Re: [HACKERS] Current sources?

>> PPC/Linux has been broken too.
>
>Please let me know what the problem was, even if it was just the 'global tas'
>thing. I am trying to make sure this works on all platforms. Thanks.

Here are patches for s_lock.c (against May23 snapshot).
----------------------------------------------------------
*** s_lock.c.orig    Mon May 25 18:08:20 1998
--- s_lock.c    Mon May 25 18:08:57 1998
***************
*** 151,161 ****

  #if defined(PPC)

! static int
! tas_dummy()
  {
      __asm__("                \n\
- tas:                        \n\
              lwarx    5,0,3    \n\
              cmpwi    5,0        \n\
              bne        fail    \n\
--- 151,160 ----

  #if defined(PPC)

! int
! tas(slock_t *lock)
  {
      __asm__("                \n\
              lwarx    5,0,3    \n\
              cmpwi    5,0        \n\
              bne        fail    \n\
----------------------------------------------------------
>> I have tested the s_lock_test and seems it is working. However I have
>> lots of failure with various SQL's including 'drop database', 'delete
>> from'.
>> Have you succeeded in running regression tests? If so, what kind of
>> platforms are you using?
>
>I made this patch against 6.3.2 and ran regression successfully. This on a
>glibc Linux x86 system. I just rebuilt against the latest CVS (from anoncvs)
>and see 27 tests that fail, many with dropconns. I looked a little into the
>'drop database failure' and it does not look related to spinlocks as far as
>I looked.

I see. BTW, I have tested on FreeBSD box and found exactly same thing
has occured.
--
Tatsuo Ishii
t-ishii@sra.co.jp