public inbox for glibc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug ports/26932] New: sh: Multiple floating point functions defined as stubs only since 2.31
@ 2020-11-23  1:45 glaubitz at physik dot fu-berlin.de
  2020-11-23 14:56 ` [Bug ports/26932] " glaubitz at physik dot fu-berlin.de
                   ` (7 more replies)
  0 siblings, 8 replies; 9+ messages in thread
From: glaubitz at physik dot fu-berlin.de @ 2020-11-23  1:45 UTC (permalink / raw)
  To: glibc-bugs

https://sourceware.org/bugzilla/show_bug.cgi?id=26932

            Bug ID: 26932
           Summary: sh: Multiple floating point functions defined as stubs
                    only since 2.31
           Product: glibc
           Version: 2.31
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: ports
          Assignee: unassigned at sourceware dot org
          Reporter: glaubitz at physik dot fu-berlin.de
                CC: adhemerval.zanella at linaro dot org, aurelien at aurel32 dot net,
                    carlos at redhat dot com,
                    sourceware-bugzilla at mkarcher dot dialup.fu-berlin.de
  Target Milestone: ---

Recently firebird3.0 started to fail to build from source on sh4 [1].

After some debugging, it turned out that the problem is that with glibc 2.31,
multiple floating point functions, including fegetenv() are no longer available
and are hence defined as __stub_ macros.

The complete list of missing functions can be seen by diffing the stubs.h
header:

--- stubs-2.30.h        2020-03-20 14:24:22.000000000 +0100
+++ stubs-2.31.h        2020-10-19 16:37:57.000000000 +0200
@@ -12,10 +12,25 @@
 #define __stub___compat_query_module
 #define __stub_chflags
 #define __stub_fchflags
+#define __stub_feclearexcept
+#define __stub_fedisableexcept
+#define __stub_feenableexcept
+#define __stub_fegetenv
+#define __stub_fegetexcept
+#define __stub_fegetexceptflag
+#define __stub_fegetmode
+#define __stub_fegetround
+#define __stub_feholdexcept
+#define __stub_feraiseexcept
+#define __stub_fesetenv
+#define __stub_fesetexcept
+#define __stub_fesetexceptflag
+#define __stub_fesetmode
+#define __stub_fesetround
+#define __stub_fetestexcept
+#define __stub_feupdateenv
 #define __stub_gtty
 #define __stub_lchmod
-#define __stub_pkey_alloc
-#define __stub_pkey_free
 #define __stub_revoke
 #define __stub_setlogin
 #define __stub_sigreturn

Bisecting led me to this commit as the culprit:

commit 605f38177dba31fe0d0183abb67e25a28fd37d15 (refs/bisect/bad)
Author: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Date:   Tue Jul 23 17:25:57 2019 -0300

    sh: Split BE/LE abilist

    The fix for BZ#18231 requires new symbols only for sh4eb.  This patch
    adds the required folder and files for both BE and LE abilist.  No
    semantic changes are expected.

    Checked with check-abi for sh4eb-linux-gnu and sh4-linux-gnu.

            * sysdeps/sh/preconfigure.ac: New file.
            * sysdeps/sh/preconfigure: Regenerate.
            * sysdeps/sh/be/sh3/Implies: New file.
            * sysdeps/sh/be/sh4/Implies: Likewise.
            * sysdeps/sh/le/sh3/Implies: Likewise.
            * sysdeps/sh/le/sh4/Implies: Likewise.
            * sysdeps/unix/sysv/linux/sh/le/sh3/Implies: Likewise.
            * sysdeps/unix/sysv/linux/sh/le/sh4/Implies: Likewise.
            * sysdeps/unix/sysv/linux/sh/*.abilist: Move to
            sysdeps/unix/sysv/linux/sh/le/*.abilist.
            * sysdeps/unix/sysv/linux/sh/be/*.abilist: New files.

It looks like the files sysdeps/sh/{le,be}/Implies are missing, I added them
with the content "sh" but that didn't help but chances are I did something
wrong as I don't fully understand the mechanism.

Since this affects 2.31 and higher, it would be nice if the fix could get
backported to the 2.31 branch.

> [1] https://buildd.debian.org/status/fetch.php?pkg=firebird3.0&arch=sh4&ver=3.0.7.33374.ds4-1&stamp=1603219363&raw=0

-- 
You are receiving this mail because:
You are on the CC list for the bug.

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Bug ports/26932] sh: Multiple floating point functions defined as stubs only since 2.31
  2020-11-23  1:45 [Bug ports/26932] New: sh: Multiple floating point functions defined as stubs only since 2.31 glaubitz at physik dot fu-berlin.de
@ 2020-11-23 14:56 ` glaubitz at physik dot fu-berlin.de
  2020-11-23 15:09 ` [Bug libc/26932] " schwab@linux-m68k.org
                   ` (6 subsequent siblings)
  7 siblings, 0 replies; 9+ messages in thread
From: glaubitz at physik dot fu-berlin.de @ 2020-11-23 14:56 UTC (permalink / raw)
  To: glibc-bugs

https://sourceware.org/bugzilla/show_bug.cgi?id=26932

--- Comment #1 from John Paul Adrian Glaubitz <glaubitz at physik dot fu-berlin.de> ---
I tried the following patch but that didn't help, unfortunately:

diff --git a/sysdeps/sh/be/Implies b/sysdeps/sh/be/Implies
new file mode 100644
index 0000000000..b001cf72a3
--- /dev/null
+++ b/sysdeps/sh/be/Implies
@@ -0,0 +1 @@
+sh
diff --git a/sysdeps/sh/le/Implies b/sysdeps/sh/le/Implies
new file mode 100644
index 0000000000..b001cf72a3
--- /dev/null
+++ b/sysdeps/sh/le/Implies
@@ -0,0 +1 @@
+sh
diff --git a/sysdeps/unix/sysv/linux/sh/be/Implies
b/sysdeps/unix/sysv/linux/sh/be/Implies
new file mode 100644
index 0000000000..1744ed820b
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/sh/be/Implies
@@ -0,0 +1 @@
+unix/sysv/linux/sh
diff --git a/sysdeps/unix/sysv/linux/sh/le/Implies
b/sysdeps/unix/sysv/linux/sh/le/Implies
new file mode 100644
index 0000000000..1744ed820b
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/sh/le/Implies
@@ -0,0 +1 @@
+unix/sysv/linux/sh

-- 
You are receiving this mail because:
You are on the CC list for the bug.

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Bug libc/26932] sh: Multiple floating point functions defined as stubs only since 2.31
  2020-11-23  1:45 [Bug ports/26932] New: sh: Multiple floating point functions defined as stubs only since 2.31 glaubitz at physik dot fu-berlin.de
  2020-11-23 14:56 ` [Bug ports/26932] " glaubitz at physik dot fu-berlin.de
@ 2020-11-23 15:09 ` schwab@linux-m68k.org
  2020-11-23 15:15 ` glaubitz at physik dot fu-berlin.de
                   ` (5 subsequent siblings)
  7 siblings, 0 replies; 9+ messages in thread
From: schwab@linux-m68k.org @ 2020-11-23 15:09 UTC (permalink / raw)
  To: glibc-bugs

https://sourceware.org/bugzilla/show_bug.cgi?id=26932

Andreas Schwab <schwab@linux-m68k.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |drepper.fsp at gmail dot com
          Component|ports                       |libc

--- Comment #2 from Andreas Schwab <schwab@linux-m68k.org> ---
Make sure your compiler defines __SH_FPU_ANY__.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Bug libc/26932] sh: Multiple floating point functions defined as stubs only since 2.31
  2020-11-23  1:45 [Bug ports/26932] New: sh: Multiple floating point functions defined as stubs only since 2.31 glaubitz at physik dot fu-berlin.de
  2020-11-23 14:56 ` [Bug ports/26932] " glaubitz at physik dot fu-berlin.de
  2020-11-23 15:09 ` [Bug libc/26932] " schwab@linux-m68k.org
@ 2020-11-23 15:15 ` glaubitz at physik dot fu-berlin.de
  2020-11-25 17:58 ` adhemerval.zanella at linaro dot org
                   ` (4 subsequent siblings)
  7 siblings, 0 replies; 9+ messages in thread
From: glaubitz at physik dot fu-berlin.de @ 2020-11-23 15:15 UTC (permalink / raw)
  To: glibc-bugs

https://sourceware.org/bugzilla/show_bug.cgi?id=26932

--- Comment #3 from John Paul Adrian Glaubitz <glaubitz at physik dot fu-berlin.de> ---
(In reply to Andreas Schwab from comment #2)
> Make sure your compiler defines __SH_FPU_ANY__.

It does:

root@tirpitz:~> echo | gcc -E -dM - | grep FPU
#define __SH_FPU_DOUBLE__ 1
#define __SH_FPU_ANY__ 1
root@tirpitz:~>

-- 
You are receiving this mail because:
You are on the CC list for the bug.

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Bug libc/26932] sh: Multiple floating point functions defined as stubs only since 2.31
  2020-11-23  1:45 [Bug ports/26932] New: sh: Multiple floating point functions defined as stubs only since 2.31 glaubitz at physik dot fu-berlin.de
                   ` (2 preceding siblings ...)
  2020-11-23 15:15 ` glaubitz at physik dot fu-berlin.de
@ 2020-11-25 17:58 ` adhemerval.zanella at linaro dot org
  2020-11-26 21:26 ` glaubitz at physik dot fu-berlin.de
                   ` (3 subsequent siblings)
  7 siblings, 0 replies; 9+ messages in thread
From: adhemerval.zanella at linaro dot org @ 2020-11-25 17:58 UTC (permalink / raw)
  To: glibc-bugs

https://sourceware.org/bugzilla/show_bug.cgi?id=26932

--- Comment #4 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
I forgot to handle the fpu support for SH4, since now le/be is handled as new
ABI for the sysnames loop in configure.ac.  I think this should fix it:

diff --git a/sysdeps/sh/be/sh4/fpu/Implies b/sysdeps/sh/be/sh4/fpu/Implies
new file mode 100644
index 0000000000..71b28ee1a4
--- /dev/null
+++ b/sysdeps/sh/be/sh4/fpu/Implies
@@ -0,0 +1 @@
+sh/sh4/fpu
diff --git a/sysdeps/sh/le/sh4/fpu/Implies b/sysdeps/sh/le/sh4/fpu/Implies
new file mode 100644
index 0000000000..71b28ee1a4
--- /dev/null
+++ b/sysdeps/sh/le/sh4/fpu/Implies
@@ -0,0 +1 @@
+sh/sh4/fpu
diff --git a/sysdeps/unix/sysv/linux/sh/be/sh4/fpu/Implies
b/sysdeps/unix/sysv/linux/sh/be/sh4/fpu/Implies
new file mode 100644
index 0000000000..7eeaf15a5a
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/sh/be/sh4/fpu/Implies
@@ -0,0 +1 @@
+unix/sysv/linux/sh/sh4/fpu
diff --git a/sysdeps/unix/sysv/linux/sh/le/sh4/fpu/Implies
b/sysdeps/unix/sysv/linux/sh/le/sh4/fpu/Implies
new file mode 100644
index 0000000000..7eeaf15a5a
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/sh/le/sh4/fpu/Implies
@@ -0,0 +1 @@
+unix/sysv/linux/sh/sh4/fpu

-- 
You are receiving this mail because:
You are on the CC list for the bug.

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Bug libc/26932] sh: Multiple floating point functions defined as stubs only since 2.31
  2020-11-23  1:45 [Bug ports/26932] New: sh: Multiple floating point functions defined as stubs only since 2.31 glaubitz at physik dot fu-berlin.de
                   ` (3 preceding siblings ...)
  2020-11-25 17:58 ` adhemerval.zanella at linaro dot org
@ 2020-11-26 21:26 ` glaubitz at physik dot fu-berlin.de
  2020-11-26 21:27 ` glaubitz at physik dot fu-berlin.de
                   ` (2 subsequent siblings)
  7 siblings, 0 replies; 9+ messages in thread
From: glaubitz at physik dot fu-berlin.de @ 2020-11-26 21:26 UTC (permalink / raw)
  To: glibc-bugs

https://sourceware.org/bugzilla/show_bug.cgi?id=26932

--- Comment #5 from John Paul Adrian Glaubitz <glaubitz at physik dot fu-berlin.de> ---
(In reply to Adhemerval Zanella from comment #4)
> I forgot to handle the fpu support for SH4, since now le/be is handled as
> new ABI for the sysnames loop in configure.ac.  I think this should fix it:

I can confirm that this fixes the issue for me. The firebird3.0 package builds
successfully on sh4 again.

Thanks!

-- 
You are receiving this mail because:
You are on the CC list for the bug.

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Bug libc/26932] sh: Multiple floating point functions defined as stubs only since 2.31
  2020-11-23  1:45 [Bug ports/26932] New: sh: Multiple floating point functions defined as stubs only since 2.31 glaubitz at physik dot fu-berlin.de
                   ` (4 preceding siblings ...)
  2020-11-26 21:26 ` glaubitz at physik dot fu-berlin.de
@ 2020-11-26 21:27 ` glaubitz at physik dot fu-berlin.de
  2020-11-27 12:32 ` adhemerval.zanella at linaro dot org
  2020-11-27 12:33 ` adhemerval.zanella at linaro dot org
  7 siblings, 0 replies; 9+ messages in thread
From: glaubitz at physik dot fu-berlin.de @ 2020-11-26 21:27 UTC (permalink / raw)
  To: glibc-bugs

https://sourceware.org/bugzilla/show_bug.cgi?id=26932

--- Comment #6 from John Paul Adrian Glaubitz <glaubitz at physik dot fu-berlin.de> ---
PS: Please don't forget to backport to 2.31 - if possible - as Debian is still
on 2.31.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Bug libc/26932] sh: Multiple floating point functions defined as stubs only since 2.31
  2020-11-23  1:45 [Bug ports/26932] New: sh: Multiple floating point functions defined as stubs only since 2.31 glaubitz at physik dot fu-berlin.de
                   ` (5 preceding siblings ...)
  2020-11-26 21:27 ` glaubitz at physik dot fu-berlin.de
@ 2020-11-27 12:32 ` adhemerval.zanella at linaro dot org
  2020-11-27 12:33 ` adhemerval.zanella at linaro dot org
  7 siblings, 0 replies; 9+ messages in thread
From: adhemerval.zanella at linaro dot org @ 2020-11-27 12:32 UTC (permalink / raw)
  To: glibc-bugs

https://sourceware.org/bugzilla/show_bug.cgi?id=26932

Adhemerval Zanella <adhemerval.zanella at linaro dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|unassigned at sourceware dot org   |adhemerval.zanella at linaro dot o
                   |                            |rg
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|---                         |FIXED
   Target Milestone|---                         |2.23

--- Comment #7 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
Fixed on 2.33.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Bug libc/26932] sh: Multiple floating point functions defined as stubs only since 2.31
  2020-11-23  1:45 [Bug ports/26932] New: sh: Multiple floating point functions defined as stubs only since 2.31 glaubitz at physik dot fu-berlin.de
                   ` (6 preceding siblings ...)
  2020-11-27 12:32 ` adhemerval.zanella at linaro dot org
@ 2020-11-27 12:33 ` adhemerval.zanella at linaro dot org
  7 siblings, 0 replies; 9+ messages in thread
From: adhemerval.zanella at linaro dot org @ 2020-11-27 12:33 UTC (permalink / raw)
  To: glibc-bugs

https://sourceware.org/bugzilla/show_bug.cgi?id=26932

Adhemerval Zanella <adhemerval.zanella at linaro dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|2.23                        |2.33

-- 
You are receiving this mail because:
You are on the CC list for the bug.

^ permalink raw reply	[flat|nested] 9+ messages in thread

end of thread, other threads:[~2020-11-27 12:33 UTC | newest]

Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-11-23  1:45 [Bug ports/26932] New: sh: Multiple floating point functions defined as stubs only since 2.31 glaubitz at physik dot fu-berlin.de
2020-11-23 14:56 ` [Bug ports/26932] " glaubitz at physik dot fu-berlin.de
2020-11-23 15:09 ` [Bug libc/26932] " schwab@linux-m68k.org
2020-11-23 15:15 ` glaubitz at physik dot fu-berlin.de
2020-11-25 17:58 ` adhemerval.zanella at linaro dot org
2020-11-26 21:26 ` glaubitz at physik dot fu-berlin.de
2020-11-26 21:27 ` glaubitz at physik dot fu-berlin.de
2020-11-27 12:32 ` adhemerval.zanella at linaro dot org
2020-11-27 12:33 ` adhemerval.zanella at linaro dot org

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).