standards/186028: incorrect return values for posix_fallocate()

[Bob Bishop]

Hi,

On 23 Jan 2014, at 09:40, Konstantin Belousov <kostikbel at gmail.com> wrote:

> On Thu, Jan 23, 2014 at 08:58:58AM +0000, Gennady Proskurin wrote:
>> 
>>> Number:         186028
>>> Category:       standards
>>> Synopsis:       incorrect return values for posix_fallocate()
>>> Confidential:   no
>>> Severity:       non-critical
>>> Priority:       low
>>> Responsible:    freebsd-standards
>>> State:          open
>>> Quarter:        
>>> Keywords:       
>>> Date-Required:
>>> Class:          sw-bug
>>> Submitter-Id:   current-users
>>> Arrival-Date:   Thu Jan 23 09:00:00 UTC 2014
>>> Closed-Date:
>>> Last-Modified:
>>> Originator:     Gennady Proskurin
>>> Release:        FreeBSD 11.0-CURRENT
>>> Organization:
>>> Environment:
>> FreeBSD gpr.nnz-home.ru 11.0-CURRENT FreeBSD 11.0-CURRENT #0 r260472+743aa78(svn_head): Fri Jan 10 05:28:05 MSK 2014     gpr at gpr.nnz-home.ru:/usr/obj/usr/src/freebsd-head/sys/GPR  amd64
>>> Description:
>> In case of error, posix_fallocate() should return error code itself, but in FreeBSD it returns -1 and sets errno.

FWIW at least posix_madvise() and posix_fadvise() have the same problem.

>> Quote from standard:
>> http://pubs.opengroup.org/onlinepubs/009695399/functions/posix_fallocate.html
>> RETURN VALUE
>>    Upon successful completion, posix_fallocate() shall return zero; otherwise, an error number shall be returned to indicate the error.
>> 
>> 
>> Quote from freebsd man:
>> RETURN VALUES
>>     If successful, posix_fallocate() returns zero.  It returns -1 on failure,
>>     and sets errno to indicate the error.
>> 
>>> How-To-Repeat:
>> test program attached
>>> Fix:
>> 
>> 
>> Patch attached with submission follows:
>> 
>> #include <fcntl.h>
>> #include <errno.h>
>> #include <string.h>
>> #include <stdio.h>
>> 
>> int main()
>> {
>> 	int ret;
>> 	int err;
>> 
>> 	errno = 0;
>> 	ret = posix_fallocate(-1 /* emulate EBADF error */, 0, 1);
>> 	err = errno;
>> 	printf("return value : %i   strerror: %s\n", ret, strerror(ret));
>> 	printf("errno        : %i   strerror: %s\n", err, strerror(err));
>> }
>> 
> 
> Indeed.  Linux also seems to have the conforming behaviour, according
> to their man page, which also explicitely states that errno is not set.
> 
> Try this.
> 
> diff --git a/lib/libc/sys/posix_fallocate.2 b/lib/libc/sys/posix_fallocate.2
> index 087c68c..ee6fcc4 100644
> --- a/lib/libc/sys/posix_fallocate.2
> +++ b/lib/libc/sys/posix_fallocate.2
> @@ -83,9 +83,9 @@ that reduces the file size to a size smaller than
> If successful,
> .Fn posix_fallocate
> returns zero.
> -It returns -1 on failure, and sets
> +It returns error on failure, without setting
> .Va errno
> -to indicate the error.
> +variable.
> .Sh ERRORS
> Possible failure conditions:
> .Bl -tag -width Er
> diff --git a/sys/compat/freebsd32/freebsd32_misc.c b/sys/compat/freebsd32/freebsd32_misc.c
> index 719a057..e4ffbe4 100644
> --- a/sys/compat/freebsd32/freebsd32_misc.c
> +++ b/sys/compat/freebsd32/freebsd32_misc.c
> @@ -2995,8 +2995,9 @@ freebsd32_posix_fallocate(struct thread *td,
>     struct freebsd32_posix_fallocate_args *uap)
> {
> 
> -	return (kern_posix_fallocate(td, uap->fd,
> -	    PAIR32TO64(off_t, uap->offset), PAIR32TO64(off_t, uap->len)));
> +	td->td_retval[0] = kern_posix_fallocate(td, uap->fd,
> +	    PAIR32TO64(off_t, uap->offset), PAIR32TO64(off_t, uap->len));
> +	return (0);
> }
> 
> int
> diff --git a/sys/kern/vfs_syscalls.c b/sys/kern/vfs_syscalls.c
> index dbad1ae..b864c90 100644
> --- a/sys/kern/vfs_syscalls.c
> +++ b/sys/kern/vfs_syscalls.c
> @@ -4584,7 +4584,9 @@ int
> sys_posix_fallocate(struct thread *td, struct posix_fallocate_args *uap)
> {
> 
> -	return (kern_posix_fallocate(td, uap->fd, uap->offset, uap->len));
> +	td->td_retval[0] = kern_posix_fallocate(td, uap->fd, uap->offset,
> +	    uap->len);
> +	return (0);
> }
> 
> /*


--
Bob Bishop          +44 (0)118 940 1243
rb at gid.co.uk    fax +44 (0)118 940 1295
             mobile +44 (0)783 626 4518