The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
sizeof(flexible-array-member) triggers a warning because flexible array
members have incomplete type[1]. There are some instances of code in
which the sizeof operator is being incorrectly/erroneously applied to
zero-length arrays and the result is zero. Such instances may be hiding
some bugs. So, this work (flexible-array member conversions) will also
help to get completely rid of those sorts of issues.
Also, the following issue shows up due to the flexible-array member
having incomplete type[4]:
drivers/char/tpm/eventlog/tpm2.c: In function ‘tpm2_bios_measurements_start’:
drivers/char/tpm/eventlog/tpm2.c:54:46: error: invalid application of ‘sizeof’ to incomplete type ‘u8[]’ {aka ‘unsigned char[]’}
54 | size = sizeof(struct tcg_pcr_event) - sizeof(event_header->event)
| ^
drivers/char/tpm/eventlog/tpm2.c: In function ‘tpm2_bios_measurements_next’:
drivers/char/tpm/eventlog/tpm2.c:102:10: error: invalid application of ‘sizeof’ to incomplete type ‘u8[]’ {aka ‘unsigned char[]’}
102 | sizeof(event_header->event) + event_header->event_size;
| ^
drivers/char/tpm/eventlog/tpm2.c: In function ‘tpm2_binary_bios_measurements_show’:
drivers/char/tpm/eventlog/tpm2.c:140:10: error: invalid application of ‘sizeof’ to incomplete type ‘u8[]’ {aka ‘unsigned char[]’}
140 | sizeof(event_header->event) + event_header->event_size;
| ^
scripts/Makefile.build:266: recipe for target 'drivers/char/tpm/eventlog/tpm2.o' failed
make[3]: *** [drivers/char/tpm/eventlog/tpm2.o] Error 1
As mentioned above: "Flexible array members have incomplete type, and
so the sizeof operator may not be applied. As a quirk of the original
implementation of zero-length arrays, sizeof evaluates to zero."[1] So,
the sizeof(flexible-array) can be safely removed to fix the error above.
Lastly, prefer sizeof(*ptr) over sizeof(struct foo).
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")
[4] https://github.com/KSPP/linux/issues/43
Signed-off-by: Gustavo A. R. Silva <[email protected]>
---
Hi,
I'm resending this because LKML is eating some messages, recently.
Sorry for the noise in case you've already received this patch.
Thanks
drivers/char/tpm/eventlog/tpm2.c | 10 +++-------
include/linux/tpm_eventlog.h | 2 +-
2 files changed, 4 insertions(+), 8 deletions(-)
diff --git a/drivers/char/tpm/eventlog/tpm2.c b/drivers/char/tpm/eventlog/tpm2.c
index e741b1157525..351a2989b3c6 100644
--- a/drivers/char/tpm/eventlog/tpm2.c
+++ b/drivers/char/tpm/eventlog/tpm2.c
@@ -51,8 +51,7 @@ static void *tpm2_bios_measurements_start(struct seq_file *m, loff_t *pos)
int i;
event_header = addr;
- size = sizeof(struct tcg_pcr_event) - sizeof(event_header->event)
- + event_header->event_size;
+ size = sizeof(*event_header) + event_header->event_size;
if (*pos == 0) {
if (addr + size < limit) {
@@ -98,8 +97,7 @@ static void *tpm2_bios_measurements_next(struct seq_file *m, void *v,
event_header = log->bios_event_log;
if (v == SEQ_START_TOKEN) {
- event_size = sizeof(struct tcg_pcr_event) -
- sizeof(event_header->event) + event_header->event_size;
+ event_size = sizeof(*event_header) + event_header->event_size;
marker = event_header;
} else {
event = v;
@@ -136,9 +134,7 @@ static int tpm2_binary_bios_measurements_show(struct seq_file *m, void *v)
size_t size;
if (v == SEQ_START_TOKEN) {
- size = sizeof(struct tcg_pcr_event) -
- sizeof(event_header->event) + event_header->event_size;
-
+ size = sizeof(*event_header) + event_header->event_size;
temp_ptr = event_header;
if (size > 0)
diff --git a/include/linux/tpm_eventlog.h b/include/linux/tpm_eventlog.h
index c253461b1c4e..4f8c90c93c29 100644
--- a/include/linux/tpm_eventlog.h
+++ b/include/linux/tpm_eventlog.h
@@ -97,7 +97,7 @@ struct tcg_pcr_event {
u32 event_type;
u8 digest[20];
u32 event_size;
- u8 event[0];
+ u8 event[];
} __packed;
struct tcg_event_field {
--
2.26.2
On Wed, May 06, 2020 at 11:09:12PM -0500, Gustavo A. R. Silva wrote:
> As mentioned above: "Flexible array members have incomplete type, and
> so the sizeof operator may not be applied. As a quirk of the original
> implementation of zero-length arrays, sizeof evaluates to zero."[1] So,
> the sizeof(flexible-array) can be safely removed to fix the error above.
As in "sizeof(event_header->event) always evaluated to 0, so removing it
has no effect".
> [...]
> diff --git a/drivers/char/tpm/eventlog/tpm2.c b/drivers/char/tpm/eventlog/tpm2.c
> index e741b1157525..351a2989b3c6 100644
> --- a/drivers/char/tpm/eventlog/tpm2.c
> +++ b/drivers/char/tpm/eventlog/tpm2.c
> @@ -51,8 +51,7 @@ static void *tpm2_bios_measurements_start(struct seq_file *m, loff_t *pos)
> int i;
>
> event_header = addr;
> - size = sizeof(struct tcg_pcr_event) - sizeof(event_header->event)
> - + event_header->event_size;
> + size = sizeof(*event_header) + event_header->event_size;
That said, I think it would be better to stick to the struct_size()
idiom for dealing with flexible arrays here:
size = struct_size(event_header, event, event_size);
--
Kees Cook
On Thu, May 07, 2020 at 11:02:18AM -0700, Kees Cook wrote:
> On Wed, May 06, 2020 at 11:09:12PM -0500, Gustavo A. R. Silva wrote:
> > As mentioned above: "Flexible array members have incomplete type, and
> > so the sizeof operator may not be applied. As a quirk of the original
> > implementation of zero-length arrays, sizeof evaluates to zero."[1] So,
> > the sizeof(flexible-array) can be safely removed to fix the error above.
>
> As in "sizeof(event_header->event) always evaluated to 0, so removing it
> has no effect".
>
Thanks for this. I wanted to make a more general statement, but I'll
update the changelog text. :)
> > [...]
> > diff --git a/drivers/char/tpm/eventlog/tpm2.c b/drivers/char/tpm/eventlog/tpm2.c
> > index e741b1157525..351a2989b3c6 100644
> > --- a/drivers/char/tpm/eventlog/tpm2.c
> > +++ b/drivers/char/tpm/eventlog/tpm2.c
> > @@ -51,8 +51,7 @@ static void *tpm2_bios_measurements_start(struct seq_file *m, loff_t *pos)
> > int i;
> >
> > event_header = addr;
> > - size = sizeof(struct tcg_pcr_event) - sizeof(event_header->event)
> > - + event_header->event_size;
> > + size = sizeof(*event_header) + event_header->event_size;
>
> That said, I think it would be better to stick to the struct_size()
> idiom for dealing with flexible arrays here:
>
> size = struct_size(event_header, event, event_size);
>
Yep, I agree. I'll add this and send v2, shortly.
Thanks
--
Gustavo
On Wed, May 06, 2020 at 11:09:12PM -0500, Gustavo A. R. Silva wrote:
> The current codebase makes use of the zero-length array language
> extension to the C90 standard, but the preferred mechanism to declare
> variable-length types such as these ones is a flexible array member[1][2],
> introduced in C99:
>
> struct foo {
> int stuff;
> struct boo array[];
> };
>
> By making use of the mechanism above, we will get a compiler warning
> in case the flexible array does not occur last in the structure, which
> will help us prevent some kind of undefined behavior bugs from being
> inadvertently introduced[3] to the codebase from now on.
>
> Also, notice that, dynamic memory allocations won't be affected by
> this change:
>
> "Flexible array members have incomplete type, and so the sizeof operator
> may not be applied. As a quirk of the original implementation of
> zero-length arrays, sizeof evaluates to zero."[1]
>
> sizeof(flexible-array-member) triggers a warning because flexible array
> members have incomplete type[1]. There are some instances of code in
> which the sizeof operator is being incorrectly/erroneously applied to
> zero-length arrays and the result is zero. Such instances may be hiding
> some bugs. So, this work (flexible-array member conversions) will also
> help to get completely rid of those sorts of issues.
>
> Also, the following issue shows up due to the flexible-array member
> having incomplete type[4]:
>
> drivers/char/tpm/eventlog/tpm2.c: In function ‘tpm2_bios_measurements_start’:
> drivers/char/tpm/eventlog/tpm2.c:54:46: error: invalid application of ‘sizeof’ to incomplete type ‘u8[]’ {aka ‘unsigned char[]’}
> 54 | size = sizeof(struct tcg_pcr_event) - sizeof(event_header->event)
> | ^
> drivers/char/tpm/eventlog/tpm2.c: In function ‘tpm2_bios_measurements_next’:
> drivers/char/tpm/eventlog/tpm2.c:102:10: error: invalid application of ‘sizeof’ to incomplete type ‘u8[]’ {aka ‘unsigned char[]’}
> 102 | sizeof(event_header->event) + event_header->event_size;
> | ^
> drivers/char/tpm/eventlog/tpm2.c: In function ‘tpm2_binary_bios_measurements_show’:
> drivers/char/tpm/eventlog/tpm2.c:140:10: error: invalid application of ‘sizeof’ to incomplete type ‘u8[]’ {aka ‘unsigned char[]’}
> 140 | sizeof(event_header->event) + event_header->event_size;
> | ^
> scripts/Makefile.build:266: recipe for target 'drivers/char/tpm/eventlog/tpm2.o' failed
> make[3]: *** [drivers/char/tpm/eventlog/tpm2.o] Error 1
>
> As mentioned above: "Flexible array members have incomplete type, and
> so the sizeof operator may not be applied. As a quirk of the original
> implementation of zero-length arrays, sizeof evaluates to zero."[1] So,
> the sizeof(flexible-array) can be safely removed to fix the error above.
>
> Lastly, prefer sizeof(*ptr) over sizeof(struct foo).
>
> This issue was found with the help of Coccinelle.
>
> [1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
> [2] https://github.com/KSPP/linux/issues/21
> [3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")
> [4] https://github.com/KSPP/linux/issues/43
>
> Signed-off-by: Gustavo A. R. Silva <[email protected]>
Reviewed-by: Jarkko Sakkinen <[email protected]>
/Jarkko
On Thu, May 14, 2020 at 12:39:05AM +0300, Jarkko Sakkinen wrote:
> >
> > [1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
> > [2] https://github.com/KSPP/linux/issues/21
> > [3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")
> > [4] https://github.com/KSPP/linux/issues/43
> >
> > Signed-off-by: Gustavo A. R. Silva <[email protected]>
>
> Reviewed-by: Jarkko Sakkinen <[email protected]>
>
Hi Jarkko,
Thanks for your RB.
There is a v2 of this patch:
https://lore.kernel.org/lkml/20200508163826.GA768@embeddedor/
Thanks
--
Gustavo
On Wed, 2020-05-13 at 18:14 -0500, Gustavo A. R. Silva wrote:
> On Thu, May 14, 2020 at 12:39:05AM +0300, Jarkko Sakkinen wrote:
> > > [1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
> > > [2] https://github.com/KSPP/linux/issues/21
> > > [3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")
> > > [4] https://github.com/KSPP/linux/issues/43
> > >
> > > Signed-off-by: Gustavo A. R. Silva <[email protected]>
> >
> > Reviewed-by: Jarkko Sakkinen <[email protected]>
> >
>
> Hi Jarkko,
>
> Thanks for your RB.
>
> There is a v2 of this patch:
>
> https://lore.kernel.org/lkml/20200508163826.GA768@embeddedor/
>
> Thanks
> --
> Gustavo
Yup,
http://git.infradead.org/users/jjs/linux-tpmdd.git/commit/47c18d91f4adb2ca164c8dbee861543b4466167d
Does this look correct?
/Jarkko