strncpy() is deprecated for use on NUL-terminated destination strings
[1] and as such we should prefer more robust and less ambiguous string
interfaces.
@query is already marked as __nonstring and doesn't need to be
NUL-terminated. Since @id is a string, we can use the self-describing
string API strtomem().
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: [email protected]
Signed-off-by: Justin Stitt <[email protected]>
---
Changes in v2:
- use strtomem instead of memcpy (thanks Kees)
- Link to v1: https://lore.kernel.org/r/20240314-strncpy-drivers-soc-qcom-cmd-db-c-v1-1-70f5d5e70732@google.com
---
Note: build-tested only.
Found with: $ rg "strncpy\("
---
drivers/soc/qcom/cmd-db.c | 9 ++-------
1 file changed, 2 insertions(+), 7 deletions(-)
diff --git a/drivers/soc/qcom/cmd-db.c b/drivers/soc/qcom/cmd-db.c
index a5fd68411bed..d05f35d175bd 100644
--- a/drivers/soc/qcom/cmd-db.c
+++ b/drivers/soc/qcom/cmd-db.c
@@ -141,18 +141,13 @@ static int cmd_db_get_header(const char *id, const struct entry_header **eh,
const struct rsc_hdr *rsc_hdr;
const struct entry_header *ent;
int ret, i, j;
- u8 query[sizeof(ent->id)] __nonstring;
+ u8 query[sizeof(ent->id)] __nonstring = { 0 };
ret = cmd_db_ready();
if (ret)
return ret;
- /*
- * Pad out query string to same length as in DB. NOTE: the output
- * query string is not necessarily '\0' terminated if it bumps up
- * against the max size. That's OK and expected.
- */
- strncpy(query, id, sizeof(query));
+ strtomem(query, id);
for (i = 0; i < MAX_SLV_ID; i++) {
rsc_hdr = &cmd_db_header->header[i];
---
base-commit: fe46a7dd189e25604716c03576d05ac8a5209743
change-id: 20240314-strncpy-drivers-soc-qcom-cmd-db-c-284f3abaabb8
Best regards,
--
Justin Stitt <[email protected]>
On Mon, Mar 18, 2024 at 10:49:23PM +0000, Justin Stitt 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.
>
> @query is already marked as __nonstring and doesn't need to be
> NUL-terminated. Since @id is a string, we can use the self-describing
> string API strtomem().
>
> 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: [email protected]
> Signed-off-by: Justin Stitt <[email protected]>
Nice! A textbook use for strtomem(). :)
Reviewed-by: Kees Cook <[email protected]>
--
Kees Cook
On Mon, Mar 18, 2024 at 10:49:23PM +0000, Justin Stitt 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.
>
I don't mind changing the strncpy() in this function, but I don't think
this problem description adequately describes the problem you're
solving.
If the motivation is that we want 0 users of strncpy() in the kernel,
then say so.
> @query is already marked as __nonstring and doesn't need to be
> NUL-terminated.
You're not wrong, but in the event that strlen(id) < sizeof(ent->id) the
destination should be NUL-padded - exactly one of the well known,
normally unwanted, effects of strncpy(). strtomem() does explicitly not
do this.
> Since @id is a string, we can use the self-describing
> string API strtomem().
"self-describing"?
>
> 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: [email protected]
> Signed-off-by: Justin Stitt <[email protected]>
> ---
> Changes in v2:
> - use strtomem instead of memcpy (thanks Kees)
> - Link to v1: https://lore.kernel.org/r/20240314-strncpy-drivers-soc-qcom-cmd-db-c-v1-1-70f5d5e70732@google.com
> ---
> Note: build-tested only.
>
> Found with: $ rg "strncpy\("
> ---
> drivers/soc/qcom/cmd-db.c | 9 ++-------
> 1 file changed, 2 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/soc/qcom/cmd-db.c b/drivers/soc/qcom/cmd-db.c
> index a5fd68411bed..d05f35d175bd 100644
> --- a/drivers/soc/qcom/cmd-db.c
> +++ b/drivers/soc/qcom/cmd-db.c
> @@ -141,18 +141,13 @@ static int cmd_db_get_header(const char *id, const struct entry_header **eh,
> const struct rsc_hdr *rsc_hdr;
> const struct entry_header *ent;
> int ret, i, j;
> - u8 query[sizeof(ent->id)] __nonstring;
> + u8 query[sizeof(ent->id)] __nonstring = { 0 };
>
> ret = cmd_db_ready();
> if (ret)
> return ret;
>
> - /*
> - * Pad out query string to same length as in DB. NOTE: the output
> - * query string is not necessarily '\0' terminated if it bumps up
> - * against the max size. That's OK and expected.
> - */
> - strncpy(query, id, sizeof(query));
> + strtomem(query, id);
query needs to be NUL-padded to sizeof(ent->id) bytes (like strncpy
does), something you recognized by adding the zero-initialization above.
But why split this requirement across two non-adjacent lines? Isn't this
what strtomem_pad() is supposed to do?
Regards,
Bjorn
>
> for (i = 0; i < MAX_SLV_ID; i++) {
> rsc_hdr = &cmd_db_header->header[i];
>
> ---
> base-commit: fe46a7dd189e25604716c03576d05ac8a5209743
> change-id: 20240314-strncpy-drivers-soc-qcom-cmd-db-c-284f3abaabb8
>
> Best regards,
> --
> Justin Stitt <[email protected]>
>
Hi,
On Mon, Mar 18, 2024 at 8:37 PM Bjorn Andersson <[email protected]> wrote:
>
> On Mon, Mar 18, 2024 at 10:49:23PM +0000, Justin Stitt 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.
> >
>
> I don't mind changing the strncpy() in this function, but I don't think
> this problem description adequately describes the problem you're
> solving.
>
> If the motivation is that we want 0 users of strncpy() in the kernel,
> then say so.
Fair. You caught me in a bad case of "copy pasting this blurb into all
my patches". You are right though, the true motivation here is to rid
the kernel of strncpy.
>
> > @query is already marked as __nonstring and doesn't need to be
> > NUL-terminated.
>
> You're not wrong, but in the event that strlen(id) < sizeof(ent->id) the
> destination should be NUL-padded - exactly one of the well known,
> normally unwanted, effects of strncpy(). strtomem() does explicitly not
> do this.
>
> > Since @id is a string, we can use the self-describing
> > string API strtomem().
>
> "self-describing"?
>
In the sense that its name matches its functionality:
strncpy === string to string copy, bounded by n
strtomem === string to memory buffer
strncpy technically does the latter functionality as well but it may
not be obvious in all cases that the destination buffer is not a
string. Granted, in this case, it is extremely obvious what the
behavior is because query is marked nonstring.
> >
> > 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.enhtml [2]
> > Link: https://github.com/KSPP/linux/issues/90
> > Cc: [email protected]
> > Signed-off-by: Justin Stitt <[email protected]>
> > ---
> > Changes in v2:
> > - use strtomem instead of memcpy (thanks Kees)
> > - Link to v1: https://lore.kernel.org/r/20240314-strncpy-drivers-soc-qcom-cmd-db-c-v1-1-70f5d5e70732@google.com
> > ---
> > Note: build-tested only.
> >
> > Found with: $ rg "strncpy\("
> > ---
> > drivers/soc/qcom/cmd-db.c | 9 ++-------
> > 1 file changed, 2 insertions(+), 7 deletions(-)
> >
> > diff --git a/drivers/soc/qcom/cmd-db.c b/drivers/soc/qcom/cmd-db.c
> > index a5fd68411bed..d05f35d175bd 100644
> > --- a/drivers/soc/qcom/cmd-db.c
> > +++ b/drivers/soc/qcom/cmd-db.c
> > @@ -141,18 +141,13 @@ static int cmd_db_get_header(const char *id, const struct entry_header **eh,
> > const struct rsc_hdr *rsc_hdr;
> > const struct entry_header *ent;
> > int ret, i, j;
> > - u8 query[sizeof(ent->id)] __nonstring;
> > + u8 query[sizeof(ent->id)] __nonstring = { 0 };
> >
> > ret = cmd_db_ready();
> > if (ret)
> > return ret;
> >
> > - /*
> > - * Pad out query string to same length as in DB. NOTE: the output
> > - * query string is not necessarily '\0' terminated if it bumps up
> > - * against the max size. That's OK and expected.
> > - */
> > - strncpy(query, id, sizeof(query));
> > + strtomem(query, id);
>
> query needs to be NUL-padded to sizeof(ent->id) bytes (like strncpy
> does), something you recognized by adding the zero-initialization above.
> But why split this requirement across two non-adjacent lines? Isn't this
> what strtomem_pad() is supposed to do?
Yes, strtomem_pad() will accomplish this task. I'll send a v3 fixing
up the commit log and use the pad version.
>
> Regards,
> Bjorn
>
> >
> > for (i = 0; i < MAX_SLV_ID; i++) {
> > rsc_hdr = &cmd_db_header->header[i];
> >
> > ---
> > base-commit: fe46a7dd189e25604716c03576d05ac8a5209743
> > change-id: 20240314-strncpy-drivers-soc-qcom-cmd-db-c-284f3abaabb8
> >
> > Best regards,
> > --
> > Justin Stitt <[email protected]>
> >
Thanks
Justin