On Wed, Mar 06, 2024 at 12:40:54PM +0000, Colin Ian King wrote:
> While exercising hfsplus with stress-ng with xattr tests the kernel
> log was spammed with many "xattr exists yet" messages. The error
> EOPNOTSUPP is returned, so the need to emit these error messages is
> not necessary; removing them reduces kernel error spamming.
Isn't that the wrong errno though? EOPNOTSUPP isn't listed as an errno
in the fsetxattr manpage. ENOTSUP should be used for "xattrs are not
supported". But this condition looks like EEXIST to me.
> Signed-off-by: Colin Ian King <colin.i.king@xxxxxxxxx>
> ---
> fs/hfsplus/xattr.c | 2 --
> 1 file changed, 2 deletions(-)
>
> diff --git a/fs/hfsplus/xattr.c b/fs/hfsplus/xattr.c
> index 9c9ff6b8c6f7..57101524fff4 100644
> --- a/fs/hfsplus/xattr.c
> +++ b/fs/hfsplus/xattr.c
> @@ -288,7 +288,6 @@ int __hfsplus_setxattr(struct inode *inode, const char *name,
>
> if (!strcmp_xattr_finder_info(name)) {
> if (flags & XATTR_CREATE) {
> - pr_err("xattr exists yet\n");
> err = -EOPNOTSUPP;
> goto end_setxattr;
> }
> @@ -335,7 +334,6 @@ int __hfsplus_setxattr(struct inode *inode, const char *name,
>
> if (hfsplus_attr_exists(inode, name)) {
> if (flags & XATTR_CREATE) {
> - pr_err("xattr exists yet\n");
> err = -EOPNOTSUPP;
> goto end_setxattr;
> }
> --
> 2.39.2
>
>
