Re: [PATCH] block: replace deprecated strncpy with strscpy

[Alison Schofield]

On Wed, Oct 18, 2023 at 03:39:59PM -0700, Justin Stitt wrote:
> I have a feeling I may have botched the subject line for this patch.
> Can anyone confirm if it's good or not?
> 
> Automated tooling told me that this was the most common patch
> prefix but it may be caused by large patch series that just
> happened to touch this file once.
> 
> Should the subject be: nvdimm/btt: ... ?
> 
> Judging from [1] I see a few "block" and a few of nvdimm/btt.

Hi Justin,

It should be nvdimm/btt because it only touches btt.c.

Here's the old school way that I use to find prefixes. Maybe you can
train your automated tooling to do this, and then share it with me ;)

I do:

~/git/linux/drivers/nvdimm$ git log --pretty=oneline --abbrev-commit btt.c

3222d8c2a7f8 block: remove ->rw_page
ba229aa8f249 nvdimm-btt: Use the enum req_op type
86947df3a923 block: Change the type of the last .rw_page() argument
8b9ab6266204 block: remove blk_cleanup_disk
3205190655ea nvdimm-btt: use bvec_kmap_local in btt_rw_integrity
322cbb50de71 block: remove genhd.h

And I see a few choices, with 'block' being pretty common.
I peek in those patches and see that block was used when the patch
included files in drivers/block AND also in nvdimm/btt.

Use nvdimm/btt for your patch.

A bit more below -

> 
> On Wed, Oct 18, 2023 at 3:35 PM Justin Stitt <justinstitt@xxxxxxxxxx> wrote:
> >
> > strncpy() is deprecated for use on NUL-terminated destination strings
> > [1] and as such we should prefer more robust and less ambiguous string
> > interfaces.
> >
> > We expect super->signature to be NUL-terminated based on its usage with
> > memcpy against a NUL-term'd buffer:
> > btt_devs.c:
> > 253 | if (memcmp(super->signature, BTT_SIG, BTT_SIG_LEN) != 0)
> > btt.h:
> > 13  | #define BTT_SIG "BTT_ARENA_INFO\0"
> >
> > NUL-padding is not required as `super` is already zero-allocated:
> > btt.c:
> > 985 | super = kzalloc(sizeof(struct btt_sb), GFP_NOIO);
> > ... rendering any additional NUL-padding superfluous.
> >
> > Considering the above, a suitable replacement is `strscpy` [2] due to
> > the fact that it guarantees NUL-termination on the destination buffer
> > without unnecessarily NUL-padding.
> >
> > Let's also use the more idiomatic strscpy usage of (dest, src,
> > sizeof(dest)) instead of (dest, src, XYZ_LEN) for buffers that the
> > compiler can determine the size of. This more tightly correlates the
> > destination buffer to the amount of bytes copied.
> >
> > Side note, this pattern of memcmp() on two NUL-terminated strings should
> > really be changed to just a strncmp(), if i'm not mistaken? I see
> > multiple instances of this pattern in this system.

I'm not following this note about memcmp() usage. Where is that?

> >
> > Link: https://www.kernel.org/doc/html/latest/process/deprecated.html#strncpy-on-nul-terminated-strings [1]
> > Link: https://manpages.debian.org/testing/linux-manual-4.8/strscpy.9.en.html [2]
> > Link: https://github.com/KSPP/linux/issues/90
> > Cc: linux-hardening@xxxxxxxxxxxxxxx
> > Signed-off-by: Justin Stitt <justinstitt@xxxxxxxxxx>
> > ---
> > Note: build-tested only.
> >
> > Found with: $ rg "strncpy\("

How you found it goes in the commit log, not below the line.

> > ---
> >  drivers/nvdimm/btt.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/nvdimm/btt.c b/drivers/nvdimm/btt.c
> > index d5593b0dc700..9372c36e8f76 100644
> > --- a/drivers/nvdimm/btt.c
> > +++ b/drivers/nvdimm/btt.c
> > @@ -986,7 +986,7 @@ static int btt_arena_write_layout(struct arena_info *arena)
> >         if (!super)
> >                 return -ENOMEM;
> >
> > -       strncpy(super->signature, BTT_SIG, BTT_SIG_LEN);
> > +       strscpy(super->signature, BTT_SIG, sizeof(super->signature));
> >         export_uuid(super->uuid, nd_btt->uuid);
> >         export_uuid(super->parent_uuid, parent_uuid);
> >         super->flags = cpu_to_le32(arena->flags);
> >
> > ---
> > base-commit: 58720809f52779dc0f08e53e54b014209d13eebb
> > change-id: 20231018-strncpy-drivers-nvdimm-btt-c-15f93879989e
> >
> > Best regards,
> > --
> > Justin Stitt <justinstitt@xxxxxxxxxx>
> >
> 
> [1]: https://lore.kernel.org/all/?q=dfn%3Adrivers%2Fnvdimm%2Fbtt.c
> 
> Thanks
> Justin
>