Skip to contentSkip to conversation

package/fakeroot: fix glibc detection on patch for new wrappers

PATCH Superseded Ryan Barnett 3 participants 4 messages Feb 21, 2021 – Feb 27, 2021

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 package/fakeroot: fix glibc detection on patch for new wrappers

Commit f45925a951318e9e53bead80b363e004301adc6f add the patch:

0003-libfakeroot.c-add-wrappers-for-new-glibc-2.33-symbol.patch

which allowed fakeroot to be compiled with GLIBC 2.33 or above. However, this introduce a bug for building with a non-GLIBC based toolchain as a GLIBC macro - __GLIBC_PREREQ - is used on the same line as the detection of GLIBC.

Fix this by backporting the fix to this incorrect macro from upstream commit:

https://salsa.debian.org/clint/fakeroot/-/commit/8090dffdad8fda86dccd47ce7a7db8840bdf7d7b

---
v1 -> v2:
- added patch as standalone patch (suggested by Yann)

---
 .../0005-fix-build-regression-on-macOS.patch  | 64 +++++++++++++++++++
 1 file changed, 64 insertions(+)
 create mode 100644 package/fakeroot/0005-fix-build-regression-on-macOS.patch
package/fakeroot/0005-fix-build-regression-on-macOS.patch new file +64 -0
new file mode 100644
index 0000000000..71152b850d
@@ -0,0 +1,64 @@
+From 3590b817df703a256f2c1de9a5f5469eaa1c86e9 Mon Sep 17 00:00:00 2001
+From: Ilya Lipnitskiy <[email protected]>
+Date: Mon, 15 Feb 2021 11:07:56 -0800
+Subject: [PATCH] fix build regression on macOS
+
+Backported from: 8090dffdad8fda86dccd47ce7a7db8840bdf7d7b
+
+Signed-off-by: Felix Fietkau <[email protected]>
+Signed-off-by: Ilya Lipnitskiy <[email protected]>
+Signed-off-by: Ryan Barnett <[email protected]>
+---
+ configure.ac  | 6 ++++++
+ libfakeroot.c | 4 +++-
+ 2 files changed, 9 insertions(+), 1 deletion(-)
+
+diff --git a/configure.ac b/configure.ac
+index d85566f..d635df1 100644
+--- a/configure.ac
++++ b/configure.ac
+@@ -181,6 +181,9 @@ AC_MSG_CHECKING([for type of arg of __xmknod])
+        #include <fcntl.h>
+        #include <unistd.h>
+   ]], [[
++#ifndef __GLIBC__
++#error no extra *
++#endif
+        int __xmknod  ( int ver,
+                        const char *pathname ,
+                        mode_t  mode ,  dev_t *dev);
+@@ -207,6 +210,9 @@ AC_MSG_CHECKING([for type of arg of __xmknodat])
+        #include <fcntl.h>
+        #include <unistd.h>
+   ]], [[
++#ifndef __GLIBC__
++#error no extra *
++#endif
+        int __xmknodat  ( int ver,
+                          int dirfd,
+                          const char *pathname ,
+diff --git a/libfakeroot.c b/libfakeroot.c
+index d75c51f..ec4e577 100644
+--- a/libfakeroot.c
++++ b/libfakeroot.c
+@@ -1352,7 +1352,8 @@ int renameat(int olddir_fd, const char *oldpath,
+ #endif /* HAVE_FSTATAT */
+ 
+ 
+-#if defined(__GLIBC__) && __GLIBC_PREREQ(2,33)
++#if defined(__GLIBC__)
++#if __GLIBC_PREREQ(2,33)
+ /* Glibc 2.33 exports symbols for these functions in the shared lib */
+   int lstat(const char *file_name, struct stat *statbuf) {
+      return WRAP_LSTAT LSTAT_ARG(_STAT_VER, file_name, statbuf);
+@@ -1397,6 +1398,7 @@ int renameat(int olddir_fd, const char *oldpath,
+        return WRAP_MKNODAT MKNODAT_ARG(_STAT_VER, dir_fd, pathname, mode, &dev);
+     }
+   #endif
++#endif /* __GLIBC__ */
+ #endif /* GLIBC_PREREQ */
+ 
+ 
+-- 
+2.25.1
+
Signature · 1 line
2.25.1
#1 lore raw

PATCH v2 1/1 package/fakeroot: fix glibc detection on patch for new wrappers

Ryan, All,

On 2021-02-21 16:38 -0600, Ryan Barnett spake thusly:

Quoted 9 lines

Commit f45925a951318e9e53bead80b363e004301adc6f add the patch:

0003-libfakeroot.c-add-wrappers-for-new-glibc-2.33-symbol.patch

which allowed fakeroot to be compiled with GLIBC 2.33 or above. However, this introduce a bug for building with a non-GLIBC based toolchain as a GLIBC macro - __GLIBC_PREREQ - is used on the same line as the detection of GLIBC.

Quoted 7 lines

Fix this by backporting the fix to this incorrect macro from upstream commit:

https://salsa.debian.org/clint/fakeroot/-/commit/8090dffdad8fda86dccd47ce7a7db8840bdf7d7b

Applied to master, thanks.

Regards,
Yann E. MORIN.

Quoted 85 lines

---
v1 -> v2:
- added patch as standalone patch (suggested by Yann)

---
 .../0005-fix-build-regression-on-macOS.patch  | 64 +++++++++++++++++++
 1 file changed, 64 insertions(+)
 create mode 100644 package/fakeroot/0005-fix-build-regression-on-macOS.patch
package/fakeroot/0005-fix-build-regression-on-macOS.patch new file +64 -0
new file mode 100644
index 0000000000..71152b850d
@@ -0,0 +1,64 @@
+From 3590b817df703a256f2c1de9a5f5469eaa1c86e9 Mon Sep 17 00:00:00 2001
+From: Ilya Lipnitskiy <[email protected]>
+Date: Mon, 15 Feb 2021 11:07:56 -0800
+Subject: [PATCH] fix build regression on macOS
+
+Backported from: 8090dffdad8fda86dccd47ce7a7db8840bdf7d7b
+
+Signed-off-by: Felix Fietkau <[email protected]>
+Signed-off-by: Ilya Lipnitskiy <[email protected]>
+Signed-off-by: Ryan Barnett <[email protected]>
+---
+ configure.ac  | 6 ++++++
+ libfakeroot.c | 4 +++-
+ 2 files changed, 9 insertions(+), 1 deletion(-)
+
+diff --git a/configure.ac b/configure.ac
+index d85566f..d635df1 100644
+--- a/configure.ac
++++ b/configure.ac
+@@ -181,6 +181,9 @@ AC_MSG_CHECKING([for type of arg of __xmknod])
+        #include <fcntl.h>
+        #include <unistd.h>
+   ]], [[
++#ifndef __GLIBC__
++#error no extra *
++#endif
+        int __xmknod  ( int ver,
+                        const char *pathname ,
+                        mode_t  mode ,  dev_t *dev);
+@@ -207,6 +210,9 @@ AC_MSG_CHECKING([for type of arg of __xmknodat])
+        #include <fcntl.h>
+        #include <unistd.h>
+   ]], [[
++#ifndef __GLIBC__
++#error no extra *
++#endif
+        int __xmknodat  ( int ver,
+                          int dirfd,
+                          const char *pathname ,
+diff --git a/libfakeroot.c b/libfakeroot.c
+index d75c51f..ec4e577 100644
+--- a/libfakeroot.c
++++ b/libfakeroot.c
+@@ -1352,7 +1352,8 @@ int renameat(int olddir_fd, const char *oldpath,
+ #endif /* HAVE_FSTATAT */
+ 
+ 
+-#if defined(__GLIBC__) && __GLIBC_PREREQ(2,33)
++#if defined(__GLIBC__)
++#if __GLIBC_PREREQ(2,33)
+ /* Glibc 2.33 exports symbols for these functions in the shared lib */
+   int lstat(const char *file_name, struct stat *statbuf) {
+      return WRAP_LSTAT LSTAT_ARG(_STAT_VER, file_name, statbuf);
+@@ -1397,6 +1398,7 @@ int renameat(int olddir_fd, const char *oldpath,
+        return WRAP_MKNODAT MKNODAT_ARG(_STAT_VER, dir_fd, pathname, mode, &dev);
+     }
+   #endif
++#endif /* __GLIBC__ */
+ #endif /* GLIBC_PREREQ */
+ 
+ 
+-- 
+2.25.1
+
Signature · 6 lines
2.25.1

_______________________________________________
buildroot mailing list
buildroot at busybox.net
http://lists.busybox.net/mailman/listinfo/buildroot
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 561 099 427 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'
#2 lore raw

PATCH v2 1/1 package/fakeroot: fix glibc detection on patch for new wrappers

In reply to Ryan Barnett
Quoted 13 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Ryan" == Ryan Barnett <[email protected]> writes:

Commit f45925a951318e9e53bead80b363e004301adc6f add the patch:
0003-libfakeroot.c-add-wrappers-for-new-glibc-2.33-symbol.patch

which allowed fakeroot to be compiled with GLIBC 2.33 or above. However, this introduce a bug for building with a non-GLIBC based toolchain as a GLIBC macro - __GLIBC_PREREQ - is used on the same line as the detection of GLIBC.

Fix this by backporting the fix to this incorrect macro from upstream commit:

Quoted 7 lines

https://salsa.debian.org/clint/fakeroot/-/commit/8090dffdad8fda86dccd47ce7a7db8840bdf7d7b

---
v1 -> v2:
- added patch as standalone patch (suggested by Yann)

Committed to 2020.11.x, thanks.

The patch does not apply to 2020.02.x. Is it needed there as well? Presumably yes - Care to send a patch?

Are you really running Buildroot under MacOS? Does that work?

Signature · 1 line
Bye, Peter Korsgaard
#3 lore raw

PATCH v2 1/1 package/fakeroot: fix glibc detection on patch for new wrappers

Peter,

Quoted 22 lines · Peter Korsgaard
Quoted 14 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

"Ryan" == Ryan Barnett <[email protected]> writes:

Commit f45925a951318e9e53bead80b363e004301adc6f add the patch:
0003-libfakeroot.c-add-wrappers-for-new-glibc-2.33-symbol.patch

which allowed fakeroot to be compiled with GLIBC 2.33 or above. However, this introduce a bug for building with a non-GLIBC based toolchain as a GLIBC macro - __GLIBC_PREREQ - is used on the same line as the detection of GLIBC.

Fix this by backporting the fix to this incorrect macro from upstream commit:

Quoted 7 lines

https://salsa.debian.org/clint/fakeroot/-/commit/8090dffdad8fda86dccd47ce7a7db8840bdf7d7b

---
v1 -> v2:
- added patch as standalone patch (suggested by Yann)

Quoted 4 lines

Committed to 2020.11.x, thanks.

The patch does not apply to 2020.02.x. Is it needed there as well? Presumably yes - Care to send a patch?

Yes it appears that this would apply to 2020.02.x since it is carry a patch to add support glibc 2.33:

https://git.busybox.net/buildroot/tree/package/fakeroot/0006-libfakeroot.c-add-wrappers-for-new-glibc-2.33-symbol.patch?h=2020.02.x

Quoted 1 line

Are you really running Buildroot under MacOS? Does that work?

I was looking at using buildroot on the bare minimum Alpine Linux Docker container to help with finding issues with missing host dependencies. Specifically looking at building on a non-glibc host machine since Alpine is a musl based distro. Eventually my goal is to have an autobuilder somewhere that actually builds with the minimum package list which is listed in the buildroot manual.

Hopefully soon I can send a patch series which describes what issues I've found and some patches for some missing host dependencies on some packages.

Thanks,
-Ryan