Re: [LTP] [PATCH 1/3] lib: introduce safe_write_fully()

[Cyril Hrubis <chrubis@suse.cz>]

Hi!
> In case there is a short (but otherwise successful) write(),
> safe_write_fully() repeats write() and attempts to resume
> with the remainder of the buffer.
> 
> Signed-off-by: Jan Stancek <jstancek@redhat.com>
> ---
>  include/tst_safe_macros.h |  5 +++++
>  lib/tst_safe_macros.c     | 19 +++++++++++++++++++
>  2 files changed, 24 insertions(+)
> 
> diff --git a/include/tst_safe_macros.h b/include/tst_safe_macros.h
> index 81c4b0844267..caee0e9cf842 100644
> --- a/include/tst_safe_macros.h
> +++ b/include/tst_safe_macros.h
> @@ -645,4 +645,9 @@ int safe_sysinfo(const char *file, const int lineno, struct sysinfo *info);
>  #define SAFE_SYSINFO(info) \
>  	safe_sysinfo(__FILE__, __LINE__, (info))
>  
> +ssize_t safe_write_fully(const char *file, const int lineno,
> +	int fildes, const void *buf, size_t nbyte);
> +#define SAFE_WRITE_FULLY(fildes, buf, nbyte) \
> +	safe_write_fully(__FILE__, __LINE__, (fildes), (buf), (nbyte))

We already have a flag for SAFE_WRITE() to fail/not-fail on partial
write, what about turning that into three way switch?

Something as:

enum safe_write_opts {
	SAFE_WRITE_PARTIAL = 0,
	SAFE_WRITE_FULL = 1,
	SAFE_WRITE_RETRY = 2,
};

Or maybe just rename the SAFE_WRITE_FULLY() to SAFE_WRITE_RETRY().

>  #endif /* SAFE_MACROS_H__ */
> diff --git a/lib/tst_safe_macros.c b/lib/tst_safe_macros.c
> index c4cdc87e7346..e4d4ef4269a4 100644
> --- a/lib/tst_safe_macros.c
> +++ b/lib/tst_safe_macros.c
> @@ -591,3 +591,22 @@ void safe_cmd(const char *file, const int lineno, const char *const argv[],
>  		tst_brk_(file, lineno, TBROK, "%s failed (%d)", argv[0], rval);
>  	}
>  }
> +
> +ssize_t safe_write_fully(const char *file, const int lineno,
> +	int fildes, const void *buf, size_t nbyte)
> +{
> +	ssize_t rval;
> +	size_t len = nbyte;
> +
> +	do {
> +		rval = write(fildes, buf, len);
> +		if (rval == -1) {
> +			tst_brk_(file, lineno, TBROK | TERRNO,
> +				"write(%d,%p,%zu) failed", fildes, buf, len);

I guess that we may print potentionally confusing output here since we
modify the buf and len in the loop. I guess that we should store the buf
pointer at the start just for a case of this message and use the nbyte
and possibly write how many bytes we have managed to write before the
failure.

> +		}
> +		buf += rval;
> +		len -= rval;
> +	} while (len > 0);
> +
> +	return nbyte;
> +}
> -- 
> 2.27.0
> 
> 
> -- 
> Mailing list info: https://lists.linux.it/listinfo/ltp

-- 
Cyril Hrubis
chrubis@suse.cz

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp