From: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
To: Alex Xu <alex_y_xu@yahoo.ca>
Cc: lttng-dev <lttng-dev@lists.lttng.org>
Subject: Re: [PATCH 2/2] Use _umtx_op for futex on FreeBSD
Date: Mon, 27 Jan 2020 13:48:24 -0500 (EST) [thread overview]
Message-ID: <1117636152.600695.1580150904852.JavaMail.zimbra@efficios.com> (raw)
In-Reply-To: <20200127182526.22881-2-alex_y_xu@yahoo.ca>
----- On Jan 27, 2020, at 1:25 PM, lttng-dev lttng-dev@lists.lttng.org wrote:
> ---
> include/urcu/futex.h | 23 +++++++++++++++++++++++
> 1 file changed, 23 insertions(+)
>
> diff --git a/include/urcu/futex.h b/include/urcu/futex.h
> index c206c6f..33dc3db 100644
> --- a/include/urcu/futex.h
> +++ b/include/urcu/futex.h
> @@ -24,6 +24,7 @@
> */
>
> #include <urcu/config.h>
> +#include <errno.h>
> #include <stdint.h>
> #include <time.h>
>
> @@ -103,6 +104,28 @@ static inline int futex_async(int32_t *uaddr, int op,
> int32_t val,
> return ret;
> }
>
> +#elif defined(__FreeBSD__)
> +
> +#include <sys/umtx.h>
> +
> +static inline int futex_noasync(int32_t *uaddr, int op, int32_t val,
> + const struct timespec *timeout, int32_t *uaddr2, int32_t val3)
> +{
> + return futex_async(uaddr, op, val, timeout, uaddr2, val3);
> +}
Please move the futex_noasync definition after futex_async, considering that
futex_noasync uses futex_async().
> +
> +static inline int futex_async(int32_t *uaddr, int op, int32_t val,
> + const struct timespec *timeout, int32_t *uaddr2, int32_t val3)
> +{
Please declare a separate variable for umtx, e.g.:
int umtx_op;
> + switch (op) {
> + case FUTEX_WAIT: op = UMTX_OP_WAIT_UINT; break;
> + case FUTEX_WAKE: op = UMTX_OP_WAKE; break;
> + default: errno = EINVAL; return -1;
> + }
Please follow the liburcu coding style, based on the Linux kernel
coding style:
switch (op) {
case FUTEX_WAIT:
umtx_op = UMTX_OP_WAIT_UINT;
break;
case FUTEX_WAKE:
umtx_op = UMTX_OP_WAKE;
break;
default:
errno = EINVAL;
return -1;
}
Also, do you have a link to the API documentation of _umtx_op() ?
I would like to double-check that all its return values are
expected by the callers.
> + return _umtx_op(uaddr, op, val,
> + (void *)sizeof(*timeout), (void *)timeout);
return _umtx_op(uaddr, umtx_op, val,
(void *)sizeof(*timeout), (void *)timeout);
Is it OK that uaddr is a int32_t ? Is it expected to be some
other layout ? (e.g. struct umutex ?) I guess UMTX_OP_WAIT_UINT
somehow specifies this, but it would be good to document it
with a comment.
Thanks,
Mathieu
> +}
> +
> #elif defined(__CYGWIN__)
>
> /*
> --
> 2.25.0
>
> _______________________________________________
> lttng-dev mailing list
> lttng-dev@lists.lttng.org
> https://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev
--
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com
next prev parent reply other threads:[~2020-01-27 18:48 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <20200127182526.22881-1-alex_y_xu.ref@yahoo.ca>
2020-01-27 18:25 ` [PATCH 1/2] Add FreeBSD, DragonFly to syscall-compat.h Alex Xu (Hello71) via lttng-dev
2020-01-27 18:25 ` [PATCH 2/2] Use _umtx_op for futex on FreeBSD Alex Xu (Hello71) via lttng-dev
2020-01-27 18:48 ` Mathieu Desnoyers [this message]
2020-01-27 19:43 ` Alex Xu via lttng-dev
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=1117636152.600695.1580150904852.JavaMail.zimbra@efficios.com \
--to=mathieu.desnoyers@efficios.com \
--cc=alex_y_xu@yahoo.ca \
--cc=lttng-dev@lists.lttng.org \
/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 NNTP newsgroup(s).