[v2] gdb/source.c: Fix undefined behaviour dereferencing empty string

  When a source file's dirname is solely made up of directory separators
we end up trying to dereference the last character of an empty string
with std::string::back, which results in undefined behaviour. A typical
use case where this can happen is when the root directory "/" is used as
a compilation directory.

With libstdc++.so.6.0.28 we get no out-of-bounds checks and the byte
preceding the storage of the empty string is returned. The character
value of this byte depends on heap implementation and usage, but when
this byte happens to hold the value of the directory separator character
we go on to call std::string::pop_back on the empty string which results
in an out_of_range exception which terminates GDB.

Fix this by using path_join. prepare_path_for_appending ensures that the
filename component is relative.

The testsuite has been run before and after the change and no
regressions were found.
---
 gdb/source.c | 10 +---------
 1 file changed, 1 insertion(+), 9 deletions(-)
  

On 2022-09-21 10:58, Magne Hov via Gdb-patches wrote:
> When a source file's dirname is solely made up of directory separators
> we end up trying to dereference the last character of an empty string
> with std::string::back, which results in undefined behaviour. A typical
> use case where this can happen is when the root directory "/" is used as
> a compilation directory.
> 
> With libstdc++.so.6.0.28 we get no out-of-bounds checks and the byte
> preceding the storage of the empty string is returned. The character
> value of this byte depends on heap implementation and usage, but when
> this byte happens to hold the value of the directory separator character
> we go on to call std::string::pop_back on the empty string which results
> in an out_of_range exception which terminates GDB.
> 
> Fix this by using path_join. prepare_path_for_appending ensures that the
> filename component is relative.
> 
> The testsuite has been run before and after the change and no
> regressions were found.
> ---
>  gdb/source.c | 10 +---------
>  1 file changed, 1 insertion(+), 9 deletions(-)
> 
> diff --git a/gdb/source.c b/gdb/source.c
> index 3f498d552c4..25ad1ecb3da 100644
> --- a/gdb/source.c
> +++ b/gdb/source.c
> @@ -1146,15 +1146,7 @@ find_and_open_source (const char *filename,
>  	 helpful if part of the compilation directory was removed,
>  	 e.g. using gcc's -fdebug-prefix-map, and we have added the missing
>  	 prefix to source_path.  */
> -      std::string cdir_filename (dirname);
> -
> -      /* Remove any trailing directory separators.  */
> -      while (IS_DIR_SEPARATOR (cdir_filename.back ()))
> -	cdir_filename.pop_back ();
> -
> -      /* Add our own directory separator.  */
> -      cdir_filename.append (SLASH_STRING);
> -      cdir_filename.append (filename_start);
> +      std::string cdir_filename = path_join (dirname, filename_start);
>  
>        result = openp (path, OPF_SEARCH_IN_PATH | OPF_RETURN_REALPATH,
>  		      cdir_filename.c_str (), OPEN_MODE, fullname);

Thanks, this is OK, nice cleanup.

Simon
  

On Wed, Sep 21 2022, Simon Marchi wrote:
>>        result = openp (path, OPF_SEARCH_IN_PATH | OPF_RETURN_REALPATH,
>>  		      cdir_filename.c_str (), OPEN_MODE, fullname);
>
> Thanks, this is OK, nice cleanup.

Thanks, I'll push the patch tomorrow morning if no other concerns are raised.

>
> Simon
  

On Wed, Sep 21 2022, Magne Hov wrote:

> When a source file's dirname is solely made up of directory separators
> we end up trying to dereference the last character of an empty string
> with std::string::back, which results in undefined behaviour. A typical
> use case where this can happen is when the root directory "/" is used as
> a compilation directory.
>
> With libstdc++.so.6.0.28 we get no out-of-bounds checks and the byte
> preceding the storage of the empty string is returned. The character
> value of this byte depends on heap implementation and usage, but when
> this byte happens to hold the value of the directory separator character
> we go on to call std::string::pop_back on the empty string which results
> in an out_of_range exception which terminates GDB.
>
> Fix this by using path_join. prepare_path_for_appending ensures that the
> filename component is relative.
>
> The testsuite has been run before and after the change and no
> regressions were found.
> ---
>  gdb/source.c | 10 +---------
>  1 file changed, 1 insertion(+), 9 deletions(-)
>
> diff --git a/gdb/source.c b/gdb/source.c
> index 3f498d552c4..25ad1ecb3da 100644
> --- a/gdb/source.c
> +++ b/gdb/source.c
> @@ -1146,15 +1146,7 @@ find_and_open_source (const char *filename,
>  	 helpful if part of the compilation directory was removed,
>  	 e.g. using gcc's -fdebug-prefix-map, and we have added the missing
>  	 prefix to source_path.  */
> -      std::string cdir_filename (dirname);
> -
> -      /* Remove any trailing directory separators.  */
> -      while (IS_DIR_SEPARATOR (cdir_filename.back ()))
> -	cdir_filename.pop_back ();
> -
> -      /* Add our own directory separator.  */
> -      cdir_filename.append (SLASH_STRING);
> -      cdir_filename.append (filename_start);
> +      std::string cdir_filename = path_join (dirname, filename_start);
>  
>        result = openp (path, OPF_SEARCH_IN_PATH | OPF_RETURN_REALPATH,
>  		      cdir_filename.c_str (), OPEN_MODE, fullname);
> -- 
> 2.25.1
>

This has been pushed now. Thank you all for the review.