[ISSUE] `su` segfaults

[ISSUE] `su` segfaults

[balejk]

[-- Attachment #1: Type: text/plain, Size: 790 bytes --]

New issue by balejk on void-packages repository

https://github.com/void-linux/void-packages/issues/48758

Description:
# Is this a new report?

Yes

# System Info

`Void 6.7.4_1 x86_64-musl GenuineIntel uptodate FFFFFFFF`

with `dracut` and `sudo` ignored.

# Package(s) Affected

`util-linux-2.38.1_6`

# Does a report exist for this bug with the project's home (upstream) and/or another distro?

No response

# Expected behaviour

`su -` opens a root shell when correct password is entered.

# Actual behaviour

`su` segfaults.

# Steps to reproduce

Update the system, run `su -`, enter correct password.

---

Downgrading to `util-linux-2.38.1_5` solves the problem. Since the latest revision of the package only changed the manpages, the problem is likely not in the package itself.

Re: `su` segfaults

[classabbyamp]

[-- Attachment #1: Type: text/plain, Size: 191 bytes --]

New comment by classabbyamp on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1948190565

Comment:
can you run it in gdb and get a backtrace?

Re: `su` segfaults

[sgn]

[-- Attachment #1: Type: text/plain, Size: 232 bytes --]

New comment by sgn on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1948606880

Comment:
Can you run gdb on the corefile and post the backtrace. I don't think `gdb` works with `su`.

Re: `su` segfaults

[balejk]

[-- Attachment #1: Type: text/plain, Size: 468 bytes --]

New comment by balejk on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1949104026

Comment:
`# gdb --args su - user` does segfault but `bt` says "No stack". `gdb`
also prints "Detaching after fork from child process <PID>" -- could
that be the reason for the missing backtrace?

Unless I'm doing something wrong (`# ulimit -c unlimited; su - user`,
works for other binaries), there seems to be no core dumped either.

Re: `su` segfaults

[sgn]

[-- Attachment #1: Type: text/plain, Size: 164 bytes --]

New comment by sgn on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1950980499

Comment:
Because `su` is `setuid`

Re: `su` segfaults

[catinello]

[-- Attachment #1: Type: text/plain, Size: 460 bytes --]

New comment by catinello on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951035636

Comment:
Experiencing almost the same.

```
su
Password:
su: failed to execute /bin/bash: Bad address
```
This happens independently of the used shell (bash/dash). 


If I'm root and try to call a specific command for a user via su I get a segmentation fault.

```
su -c /bin/sh - USER
Segmentation fault
```

Re: `su` segfaults

[catinello]

[-- Attachment #1: Type: text/plain, Size: 505 bytes --]

New comment by catinello on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951035636

Comment:
Experiencing almost the same.

```
su
Password:
su: failed to execute /bin/bash: Bad address
```
This happens independently of the used shell (bash/dash). 


If I'm root and try to call a specific command for a user via su I get a segmentation fault.

```
su -c /bin/sh - USER
Segmentation fault
```

System is musl / x86_64 with all updates.

Re: `su` segfaults

[sgn]

[-- Attachment #1: Type: text/plain, Size: 290 bytes --]

New comment by sgn on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951334729

Comment:
```
failed to execute /bin/bash: Bad address
```

This is broken during `fork(2)` some invalid addresses was used. No corefile would be generated.

Re: `su` segfaults

[sgn]

[-- Attachment #1: Type: text/plain, Size: 333 bytes --]

New comment by sgn on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951340641

Comment:
```
850     in login-utils/su-common.c
(gdb) p args[0]
$4 = 0x555660e5 <error: Cannot access memory at address 0x555660e5>
```

A classic 64 bits pointer truncated to 32 bit int, I think.

Re: [ISSUE] [CLOSED] `su` segfaults

[sgn]

[-- Attachment #1: Type: text/plain, Size: 793 bytes --]

Closed issue by balejk on void-packages repository

https://github.com/void-linux/void-packages/issues/48758

Description:
# Is this a new report?

Yes

# System Info

`Void 6.7.4_1 x86_64-musl GenuineIntel uptodate FFFFFFFF`

with `dracut` and `sudo` ignored.

# Package(s) Affected

`util-linux-2.38.1_6`

# Does a report exist for this bug with the project's home (upstream) and/or another distro?

No response

# Expected behaviour

`su -` opens a root shell when correct password is entered.

# Actual behaviour

`su` segfaults.

# Steps to reproduce

Update the system, run `su -`, enter correct password.

---

Downgrading to `util-linux-2.38.1_5` solves the problem. Since the latest revision of the package only changed the manpages, the problem is likely not in the package itself.

Re: `su` segfaults

[balejk]

[-- Attachment #1: Type: text/plain, Size: 312 bytes --]

New comment by balejk on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951343930

Comment:
> Experiencing almost the same.
>
> ```
> su
> Password:
> su: failed to execute /bin/bash: Bad address
> ```

I can confirm that this is the same thing as I am seeing.

Re: `su` segfaults

[sgn]

[-- Attachment #1: Type: text/plain, Size: 182 bytes --]

New comment by sgn on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951344277

Comment:
Fixed, please update and test in a moment.

Re: `su` segfaults

[sgn]

[-- Attachment #1: Type: text/plain, Size: 190 bytes --]

New comment by sgn on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951344555

Comment:
https://github.com/util-linux/util-linux/pull/2615

Re: `su` segfaults

[sgn]

[-- Attachment #1: Type: text/plain, Size: 292 bytes --]

New comment by sgn on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951334729

Comment:
```
failed to execute /bin/bash: Bad address
```

This is broken during `execve(2)` some invalid addresses was used. No corefile would be generated.

Re: `su` segfaults

[sgn]

[-- Attachment #1: Type: text/plain, Size: 242 bytes --]

New comment by sgn on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951345143

Comment:
The problem, as shown, is in the package itself, the old revision was compiled by different compilers.

Re: `su` segfaults

[balejk]

[-- Attachment #1: Type: text/plain, Size: 233 bytes --]

New comment by balejk on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951346151

Comment:
> Fixed, please update and test in a moment.

The fix seems to work, thank you very much.

Re: `su` segfaults

[oreo639]

[-- Attachment #1: Type: text/plain, Size: 340 bytes --]

New comment by oreo639 on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951459179

Comment:
Sorry for not noticing this issue earlier.
Btw for the future, such issues can be mitigated with `XBPS_CFLAGS="-O2 -pipe -Werror=implicit-function-declaration -Werror=implicit-int"` until gcc 14.

Re: `su` segfaults

[oreo639]

[-- Attachment #1: Type: text/plain, Size: 339 bytes --]

New comment by oreo639 on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951459179

Comment:
Sorry for not noticing the error earlier.
Btw for the future, such issues can be mitigated with `XBPS_CFLAGS="-O2 -pipe -Werror=implicit-function-declaration -Werror=implicit-int"` until gcc 14.

Re: `su` segfaults

[oreo639]

[-- Attachment #1: Type: text/plain, Size: 374 bytes --]

New comment by oreo639 on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951459179

Comment:
Sorry for not noticing the error earlier.
Btw, for the future, such issues can be mitigated with `XBPS_CFLAGS="-O2 -pipe -Werror=implicit-function-declaration -Werror=implicit-int"` until gcc 14 (where this becomes a hard error).

Re: `su` segfaults

[oreo639]

[-- Attachment #1: Type: text/plain, Size: 381 bytes --]

New comment by oreo639 on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951459179

Comment:
Sorry for not noticing the error earlier.
Btw, for the future, such issues can be mitigated with `XBPS_CFLAGS="-O2 -pipe -Werror=implicit-function-declaration -Werror=implicit-int"` until gcc 14 (where this becomes an error by default).

Re: `su` segfaults

[oreo639]

[-- Attachment #1: Type: text/plain, Size: 381 bytes --]

New comment by oreo639 on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951459179

Comment:
Sorry for not catching the error earlier.
Btw, for the future, such issues can be mitigated with `XBPS_CFLAGS="-O2 -pipe -Werror=implicit-function-declaration -Werror=implicit-int"` until gcc 14 (where this becomes an error by default).

Re: `su` segfaults

[oreo639]

[-- Attachment #1: Type: text/plain, Size: 406 bytes --]

New comment by oreo639 on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951459179

Comment:
Sorry for not catching the error earlier.
Btw, for the future, such issues can be mitigated with `XBPS_CFLAGS="-O2 -pipe -Werror=implicit-function-declaration -Werror=implicit-int"` until gcc 14 (where, without `-fpermissive`, this becomes an error by default).

Re: `su` segfaults

[sgn]

[-- Attachment #1: Type: text/plain, Size: 301 bytes --]

New comment by sgn on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951736730

Comment:
For `-Werror=implicit-function-declaration -Werror=implicit-int`

https://www.mail-archive.com/autoconf@gnu.org/msg24713.html
https://bugs.gentoo.org/870412

Re: `su` segfaults

[sgn]

[-- Attachment #1: Type: text/plain, Size: 337 bytes --]

New comment by sgn on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951740377

Comment:
Thus, I think in order to have  `-Werror=implicit-function-declaration -Werror=implicit-int`, we need:
- Update autoconf to latest version
- `gnu-configure`'s `do_configure` need `autoreconf -fi`

Re: `su` segfaults

[sgn]

[-- Attachment #1: Type: text/plain, Size: 225 bytes --]

New comment by sgn on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951741171

Comment:
But, even that, tons of autoconf script out there is their own m4 versions. :shrug:

Re: `su` segfaults

[oreo639]

[-- Attachment #1: Type: text/plain, Size: 601 bytes --]

New comment by oreo639 on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951752022

Comment:
Either way, will have to deal with it in the near future. I'll look in to it more later.

(and yeah, a lot of applications will need their tests updated for ` -Werror=implicit-function-declaration -Werror=implicit-int`, including for example gimp. Fedora is doing a lot of work with testing, previously with ` -Werror=implicit-function-declaration -Werror=implicit-int` and currently with gcc 14. https://fedoraproject.org/wiki/Toolchain/PortingToModernC)

Re: `su` segfaults

[oreo639]

[-- Attachment #1: Type: text/plain, Size: 599 bytes --]

New comment by oreo639 on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951752022

Comment:
Either way, will have to deal with it in the near future. I'll look in to it more later.

(and yeah, a lot of applications will need their tests updated for ` -Werror=implicit-function-declaration -Werror=implicit-int`, including for example gimp. Fedora is doing a lot of work with testing, previously with ` -Werror=implicit-function-declaration -Werror=implicit-int` and currently with gcc 14. https://fedoraproject.org/wiki/Changes/PortingToModernC)

Re: `su` segfaults

[oreo639]

[-- Attachment #1: Type: text/plain, Size: 711 bytes --]

New comment by oreo639 on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951752022

Comment:
Either way, will have to deal with it in the near future. I'll look in to it more later.

(and yeah, a lot of applications will need their tests updated for ` -Werror=implicit-function-declaration -Werror=implicit-int`, including for example gimp. Fedora is doing a lot of work with testing, previously with ` -Werror=implicit-function-declaration -Werror=implicit-int` and currently with gcc 14. https://fedoraproject.org/wiki/Changes/PortingToModernC)

Most stuff should still compile though (only actual compile error I noticed that isn't a bug was `virtuoso`)

Re: `su` segfaults

[oreo639]

[-- Attachment #1: Type: text/plain, Size: 694 bytes --]

New comment by oreo639 on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951752022

Comment:
Either way, will have to deal with it in the near future. I'll look in to it more later.

(and yeah, a lot of applications will need their tests updated for ` -Werror=implicit-function-declaration -Werror=implicit-int`, including for example gimp. Fedora is doing a lot of work with testing, previously with ` -Werror=implicit-function-declaration -Werror=implicit-int` and currently with gcc 14. https://fedoraproject.org/wiki/Changes/PortingToModernC)

Most stuff should still compile though (only actual compile error I noticed was `virtuoso`)

Re: `su` segfaults

[oreo639]

[-- Attachment #1: Type: text/plain, Size: 726 bytes --]

New comment by oreo639 on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951752022

Comment:
Either way, will have to deal with it in the near future. I'll look in to it more later.

(and yeah, a lot of applications will need their tests updated for ` -Werror=implicit-function-declaration -Werror=implicit-int`, including for example gimp. Fedora is doing a lot of work with testing, previously with ` -Werror=implicit-function-declaration -Werror=implicit-int` and currently with gcc 14. https://fedoraproject.org/wiki/Changes/PortingToModernC)

Most stuff should still compile though (only actual compile error I noticed, outside of the basename stuff, was `virtuoso`)

Re: `su` segfaults

[oreo639]

[-- Attachment #1: Type: text/plain, Size: 757 bytes --]

New comment by oreo639 on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951752022

Comment:
Either way, will have to deal with it in the near future. I'll look in to it more later.

(and yeah, a lot of applications will need their tests updated for ` -Werror=implicit-function-declaration -Werror=implicit-int`, including for example gimp. Fedora is doing a lot of work with testing, previously with ` -Werror=implicit-function-declaration -Werror=implicit-int` and currently with gcc 14. https://fedoraproject.org/wiki/Changes/PortingToModernC)

Most stuff should still compile though (only actual compile error I noticed, outside of the basename stuff, was `virtuoso`, which would need `-Wno-error`)

Re: `su` segfaults

[oreo639]

[-- Attachment #1: Type: text/plain, Size: 761 bytes --]

New comment by oreo639 on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951752022

Comment:
Either way, will have to deal with it in the near future. I'll look in to it more later.

(and yeah, a lot of applications will need their tests updated for ` -Werror=implicit-function-declaration -Werror=implicit-int`, including for example gimp. Fedora is doing a lot of work with testing, previously with ` -Werror=implicit-function-declaration -Werror=implicit-int` and currently with gcc 14. https://fedoraproject.org/wiki/Changes/PortingToModernC)

Most stuff should still compile though (only actual compile error I noticed, outside of the basename stuff, was `virtuoso`, which would need `-Wno-error=...`)

Re: `su` segfaults

[oreo639]

[-- Attachment #1: Type: text/plain, Size: 753 bytes --]

New comment by oreo639 on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951752022

Comment:
Either way, will have to deal with it eventually. I'll look in to it more later.

(and yeah, a lot of applications will need their tests updated for ` -Werror=implicit-function-declaration -Werror=implicit-int`, including for example gimp. Fedora is doing a lot of work with testing, previously with ` -Werror=implicit-function-declaration -Werror=implicit-int` and currently with gcc 14. https://fedoraproject.org/wiki/Changes/PortingToModernC)

Most stuff should still compile though (only actual compile error I noticed, outside of the basename stuff, was `virtuoso`, which would need `-Wno-error=...`)

Re: `su` segfaults

[oreo639]

[-- Attachment #1: Type: text/plain, Size: 761 bytes --]

New comment by oreo639 on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951752022

Comment:
Either way, will have to deal with it eventually. I'll look in to it more later. Thanks.

(and yeah, a lot of applications will need their tests updated for ` -Werror=implicit-function-declaration -Werror=implicit-int`, including for example gimp. Fedora is doing a lot of work with testing, previously with ` -Werror=implicit-function-declaration -Werror=implicit-int` and currently with gcc 14. https://fedoraproject.org/wiki/Changes/PortingToModernC)

Most stuff should still compile though (only actual compile error I noticed, outside of the basename stuff, was `virtuoso`, which would need `-Wno-error=...`)

Re: `su` segfaults

[oreo639]

[-- Attachment #1: Type: text/plain, Size: 607 bytes --]

New comment by oreo639 on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951752022

Comment:
Either way, will have to deal with it eventually. I'll look in to it more later. Thanks.

(and yeah, a lot of applications will need their tests updated for ` -Werror=implicit-function-declaration -Werror=implicit-int`, including for example gimp and dia. Fedora is doing a lot of work with testing, previously with ` -Werror=implicit-function-declaration -Werror=implicit-int` and currently with gcc 14. https://fedoraproject.org/wiki/Changes/PortingToModernC)

Re: `su` segfaults

[oreo639]

[-- Attachment #1: Type: text/plain, Size: 650 bytes --]

New comment by oreo639 on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951752022

Comment:
Either way, will have to deal with it eventually. I'll look in to it more later. Thanks.

(and yeah, a lot of applications will need their tests updated for ` -Werror=implicit-function-declaration -Werror=implicit-int`, including for example gimp and dia. Fedora is doing a lot of work with testing, previously with ` -Werror=implicit-function-declaration -Werror=implicit-int` and currently with gcc 14. https://fedoraproject.org/wiki/Changes/PortingToModernC https://gcc.gnu.org/gcc-14/porting_to.html)

Re: `su` segfaults

[oreo639]

[-- Attachment #1: Type: text/plain, Size: 708 bytes --]

New comment by oreo639 on void-packages repository

https://github.com/void-linux/void-packages/issues/48758#issuecomment-1951752022

Comment:
Either way, will have to deal with it eventually. I'll look in to it more later. Thanks.

(and yeah, a lot of applications will need their tests updated for ` -Werror=implicit-function-declaration -Werror=implicit-int`, including for example gimp and dia. Fedora is doing a lot of work with testing, previously with ` -Werror=implicit-function-declaration -Werror=implicit-int -Werror=incompatible-pointer-types -Werror=int-conversion` and currently with gcc 14. https://fedoraproject.org/wiki/Changes/PortingToModernC https://gcc.gnu.org/gcc-14/porting_to.html)