Tony Luck wrote: > But I still can't get sys_mkdirat() > to make a directory ... I just get a regular file. Does normal mkdir() work? If yes, it a problem in the way you wire the syscalls up. sys_mkdir is just calling sys_mkdirat. Anyway, hold off on submitting code for the *at syscalls for a bit longer. I just found another problem. Not a bug, but we must expose the stat64 data instead of stat on 32bit platforms. I completely missed this in my patch. The kernel's stat-related code is in almost a bad of a shape as glibc's. -- ➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA ❖
Attachment:
signature.asc
Description: OpenPGP digital signature
- References:
- Re: [PATCH 1/3] NEW VERSION: *at syscalls: Implementation
- From: Tony Luck <[email protected]>
- Re: [PATCH 1/3] NEW VERSION: *at syscalls: Implementation
- From: Tony Luck <[email protected]>
- Re: [PATCH 1/3] NEW VERSION: *at syscalls: Implementation
- Prev by Date: Re: [PATCH] stop ==== emergency
- Next by Date: Re: [PATCH] Prevent spinlock debug from timing out too early
- Previous by thread: Re: [PATCH 1/3] NEW VERSION: *at syscalls: Implementation
- Next by thread: Re: [PATCH 1/3] NEW VERSION: *at syscalls: Implementation
- Index(es):