && ((!TS_SPECIAL (tsp64[0])
> > > - && !in_time_t_range (tsp64[0].tv_sec))
> > > + && !in_int32_t_range (tsp64[0].tv_sec))
> > > || (!TS_SPECIAL (tsp64[1])
> > > - && !in_time_t_range (tsp64[1].tv_sec)));
> > > + && !in_int32_t_range (tsp64[1].tv_sec)));
> > > if (need_time64)
> > > {
> > > int r = INLINE_SYSCALL_CALL (utimensat_time64, fd, file, &tsp64[0],
prev parent reply other threads:[~2022-11-08 2:39 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-11-04 1:47 YunQiang Su
2022-11-07 19:49 ` Adhemerval Zanella Netto
2022-11-08 2:26 ` YunQiang Su
2022-11-08 2:39 ` YunQiang Su [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=CAKcpw6V27brbNV1WBFEQ+3Ksv3O9cW8ajqbKMPqSxyrxbGc1gw@mail.gmail.com \
--to=syq@debian.org \
--cc=adhemerval.zanella@linaro.org \
--cc=aurelien@aurel32.net \
--cc=jiaxun.yang@flygoat.com \
--cc=libc-alpha@sourceware.org \
--cc=macro@orcam.me.uk \
--cc=yunqiang.su@cipunited.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).