2013-05-19 13:00:06

by Dan Carpenter

Subject: [patch] gru: fix a sanity test in gru_set_context_option()

"req.val1 == -1" is valid but it doesn't make sense to check
gru_base[-1]. gru_base[] is a global array.

Signed-off-by: Dan Carpenter <[email protected]>

diff --git a/drivers/misc/sgi-gru/grufault.c b/drivers/misc/sgi-gru/grufault.c
index c4acac7..e7c37b2 100644
--- a/drivers/misc/sgi-gru/grufault.c
+++ b/drivers/misc/sgi-gru/grufault.c
@@ -876,7 +876,8 @@ int gru_set_context_option(unsigned long arg)
switch (req.op) {
case sco_blade_chiplet:
/* Select blade/chiplet for GRU context */
- if (req.val1 < -1 || req.val1 >= GRU_MAX_BLADES || !gru_base[req.val1] ||
+ if (req.val1 < -1 || req.val1 >= GRU_MAX_BLADES ||
+ (req.val1 != -1 && !gru_base[req.val1]) ||
req.val0 < -1 || req.val0 >= GRU_CHIPLETS_PER_HUB) {
ret = -EINVAL;
} else {

2013-05-19 14:44:46

by Dimitri Sivanich

Subject: Re: [patch] gru: fix a sanity test in gru_set_context_option()

On Sun, May 19, 2013 at 03:59:50PM +0300, Dan Carpenter wrote:
> "req.val1 == -1" is valid but it doesn't make sense to check
> gru_base[-1]. gru_base[] is a global array.
>
> Signed-off-by: Dan Carpenter <[email protected]>
>
> diff --git a/drivers/misc/sgi-gru/grufault.c b/drivers/misc/sgi-gru/grufault.c
> index c4acac7..e7c37b2 100644
> --- a/drivers/misc/sgi-gru/grufault.c
> +++ b/drivers/misc/sgi-gru/grufault.c
> @@ -876,7 +876,8 @@ int gru_set_context_option(unsigned long arg)
> switch (req.op) {
> case sco_blade_chiplet:
> /* Select blade/chiplet for GRU context */
> - if (req.val1 < -1 || req.val1 >= GRU_MAX_BLADES || !gru_base[req.val1] ||
> + if (req.val1 < -1 || req.val1 >= GRU_MAX_BLADES ||
> + (req.val1 != -1 && !gru_base[req.val1]) ||
> req.val0 < -1 || req.val0 >= GRU_CHIPLETS_PER_HUB) {
> ret = -EINVAL;
> } else {

Since we're here, maybe we should neaten this up a little?

Signed-off-by: Dimitri Sivanich <[email protected]>
---
drivers/misc/sgi-gru/grufault.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)

Index: linux/drivers/misc/sgi-gru/grufault.c
===================================================================
--- linux.orig/drivers/misc/sgi-gru/grufault.c
+++ linux/drivers/misc/sgi-gru/grufault.c
@@ -876,8 +876,9 @@ int gru_set_context_option(unsigned long
switch (req.op) {
case sco_blade_chiplet:
/* Select blade/chiplet for GRU context */
- if (req.val1 < -1 || req.val1 >= GRU_MAX_BLADES || !gru_base[req.val1] ||
- req.val0 < -1 || req.val0 >= GRU_CHIPLETS_PER_HUB) {
+ if (req.val0 < -1 || req.val0 >= GRU_CHIPLETS_PER_HUB ||
+ req.val1 < -1 || req.val1 >= GRU_MAX_BLADES ||
+ (req.val1 >= 0 && !gru_base[req.val1])) {
ret = -EINVAL;
} else {
gts->ts_user_blade_id = req.val1;

2013-05-19 18:30:49

by Dan Carpenter

Subject: Re: [patch] gru: fix a sanity test in gru_set_context_option()

On Sun, May 19, 2013 at 09:44:42AM -0500, Dimitri Sivanich wrote:
>
> Since we're here, maybe we should neaten this up a little?
>
> Signed-off-by: Dimitri Sivanich <[email protected]>
> ---
> drivers/misc/sgi-gru/grufault.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> Index: linux/drivers/misc/sgi-gru/grufault.c
> ===================================================================
> --- linux.orig/drivers/misc/sgi-gru/grufault.c
> +++ linux/drivers/misc/sgi-gru/grufault.c
> @@ -876,8 +876,9 @@ int gru_set_context_option(unsigned long
> switch (req.op) {
> case sco_blade_chiplet:
> /* Select blade/chiplet for GRU context */
> - if (req.val1 < -1 || req.val1 >= GRU_MAX_BLADES || !gru_base[req.val1] ||
> - req.val0 < -1 || req.val0 >= GRU_CHIPLETS_PER_HUB) {
> + if (req.val0 < -1 || req.val0 >= GRU_CHIPLETS_PER_HUB ||
> + req.val1 < -1 || req.val1 >= GRU_MAX_BLADES ||
> + (req.val1 >= 0 && !gru_base[req.val1])) {
> ret = -EINVAL;

I generally prefer how in the original code all the sub conditions
were aligned nicely but I will Ack your version as well. Thanks.

regards,
dan carpenter

> } else {
> gts->ts_user_blade_id = req.val1;

2013-05-20 13:46:24

by Dimitri Sivanich

Subject: Re: [patch] gru: fix a sanity test in gru_set_context_option()

On Sun, May 19, 2013 at 09:30:41PM +0300, Dan Carpenter wrote:
> On Sun, May 19, 2013 at 09:44:42AM -0500, Dimitri Sivanich wrote:
> >
> > Since we're here, maybe we should neaten this up a little?
> >
> > Signed-off-by: Dimitri Sivanich <[email protected]>
> > ---
> > drivers/misc/sgi-gru/grufault.c | 5 +++--
> > 1 file changed, 3 insertions(+), 2 deletions(-)
> >
> > Index: linux/drivers/misc/sgi-gru/grufault.c
> > ===================================================================
> > --- linux.orig/drivers/misc/sgi-gru/grufault.c
> > +++ linux/drivers/misc/sgi-gru/grufault.c
> > @@ -876,8 +876,9 @@ int gru_set_context_option(unsigned long
> > switch (req.op) {
> > case sco_blade_chiplet:
> > /* Select blade/chiplet for GRU context */
> > - if (req.val1 < -1 || req.val1 >= GRU_MAX_BLADES || !gru_base[req.val1] ||
> > - req.val0 < -1 || req.val0 >= GRU_CHIPLETS_PER_HUB) {
> > + if (req.val0 < -1 || req.val0 >= GRU_CHIPLETS_PER_HUB ||
> > + req.val1 < -1 || req.val1 >= GRU_MAX_BLADES ||
> > + (req.val1 >= 0 && !gru_base[req.val1])) {
> > ret = -EINVAL;
>
> I generally prefer how in the original code all the sub conditions
> were aligned nicely but I will Ack your version as well. Thanks.
>

I guess I'd have to agree. How about this?

Signed-off-by: Dimitri Sivanich <[email protected]>
---
drivers/misc/sgi-gru/grufault.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)

Index: linux/drivers/misc/sgi-gru/grufault.c
===================================================================
--- linux.orig/drivers/misc/sgi-gru/grufault.c
+++ linux/drivers/misc/sgi-gru/grufault.c
@@ -876,8 +876,9 @@ int gru_set_context_option(unsigned long
switch (req.op) {
case sco_blade_chiplet:
/* Select blade/chiplet for GRU context */
- if (req.val1 < -1 || req.val1 >= GRU_MAX_BLADES || !gru_base[req.val1] ||
- req.val0 < -1 || req.val0 >= GRU_CHIPLETS_PER_HUB) {
+ if (req.val0 < -1 || req.val0 >= GRU_CHIPLETS_PER_HUB ||
+ req.val1 < -1 || req.val1 >= GRU_MAX_BLADES ||
+ (req.val1 >= 0 && !gru_base[req.val1])) {
ret = -EINVAL;
} else {
gts->ts_user_blade_id = req.val1;

2013-05-20 18:06:59

by Dan Carpenter

Subject: Re: [patch] gru: fix a sanity test in gru_set_context_option()

On Mon, May 20, 2013 at 08:46:21AM -0500, Dimitri Sivanich wrote:
> On Sun, May 19, 2013 at 09:30:41PM +0300, Dan Carpenter wrote:
> > On Sun, May 19, 2013 at 09:44:42AM -0500, Dimitri Sivanich wrote:
> > >
> > > Since we're here, maybe we should neaten this up a little?
> > >
> > > Signed-off-by: Dimitri Sivanich <[email protected]>
> > > ---
> > > drivers/misc/sgi-gru/grufault.c | 5 +++--
> > > 1 file changed, 3 insertions(+), 2 deletions(-)
> > >
> > > Index: linux/drivers/misc/sgi-gru/grufault.c
> > > ===================================================================
> > > --- linux.orig/drivers/misc/sgi-gru/grufault.c
> > > +++ linux/drivers/misc/sgi-gru/grufault.c
> > > @@ -876,8 +876,9 @@ int gru_set_context_option(unsigned long
> > > switch (req.op) {
> > > case sco_blade_chiplet:
> > > /* Select blade/chiplet for GRU context */
> > > - if (req.val1 < -1 || req.val1 >= GRU_MAX_BLADES || !gru_base[req.val1] ||
> > > - req.val0 < -1 || req.val0 >= GRU_CHIPLETS_PER_HUB) {
> > > + if (req.val0 < -1 || req.val0 >= GRU_CHIPLETS_PER_HUB ||
> > > + req.val1 < -1 || req.val1 >= GRU_MAX_BLADES ||
> > > + (req.val1 >= 0 && !gru_base[req.val1])) {
> > > ret = -EINVAL;
> >
> > I generally prefer how in the original code all the sub conditions
> > were aligned nicely but I will Ack your version as well. Thanks.
> >
>
> I guess I'd have to agree. How about this?

Looks nice. Starting with zero is obviously better than the
original.

Acked-by: Dan Carpenter <[email protected]>

regards,
dan carpenter