diff options
| author | Flavio Cruz <flaviocruz@gmail.com> | 2023-12-14 01:02:31 -0500 |
|---|---|---|
| committer | Samuel Thibault <samuel.thibault@ens-lyon.org> | 2023-12-17 23:17:27 +0100 |
| commit | 69620634858b2992e1a362e33c95d9a8ee57bce7 (patch) | |
| tree | bd9e8647ec2ee236166301e347557323d10d69c0 /i386 | |
| parent | c5f150653fd0642fdbd11b043dfbbbe473604570 (diff) | |
x86_64: Support 8 byte inlined port rights to avoid message resizing.
If a port is inlined in a message, the user has to use
mach_port_name_inlined_t to define each port. Out of line memory
continues to use mach_port_name_t since that memory has to be copied to
the kernel anyway.
Both copyinmsg and copyoutmsg can be reduced to nothing (if we ignore
USER32) as a follow up but kept this patch simple for ease of review.
Diffstat (limited to 'i386')
| -rw-r--r-- | i386/i386/copy_user.h | 12 |
1 files changed, 5 insertions, 7 deletions
diff --git a/i386/i386/copy_user.h b/i386/i386/copy_user.h index 5cdbfa80..3d1c7278 100644 --- a/i386/i386/copy_user.h +++ b/i386/i386/copy_user.h @@ -87,16 +87,14 @@ static inline int copyout_port(const mach_port_t *kaddr, mach_port_name_t *uaddr #endif /* __x86_64__ */ } -// XXX we could add another field to kmsg to store the user-side size, but then we -// should check if we can obtain it for rpc and notifications originating from -// the kernel -#ifndef __x86_64__ +#if defined(__x86_64__) && defined(USER32) +/* For 32 bit userland, kernel and user land messages are not the same size. */ +size_t msg_usize(const mach_msg_header_t *kmsg); +#else static inline size_t msg_usize(const mach_msg_header_t *kmsg) { return kmsg->msgh_size; } -#else /* __x86_64__ */ -size_t msg_usize(const mach_msg_header_t *kmsg); -#endif /* __x86_64__ */ +#endif /* __x86_64__ && USER32 */ #endif /* COPY_USER_H */ |