[PATCH v4 39/66] binfmt_elf: Take the mmap lock when walking the VMA list

[Vlastimil Babka]

On 12/1/21 15:30, Liam Howlett wrote:
> From: "Matthew Wilcox (Oracle)" <willy at infradead.org>
> 
> I'm not sure if the VMA list can change under us, but dump_vma_snapshot()
> is very careful to take the mmap_lock in write mode.  We only need to
> take it in read mode here as we do not care if the size of the stack
> VMA changes underneath us.
> 
> If it can be changed underneath us, this is a potential use-after-free
> for a multithreaded process which is dumping core.
> 
> Fixes: 2aa362c49c31 ("coredump: extend core dump note section to contain file names of mapped files")
> Signed-off-by: Matthew Wilcox (Oracle) <willy at infradead.org>
> Signed-off-by: Liam R. Howlett <Liam.Howlett at Oracle.com>

To be honest this feels out of place in this series. Send separately and CC
Jann who AFAICS added dump_vma_snapshot()?

> ---
>  fs/binfmt_elf.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c
> index d41cca755ff9..5915518c8a1d 100644
> --- a/fs/binfmt_elf.c
> +++ b/fs/binfmt_elf.c
> @@ -1652,6 +1652,7 @@ static int fill_files_note(struct memelfnote *note)
>  	name_base = name_curpos = ((char *)data) + names_ofs;
>  	remaining = size - names_ofs;
>  	count = 0;
> +	mmap_read_lock(mm);
>  	for_each_vma(vmi, vma) {
>  		struct file *file;
>  		const char *filename;
> @@ -1662,6 +1663,7 @@ static int fill_files_note(struct memelfnote *note)
>  		filename = file_path(file, name_curpos, remaining);
>  		if (IS_ERR(filename)) {
>  			if (PTR_ERR(filename) == -ENAMETOOLONG) {
> +				mmap_read_unlock(mm);
>  				kvfree(data);
>  				size = size * 5 / 4;
>  				goto alloc;
> @@ -1681,6 +1683,7 @@ static int fill_files_note(struct memelfnote *note)
>  		*start_end_ofs++ = vma->vm_pgoff;
>  		count++;
>  	}
> +	mmap_read_unlock(mm);
>  
>  	/* Now we know exact count of files, can store it */
>  	data[0] = count;