[PATCH v2] linux-user/syscall: Implement execve without execveat

Pierrick Bouvier posted 1 patch 9 months, 4 weeks ago
Patches applied successfully (tree, apply log)
git fetch https://github.com/patchew-project/qemu tags/patchew/20230705090056.610699-1-pierrick.bouvier@linaro.org
Maintainers: Laurent Vivier <laurent@vivier.eu>
There is a newer version of this series
linux-user/syscall.c | 20 ++++++++++++--------
1 file changed, 12 insertions(+), 8 deletions(-)
[PATCH v2] linux-user/syscall: Implement execve without execveat
Posted by Pierrick Bouvier 9 months, 4 weeks ago
Support for execveat syscall was implemented in 55bbe4 and is available
since QEMU 8.0.0. It relies on host execveat, which is widely available
on most of Linux kernels today.

However, this change breaks qemu-user self emulation, if "host" qemu
version is less than 8.0.0. Indeed, it does not implement yet execveat.
This strange use case happens with most of distribution today having
binfmt support.

With a concrete failing example:
$ qemu-x86_64-7.2 qemu-x86_64-8.0 /bin/bash -c /bin/ls
/bin/bash: line 1: /bin/ls: Function not implemented
-> not implemented means execve returned ENOSYS

qemu-user-static 7.2 and 8.0 can be conveniently grabbed from debian
packages qemu-user-static* [1].

One usage of this is running wine-arm64 from linux-x64 (details [2]).
This is by updating qemu embedded in docker image that we ran into this
issue.

The solution to update host qemu is not always possible. Either it's
complicated or ask you to recompile it, or simply is not accessible
(GitLab CI, GitHub Actions). Thus, it could be worth to implement execve
without relying on execveat, which is the goal of this patch.

This patch was tested with example presented in this commit message.

[1] http://ftp.us.debian.org/debian/pool/main/q/qemu/
[1] https://www.linaro.org/blog/emulate-windows-on-arm/

Signed-off-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
---
 linux-user/syscall.c | 20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)

diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index 08162cc966..4945ddd7f2 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -659,6 +659,7 @@ safe_syscall4(pid_t, wait4, pid_t, pid, int *, status, int, options, \
 #endif
 safe_syscall5(int, waitid, idtype_t, idtype, id_t, id, siginfo_t *, infop, \
               int, options, struct rusage *, rusage)
+safe_syscall3(int, execve, const char *, filename, char **, argv, char **, envp)
 safe_syscall5(int, execveat, int, dirfd, const char *, filename,
               char **, argv, char **, envp, int, flags)
 #if defined(TARGET_NR_select) || defined(TARGET_NR__newselect) || \
@@ -8615,9 +8616,9 @@ ssize_t do_guest_readlink(const char *pathname, char *buf, size_t bufsiz)
     return ret;
 }
 
-static int do_execveat(CPUArchState *cpu_env, int dirfd,
-                       abi_long pathname, abi_long guest_argp,
-                       abi_long guest_envp, int flags)
+static int do_execv(CPUArchState *cpu_env, int dirfd,
+                    abi_long pathname, abi_long guest_argp,
+                    abi_long guest_envp, int flags, bool is_execveat)
 {
     int ret;
     char **argp, **envp;
@@ -8696,11 +8697,14 @@ static int do_execveat(CPUArchState *cpu_env, int dirfd,
         goto execve_efault;
     }
 
+    const char* exe = p;
     if (is_proc_myself(p, "exe")) {
-        ret = get_errno(safe_execveat(dirfd, exec_path, argp, envp, flags));
-    } else {
-        ret = get_errno(safe_execveat(dirfd, p, argp, envp, flags));
+        exe = exec_path;
     }
+    ret = is_execveat ?
+        safe_execveat(dirfd, exe, argp, envp, flags):
+        safe_execve(exe, argp, envp);
+    ret = get_errno(ret);
 
     unlock_user(p, pathname, 0);
 
@@ -9251,9 +9255,9 @@ static abi_long do_syscall1(CPUArchState *cpu_env, int num, abi_long arg1,
         return ret;
 #endif
     case TARGET_NR_execveat:
-        return do_execveat(cpu_env, arg1, arg2, arg3, arg4, arg5);
+        return do_execv(cpu_env, arg1, arg2, arg3, arg4, arg5, true);
     case TARGET_NR_execve:
-        return do_execveat(cpu_env, AT_FDCWD, arg1, arg2, arg3, 0);
+        return do_execv(cpu_env, AT_FDCWD, arg1, arg2, arg3, 0, false);
     case TARGET_NR_chdir:
         if (!(p = lock_user_string(arg1)))
             return -TARGET_EFAULT;
-- 
2.40.1
Re: [PATCH v2] linux-user/syscall: Implement execve without execveat
Posted by Michael Tokarev 9 months, 4 weeks ago
05.07.2023 12:00, Pierrick Bouvier wrote:
...
> diff --git a/linux-user/syscall.c b/linux-user/syscall.c
> index 08162cc966..4945ddd7f2 100644
> --- a/linux-user/syscall.c
> +++ b/linux-user/syscall.c
> @@ -659,6 +659,7 @@ safe_syscall4(pid_t, wait4, pid_t, pid, int *, status, int, options, \
>   #endif
>   safe_syscall5(int, waitid, idtype_t, idtype, id_t, id, siginfo_t *, infop, \
>                 int, options, struct rusage *, rusage)
> +safe_syscall3(int, execve, const char *, filename, char **, argv, char **, envp)
>   safe_syscall5(int, execveat, int, dirfd, const char *, filename,
>                 char **, argv, char **, envp, int, flags)
>   #if defined(TARGET_NR_select) || defined(TARGET_NR__newselect) || \
> @@ -8615,9 +8616,9 @@ ssize_t do_guest_readlink(const char *pathname, char *buf, size_t bufsiz)
>       return ret;
>   }
>   
> -static int do_execveat(CPUArchState *cpu_env, int dirfd,
> -                       abi_long pathname, abi_long guest_argp,
> -                       abi_long guest_envp, int flags)
> +static int do_execv(CPUArchState *cpu_env, int dirfd,
> +                    abi_long pathname, abi_long guest_argp,
> +                    abi_long guest_envp, int flags, bool is_execveat)
>   {
>       int ret;
>       char **argp, **envp;
> @@ -8696,11 +8697,14 @@ static int do_execveat(CPUArchState *cpu_env, int dirfd,
>           goto execve_efault;
>       }
>   
> +    const char* exe = p;
>       if (is_proc_myself(p, "exe")) {
> -        ret = get_errno(safe_execveat(dirfd, exec_path, argp, envp, flags));
> -    } else {
> -        ret = get_errno(safe_execveat(dirfd, p, argp, envp, flags));
> +        exe = exec_path;
>       }
> +    ret = is_execveat ?
> +        safe_execveat(dirfd, exe, argp, envp, flags):
> +        safe_execve(exe, argp, envp);
> +    ret = get_errno(ret);


This one has 2 issues reported by checkpatch.pl:

$ git show | ./scripts/checkpatch.pl -

ERROR: "foo* bar" should be "foo *bar"
#161: FILE: linux-user/syscall.c:8700:
+    const char* exe = p;

ERROR: spaces required around that ':' (ctx:VxE)
#169: FILE: linux-user/syscall.c:8705:
+        safe_execveat(dirfd, exe, argp, envp, flags):
                                                      ^

total: 2 errors, 0 warnings, 47 lines checked

Your patch has style problems, please review.  If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.

As I mentioned in the v1, I don't remember offhand how the arithmetic if
should be styled in qemu.  At the very best, the v2 variant is difficult
to read because ":" is too close to ";" visually, an extra space before
it will make it more explicit.

Other than that, I'm fine with this version.

With the checkpatch issues fixed,

Reviewed-by: Michael Tokarev <mjt@tls.msk.ru>

Thanks,

/mjt
Re: [PATCH v2] linux-user/syscall: Implement execve without execveat
Posted by Philippe Mathieu-Daudé 9 months, 4 weeks ago
On 5/7/23 12:26, Michael Tokarev wrote:
> 05.07.2023 12:00, Pierrick Bouvier wrote:
> ...


>> @@ -8696,11 +8697,14 @@ static int do_execveat(CPUArchState *cpu_env, 
>> int dirfd,
>>           goto execve_efault;
>>       }
>> +    const char* exe = p;
>>       if (is_proc_myself(p, "exe")) {
>> -        ret = get_errno(safe_execveat(dirfd, exec_path, argp, envp, 
>> flags));
>> -    } else {
>> -        ret = get_errno(safe_execveat(dirfd, p, argp, envp, flags));
>> +        exe = exec_path;
>>       }
>> +    ret = is_execveat ?
>> +        safe_execveat(dirfd, exe, argp, envp, flags):
>> +        safe_execve(exe, argp, envp);
>> +    ret = get_errno(ret);


> ERROR: spaces required around that ':' (ctx:VxE)
> #169: FILE: linux-user/syscall.c:8705:
> +        safe_execveat(dirfd, exe, argp, envp, flags):
>                                                       ^
> 
> total: 2 errors, 0 warnings, 47 lines checked
> 
> Your patch has style problems, please review.  If any of these errors
> are false positives report them to the maintainer, see
> CHECKPATCH in MAINTAINERS.
> 
> As I mentioned in the v1, I don't remember offhand how the arithmetic if
> should be styled in qemu.  At the very best, the v2 variant is difficult
> to read because ":" is too close to ";" visually, an extra space before
> it will make it more explicit.

KISS, alternatively:

     ret = is_execveat
           ? safe_execveat(dirfd, exe, argp, envp, flags)
           : safe_execve(exe, argp, envp);


Re: [PATCH v2] linux-user/syscall: Implement execve without execveat
Posted by Pierrick Bouvier 9 months, 4 weeks ago
On 7/5/23 12:26, Michael Tokarev wrote:
> 05.07.2023 12:00, Pierrick Bouvier wrote:
> ...
>> diff --git a/linux-user/syscall.c b/linux-user/syscall.c
>> index 08162cc966..4945ddd7f2 100644
>> --- a/linux-user/syscall.c
>> +++ b/linux-user/syscall.c
>> @@ -659,6 +659,7 @@ safe_syscall4(pid_t, wait4, pid_t, pid, int *, status, int, options, \
>>    #endif
>>    safe_syscall5(int, waitid, idtype_t, idtype, id_t, id, siginfo_t *, infop, \
>>                  int, options, struct rusage *, rusage)
>> +safe_syscall3(int, execve, const char *, filename, char **, argv, char **, envp)
>>    safe_syscall5(int, execveat, int, dirfd, const char *, filename,
>>                  char **, argv, char **, envp, int, flags)
>>    #if defined(TARGET_NR_select) || defined(TARGET_NR__newselect) || \
>> @@ -8615,9 +8616,9 @@ ssize_t do_guest_readlink(const char *pathname, char *buf, size_t bufsiz)
>>        return ret;
>>    }
>>    
>> -static int do_execveat(CPUArchState *cpu_env, int dirfd,
>> -                       abi_long pathname, abi_long guest_argp,
>> -                       abi_long guest_envp, int flags)
>> +static int do_execv(CPUArchState *cpu_env, int dirfd,
>> +                    abi_long pathname, abi_long guest_argp,
>> +                    abi_long guest_envp, int flags, bool is_execveat)
>>    {
>>        int ret;
>>        char **argp, **envp;
>> @@ -8696,11 +8697,14 @@ static int do_execveat(CPUArchState *cpu_env, int dirfd,
>>            goto execve_efault;
>>        }
>>    
>> +    const char* exe = p;
>>        if (is_proc_myself(p, "exe")) {
>> -        ret = get_errno(safe_execveat(dirfd, exec_path, argp, envp, flags));
>> -    } else {
>> -        ret = get_errno(safe_execveat(dirfd, p, argp, envp, flags));
>> +        exe = exec_path;
>>        }
>> +    ret = is_execveat ?
>> +        safe_execveat(dirfd, exe, argp, envp, flags):
>> +        safe_execve(exe, argp, envp);
>> +    ret = get_errno(ret);
> 
> 
> This one has 2 issues reported by checkpatch.pl:
> 
> $ git show | ./scripts/checkpatch.pl -
> 
> ERROR: "foo* bar" should be "foo *bar"
> #161: FILE: linux-user/syscall.c:8700:
> +    const char* exe = p;
> 
> ERROR: spaces required around that ':' (ctx:VxE)
> #169: FILE: linux-user/syscall.c:8705:
> +        safe_execveat(dirfd, exe, argp, envp, flags):
>                                                        ^
> 
> total: 2 errors, 0 warnings, 47 lines checked
> 
> Your patch has style problems, please review.  If any of these errors
> are false positives report them to the maintainer, see
> CHECKPATCH in MAINTAINERS.
> 
> As I mentioned in the v1, I don't remember offhand how the arithmetic if
> should be styled in qemu.  At the very best, the v2 variant is difficult
> to read because ":" is too close to ";" visually, an extra space before
> it will make it more explicit.
> 

Sorry about it.
I'll fix it, and use style advised by Philippe.

> Other than that, I'm fine with this version.
> 
> With the checkpatch issues fixed,
> 
> Reviewed-by: Michael Tokarev <mjt@tls.msk.ru>
> 
> Thanks,
> 
> /mjt