Re: [PATCH v4 2/4] tools/bootconfig: Fix to check the write failure correctly

[Masami Hiramatsu <mhiramat@kernel.org>]

On Wed, 18 Nov 2020 11:04:37 -0500
Steven Rostedt <rostedt@goodmis.org> wrote:

> On Thu, 19 Nov 2020 00:35:35 +0900
> Masami Hiramatsu <mhiramat@kernel.org> wrote:
> 
> > Fix to check the write(2) failure including partial write
> > correctly and try to rollback the partial write, because
> > if there is no BOOTCONFIG_MAGIC string, we can not remove it.
> > 
> > Fixes: 85c46b78da58 ("bootconfig: Add bootconfig magic word for indicating bootconfig explicitly")
> > Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
> > Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
> > Tested-by: Chen Yu <yu.chen.surf@gmail.com>
> > ---
> >  tools/bootconfig/main.c |   27 +++++++++++++++++++++++----
> >  1 file changed, 23 insertions(+), 4 deletions(-)
> > 
> > diff --git a/tools/bootconfig/main.c b/tools/bootconfig/main.c
> > index 52eb2bbe8966..905bfaefae35 100644
> > --- a/tools/bootconfig/main.c
> > +++ b/tools/bootconfig/main.c
> > @@ -337,6 +337,7 @@ static int delete_xbc(const char *path)
> >  
> >  static int apply_xbc(const char *path, const char *xbc_path)
> >  {
> > +	struct stat stat;
> >  	u32 size, csum;
> >  	char *buf, *data;
> >  	int ret, fd;
> > @@ -394,16 +395,26 @@ static int apply_xbc(const char *path, const char *xbc_path)
> >  		return ret;
> >  	}
> >  	/* TODO: Ensure the @path is initramfs/initrd image */
> > +	if (fstat(fd, &stat) < 0) {
> > +		pr_err("Failed to get the size of %s\n", path);
> > +		goto out;
> > +	}
> >  	ret = write(fd, data, size + 8);
> > -	if (ret < 0) {
> > +	if (ret < size + 8) {
> > +		if (ret < 0)
> > +			ret = -errno;
> >  		pr_err("Failed to apply a boot config: %d\n", ret);
> > -		goto out;
> > +		if (ret < 0)
> > +			goto out;
> > +		goto out_rollback;
> >  	}
> >  	/* Write a magic word of the bootconfig */
> >  	ret = write(fd, BOOTCONFIG_MAGIC, BOOTCONFIG_MAGIC_LEN);
> > -	if (ret < 0) {
> > +	if (ret < BOOTCONFIG_MAGIC_LEN) {
> > +		if (ret < 0)
> > +			ret = -errno;
> >  		pr_err("Failed to apply a boot config magic: %d\n", ret);
> > -		goto out;
> > +		goto out_rollback;
> >  	}
> >  	ret = 0;
> >  out:
> > @@ -411,6 +422,14 @@ static int apply_xbc(const char *path, const char *xbc_path)
> >  	free(data);
> >  
> >  	return ret;
> > +
> > +out_rollback:
> > +	if (ftruncate(fd, stat.st_size) < 0) {
> 
> If the first write fails (doesn't modify the initrd), and then this fails,
> you will show a message of: "may be corrupted. Recommend to rebuild", when
> that would not be the case. Should a test of the size be done again, to see
> if it is already the same?

Please see the above 2 write error check. You can find

if (ret < 0)
	goto out;

only in the first check. (I know, it might a bit misleadable)
So, the first write is failed with error message, it doesn't call rollback.

> 
> out_rollback:
> 	old_size = stat.st_size;
> 	if (stat(fd, &stat) < 0) /* Shouldn't happen, it worked once before */
> 		goto print_error;
> 
> 	if (old_size < stat.st_size)
> 		if (ftruncate(fd, old_size) < 0)
> 			goto print_error;

If the old_size is current size, ftruncate must do nothing, doesn't it?

Thank you,

> 
> 	goto out;
> 
> print_error:
> 
> > +		ret = -errno;
> > +		pr_err("Failed to rollback the write error: %d\n", ret);
> > +		pr_err("The initrd %s may be corrupted. Recommend to rebuild.\n", path);
> > +	}
> > +	goto out;
> >  }
> >  
> >  static int usage(void)
> 
> -- Steve


-- 
Masami Hiramatsu <mhiramat@kernel.org>