Skip Menu |
 

History Show all quoted text — Show full headers
# Fri Feb 25 17:24:52 2005 "Arlene Berry" <aberry0364@hotmail.com> - Ticket created
From: "Arlene Berry" <aberry0364@hotmail.com>
To: krb5-bugs@mit.edu
Date: Fri, 25 Feb 2005 20:40:27 +0000
Subject: building krb5-1.4 on Solaris with native compiler
Download (untitled) / with headers
text/plain 794B
When building krb5-1.4 on Solaris with our native compiler I get numerous
warnings of the form:

/opt/SUNWspro/bin/../SC4.0/bin/fbe: "/tmp/yabeAAAEfaG1U", line 1082:
warning: label in delay slot (follows CTI)

Fbe is the assempler component of the compiler and I have never seen this
warning before. I traced the cause to some of the k5_mutex macros exceeding
the complexity level tolerated by our compiler. I converted the following
four macros in k5-thread.h to static inline functions and the warnings went
away:

k5_pthread_assert_locked
k5_os_mutex_unlock
k5_mutex_destroy
k5_mutex_unlock

Show quoted text
_________________________________________________________________
FREE pop-up blocking with the new MSN Toolbar – get it now!
http://toolbar.msn.click-url.com/go/onm00200415ave/direct/01/
# Fri Feb 25 19:15:37 2005 tlyu (Taylor Yu) - Correspondence added
To: rt@krbdev.mit.edu
Subject: Re: [krbdev.mit.edu #2945] building krb5-1.4 on Solaris with native compiler
From: Tom Yu <tlyu@MIT.EDU>
Date: Fri, 25 Feb 2005 19:15:27 -0500
RT-Send-Cc:
Download (untitled) / with headers
text/plain 418B
Show quoted text
>>>>> "Arlene" == "Arlene Berry" via RT <rt-comment@krbdev.mit.edu> writes:

Show quoted text
Arlene> /opt/SUNWspro/bin/../SC4.0/bin/fbe: "/tmp/yabeAAAEfaG1U", line 1082:
Arlene> warning: label in delay slot (follows CTI)

Note that SC4.0 is a very old version of the Sun compiler. I would
argue that this is a bug in the compiler. Have you tried using
different optimization levels to see if one of them makes the warning
go away?
# Fri Feb 25 19:30:45 2005 tlyu (Taylor Yu) - Status changed from 'new' to 'open'
# Fri Feb 25 19:30:45 2005 tlyu (Taylor Yu) - Given to tlyu (Taylor Yu)
# Fri Feb 25 19:30:46 2005 tlyu (Taylor Yu) - Component krb5-build added
# Fri Feb 25 19:30:46 2005 tlyu (Taylor Yu) - Version_reported 1.4 added
# Mon Feb 28 17:09:39 2005 "Arlene Berry" <aberry0364@hotmail.com> - Comments added
From: "Arlene Berry" <aberry0364@hotmail.com>
To: rt-comment@krbdev.mit.edu
Subject: Re: [krbdev.mit.edu #2945] building krb5-1.4 on Solaris with native compiler
Date: Mon, 28 Feb 2005 22:08:57 +0000
RT-Send-Cc:
Download (untitled) / with headers
text/plain 1.1KiB


Show quoted text
>From: "Tom Yu via RT" <rt-comment@krbdev.mit.edu>
>Reply-To: rt-comment@krbdev.mit.edu
>To: aberry0364@hotmail.com
>Subject: Re: [krbdev.mit.edu #2945] building krb5-1.4 on Solaris with
>native compiler
>Date: Fri, 25 Feb 2005 19:15:39 -0500 (EST)
>
> >>>>> "Arlene" == "Arlene Berry" via RT <rt-comment@krbdev.mit.edu>
>writes:
>
>Arlene> /opt/SUNWspro/bin/../SC4.0/bin/fbe: "/tmp/yabeAAAEfaG1U", line
>1082:
>Arlene> warning: label in delay slot (follows CTI)
>
>Note that SC4.0 is a very old version of the Sun compiler. I would
>argue that this is a bug in the compiler. Have you tried using
>different optimization levels to see if one of them makes the warning
>go away?
>

I had not tried that. -xO2 makes the warning go away. I had thought we set
that for our own projects but have found that we do not along with an old
comment that states optimization breaks the DES implementation in our code.
Some simple kinit tests using DES as well as other algorithms worked so the
build is probably okay.

Show quoted text
_________________________________________________________________
FREE pop-up blocking with the new MSN Toolbar – get it now!
http://toolbar.msn.click-url.com/go/onm00200415ave/direct/01/
# Tue Mar 22 23:48:26 2005 tlyu (Taylor Yu) - Status changed from 'open' to 'resolved'
# Tue Mar 22 23:48:27 2005 tlyu (Taylor Yu) - Tags nochange added
# Tue Mar 22 23:48:27 2005 tlyu (Taylor Yu) - Correspondence added
Download (untitled) / with headers
text/plain 74B
Closing this ticket, since it's almost certainly a compiler optimizer bug.