This is the mail archive of the
mailing list for the glibc project.
Re: [PATCH] aarch64: Fix ipc_perm definition for ILP32
- From: Szabolcs Nagy <szabolcs dot nagy at arm dot com>
- To: sellcey at cavium dot com, libc-alpha <libc-alpha at sourceware dot org>
- Cc: nd at arm dot com, Yury Norov <ynorov at caviumnetworks dot com>
- Date: Wed, 23 Aug 2017 10:55:02 +0100
- Subject: Re: [PATCH] aarch64: Fix ipc_perm definition for ILP32
- Authentication-results: sourceware.org; auth=none
- Authentication-results: spf=none (sender IP is ) smtp.mailfrom=Szabolcs dot Nagy at arm dot com;
- Nodisclaimer: True
- References: <firstname.lastname@example.org>
- Spamdiagnosticmetadata: NSPM
- Spamdiagnosticoutput: 1:99
On 22/08/17 18:46, Steve Ellcey wrote:
> Here is another aarch64 ILP32 patch. The mode field in ipc_perm in ILP32
> should be a 16 bit field, not a 32 bit one. This was out-of-sync with what the
> kernel had. This was causing sysvipc/test-sysvsem to fail in ILP32 mode.
> This change is only needed for ILP32 so it doesn't need to go in until we add
> that ABI but I am sending out for review and comments.
> 2017-08-22 Yury Norov <email@example.com>
> Steve Ellcey <firstname.lastname@example.org>
> * sysdeps/unix/sysv/linux/aarch64/bits/ipc.h (ipc_perm):
> Ifdef and pad the mode field for ILP32.
> diff --git a/sysdeps/unix/sysv/linux/aarch64/bits/ipc.h b/sysdeps/unix/sysv/linu
> index cd1f06e..cd05b74 100644
> --- a/sysdeps/unix/sysv/linux/aarch64/bits/ipc.h
> +++ b/sysdeps/unix/sysv/linux/aarch64/bits/ipc.h
> @@ -46,7 +46,12 @@ struct ipc_perm
> __gid_t gid; /* Owner's group ID. */
> __uid_t cuid; /* Creator's user ID. */
> __gid_t cgid; /* Creator's group ID. */
> +#ifdef __LP64
> unsigned int mode; /* Read/write permission. */
> + unsigned short int mode; /* Read/write permission. */
> + unsigned short int __pad0;
when did this happen?
as far as i can tell staging/ilp32-4.12 branch in
still has unsigned int __kernel_mode_t in uapi, so this is
an abi change compared to that branch.
i guess it's for 32bit compat, but i'd like to see the
kernel patch for this.