On (09/10/08 20:20), Andy Whitcroft didst pronounce:
> From: Adam Litke <[EMAIL PROTECTED]>
>
> As originally written, the functions hugetlbfs_find_path_for_size() and
> hugetlbfs_unlinked_fd_for_size() accepted 0 as a valid page size. A page size
> of zero was interpreted to mean the default size (as defined by libhugetlbfs,
> not the kernel).
>
> This behavior is undesirable in pagesize-specific functions for two reasons.
> First, the libhugetlbfs default size can be nebulous since it depends on what
> page sizes are available, which sizes have appropriate mount points, and the
> value of an environment variable. This amount of variability adds uncertainty
> to what should be an explicit API. Second, the functions
> hugetlbfs_find_path()
> and hugetlbfs_unlinked_fd() already exist as a means to delegate page size
> selection to the library. It is my view that all functions that take a page
> size paramater shall expect a valid page size.
>
> Signed-off-by: Adam Litke <[EMAIL PROTECTED]>
Acked-by: Mel Gorman <[EMAIL PROTECTED]>
> ---
> hugeutils.c | 22 +++++++++++-----------
> 1 files changed, 11 insertions(+), 11 deletions(-)
>
> diff --git a/hugeutils.c b/hugeutils.c
> index 7e730aa..724feb8 100644
> --- a/hugeutils.c
> +++ b/hugeutils.c
> @@ -241,10 +241,6 @@ int select_pool_counter(unsigned int counter, unsigned
> long pagesize,
> return -1;
> }
>
> - /* Convert a pagesize of 0 to the libhugetlbfs default size */
> - if (pagesize == 0)
> - pagesize = default_size;
> -
> /* If the user is dealing in the default page size, we can use /proc */
> if (pagesize == default_size) {
> if (meminfo_key && key) {
> @@ -515,11 +511,7 @@ const char *hugetlbfs_find_path_for_size(long page_size)
> char *path;
> int idx;
>
> - if (page_size == 0)
> - idx = hpage_sizes_default_idx;
> - else
> - idx = hpage_size_to_index(page_size);
> -
> + idx = hpage_size_to_index(page_size);
> if (idx >= 0) {
> path = hpage_sizes[idx].mount;
> if (strlen(path))
> @@ -530,7 +522,11 @@ const char *hugetlbfs_find_path_for_size(long page_size)
>
> const char *hugetlbfs_find_path(void)
> {
> - return hugetlbfs_find_path_for_size(0);
> + long hpage_size = gethugepagesize();
> + if (hpage_size > 0)
> + return hugetlbfs_find_path_for_size(hpage_size);
> + else
> + return NULL;
> }
>
> int hugetlbfs_unlinked_fd_for_size(long page_size)
> @@ -563,7 +559,11 @@ int hugetlbfs_unlinked_fd_for_size(long page_size)
>
> int hugetlbfs_unlinked_fd(void)
> {
> - return hugetlbfs_unlinked_fd_for_size(0);
> + long hpage_size = gethugepagesize();
> + if (hpage_size > 0)
> + return hugetlbfs_unlinked_fd_for_size(hpage_size);
> + else
> + return -1;
> }
>
> #define IOV_LEN 64
> --
> 1.6.0.1.451.gc8d31
>
--
Mel Gorman
Part-time Phd Student Linux Technology Center
University of Limerick IBM Dublin Software Lab
-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Libhugetlbfs-devel mailing list
Libhugetlbfs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libhugetlbfs-devel
