This is the mail archive of the
mailing list for the glibc project.
Re: [PATCH] Define GEN_AS_CONST_HEADERS when generating header files [BZ #22792]
On 02/21/2018 06:19 PM, H.J. Lu wrote:
On Wed, Feb 21, 2018 at 7:40 AM, H.J. Lu <email@example.com> wrote:
On Tue, Feb 20, 2018 at 11:09 PM, Florian Weimer <firstname.lastname@example.org> wrote:
On 02/08/2018 06:58 PM, H.J. Lu wrote:
* Makerules ($(common-objpfx)%.h): Pass -DGEN_AS_CONST_HEADERS
* sysdeps/unix/sysv/linux/i386/lowlevellock.h: Include
<tcb-offsets.h> only if GEN_AS_CONST_HEADERS isn't defined.
* sysdeps/unix/sysv/linux/x86_64/lowlevellock.h: Don't include
It looks to me that a cleaner fix would be to move the definition of the TCB
types to a separate header.
I will take a look.
RESULT offsetof (struct pthread, result)
TID offsetof (struct pthread, tid)
CANCELHANDLING offsetof (struct pthread, cancelhandling)
CLEANUP_JMP_BUF offsetof (struct pthread, cleanup_jmp_buf)
MULTIPLE_THREADS_OFFSET offsetof (tcbhead_t, multiple_threads)
SYSINFO_OFFSET offsetof (tcbhead_t, sysinfo)
CLEANUP offsetof (struct pthread, cleanup)
CLEANUP_PREV offsetof (struct _pthread_cleanup_buffer, __prev)
MUTEX_FUTEX offsetof (pthread_mutex_t, __data.__lock)
POINTER_GUARD offsetof (tcbhead_t, pointer_guard)
PRIVATE_FUTEX offsetof (tcbhead_t, private_futex)
It covers more than just tcbhead_t. A separate header file for tcbhead_t
won't help here.
Yes, it's more involved.
The circular dependency is just a cosmetic warning from make, right?
It depends on how unlucky you are. In one case on a many-core machine
with make -j60, I had to kill glibc build since it never stopped.
If it fixes an actual build issue, then I think your patch is
acceptable, although I consider it quite hackish.
But please do add a comment explaining the issue before:
+# ifndef GEN_AS_CONST_HEADERS