diff mbox series

[v3,1/2] Gold/MIPS: Improve MIPS support in configure.tgt

Message ID 20230820171457.1377429-1-yunqiang.su@cipunited.com
State Accepted
Headers
Series [v3,1/2] Gold/MIPS: Improve MIPS support in configure.tgt |

Checks

Context Check Description
snail/binutils-gdb-check success Github commit url

Commit Message

YunQiang Su Aug. 20, 2023, 5:14 p.m. UTC 
  1. Drop mips*le triple pattern, which will consider as big endian
   by other components of binutils.
2. Use EM_MIPS instead of EM_MIPS_RS3_LE for little endian.
   EM_MIPS_RS3_LE has been deprecated quite long ago, and in fact
   most of current LE ELF files are using EM_MIPS.
3. Add targ_extra_size=64 for mips32 triples.
4. Add targ_extra_little_endian=true for big endian triples.
5. Add mips64 triples support, and define targ_extra_little_endian/
   targ_extra_big_endian, and targ_extra_size=32.

Let's use `targ_extra_little_endian` in configure.ac/configure,
so that littlen endian can be enabled if the targets are all big endian.
---
 gold/configure     |  8 ++++++++
 gold/configure.ac  |  8 ++++++++
 gold/configure.tgt | 24 +++++++++++++++++++++---
 3 files changed, 37 insertions(+), 3 deletions(-)

Comments

YunQiang Su Aug. 25, 2023, 2:57 a.m. UTC | #1 
@Maciej W. Rozycki How about this patch?

@Nick Clifton Is there any policy of backport policy of 2.41?
If OK, I'd like to backport this patch to 2.41. As it is really a bug:
   ../configure --enable-targets=mipsel-linux-gnu only will fail to build.

See:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1049455
and
https://buildd.debian.org/status/fetch.php?pkg=binutils-mipsen&arch=amd64&ver=10%2Bc5&stamp=1692086940&raw=0
Maciej W. Rozycki Aug. 25, 2023, 3:31 a.m. UTC | #2 
On Fri, 25 Aug 2023, YunQiang Su wrote:

> @Maciej W. Rozycki How about this patch?

 This will have to be split as I previously outlined.  Also I won't have 
any time to review patches in the next few weeks as I am out of capacity 
right now.

  Maciej
YunQiang Su Aug. 29, 2023, 10:49 p.m. UTC | #3 
Maciej W. Rozycki <macro@orcam.me.uk> 于2023年8月25日周五 11:31写道：
>
> On Fri, 25 Aug 2023, YunQiang Su wrote:
>
> > @Maciej W. Rozycki How about this patch?
>
>  This will have to be split as I previously outlined.  Also I won't have

Generally, we should split patches by code logic, instead of make it fragments.
For distrubutions, cherry-pick patches from master is needed usally.
Fragmental patches will make lots of trobule to them.

We should have a trade off, instead of be extrame.

> any time to review patches in the next few weeks as I am out of capacity
> right now.
>
>   Maciej
diff mbox series

Patch

diff --git a/gold/configure b/gold/configure
index 5c6fe2a5607..ccb1f2269b2 100755
--- a/gold/configure
+++ b/gold/configure
@@ -5266,6 +5266,10 @@  for targ in $target $canon_targets; do
 	      -o "$targ_extra_big_endian" = "true"; then
 	      targ_32_big=yes
 	  fi
+	  if test "$targ_little_endian" = "true" \
+	      -o "$targ_extra_little_endian" = "true"; then
+	      targ_32_little=yes
+	  fi
 	  if test "$targ_big_endian" = "false" \
 	      -o "$targ_extra_big_endian" = "false"; then
 	      targ_32_little=yes
@@ -5276,6 +5280,10 @@  for targ in $target $canon_targets; do
 	      -o "$targ_extra_big_endian" = "true"; then
 	      targ_64_big=yes
 	  fi
+	  if test "$targ_little_endian" = "true" \
+	      -o "$targ_extra_little_endian" = "true"; then
+	      targ_64_little=yes
+	  fi
 	  if test "$targ_big_endian" = "false" \
 	      -o "$targ_extra_big_endian" = "false"; then
 	      targ_64_little=yes
diff --git a/gold/configure.ac b/gold/configure.ac
index cafd3503b3b..e04e4086140 100644
--- a/gold/configure.ac
+++ b/gold/configure.ac
@@ -193,6 +193,10 @@  for targ in $target $canon_targets; do
 	      -o "$targ_extra_big_endian" = "true"; then
 	      targ_32_big=yes
 	  fi
+	  if test "$targ_little_endian" = "true" \
+	      -o "$targ_extra_little_endian" = "true"; then
+	      targ_32_little=yes
+	  fi
 	  if test "$targ_big_endian" = "false" \
 	      -o "$targ_extra_big_endian" = "false"; then
 	      targ_32_little=yes
@@ -203,6 +207,10 @@  for targ in $target $canon_targets; do
 	      -o "$targ_extra_big_endian" = "true"; then
 	      targ_64_big=yes
 	  fi
+	  if test "$targ_little_endian" = "true" \
+	      -o "$targ_extra_little_endian" = "true"; then
+	      targ_64_little=yes
+	  fi
 	  if test "$targ_big_endian" = "false" \
 	      -o "$targ_extra_big_endian" = "false"; then
 	      targ_64_little=yes
diff --git a/gold/configure.tgt b/gold/configure.tgt
index 4b54e08d27f..e2e131bbd84 100644
--- a/gold/configure.tgt
+++ b/gold/configure.tgt
@@ -153,19 +153,37 @@  aarch64*-*)
  targ_big_endian=false
  targ_extra_big_endian=true
  ;;
-mips*el*-*-*|mips*le*-*-*)
+mips64*el-*-*)
  targ_obj=mips
- targ_machine=EM_MIPS_RS3_LE
+ targ_machine=EM_MIPS
+ targ_size=64
+ targ_extra_size=32
+ targ_big_endian=false
+ targ_extra_big_endian=true
+ ;;
+mips*el-*-*)
+ targ_obj=mips
+ targ_machine=EM_MIPS
  targ_size=32
+ targ_extra_size=64
  targ_big_endian=false
  targ_extra_big_endian=true
  ;;
+mips64*-*-*)
+ targ_obj=mips
+ targ_machine=EM_MIPS
+ targ_size=64
+ targ_extra_size=32
+ targ_big_endian=true
+ targ_extra_little_endian=true
+ ;;
 mips*-*-*)
  targ_obj=mips
  targ_machine=EM_MIPS
  targ_size=32
+ targ_extra_size=64
  targ_big_endian=true
- targ_extra_big_endian=false
+ targ_extra_little_endian=true
  ;;
 s390-*-*)
  targ_obj=s390