* [PATCH] rtc: s5m: Remove VLA usage
@ 2018-03-08 17:20 Gustavo A. R. Silva
2018-03-08 17:58 ` Kees Cook
0 siblings, 1 reply; 5+ messages in thread
From: Gustavo A. R. Silva @ 2018-03-08 17:20 UTC (permalink / raw)
To: Sangbeom Kim, Krzysztof Kozlowski, Bartlomiej Zolnierkiewicz,
Alessandro Zummo, Alexandre Belloni
Cc: linux-kernel, linux-samsung-soc, linux-rtc, kernel-hardening,
Kees Cook, Gustavo A. R. Silva
In preparation to enabling -Wvla, remove VLAs and replace them
with fixed-length arrays instead.
>From a security viewpoint, the use of Variable Length Arrays can be
a vector for stack overflow attacks. Also, in general, as the code
evolves it is easy to lose track of how big a VLA can get. Thus, we
can end up having segfaults that are hard to debug.
Also, fixed as part of the directive to remove all VLAs from
the kernel: https://lkml.org/lkml/2018/3/7/621
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
---
drivers/rtc/rtc-s5m.c | 15 +++++++++------
1 file changed, 9 insertions(+), 6 deletions(-)
diff --git a/drivers/rtc/rtc-s5m.c b/drivers/rtc/rtc-s5m.c
index 6deae10..2b5f4f7 100644
--- a/drivers/rtc/rtc-s5m.c
+++ b/drivers/rtc/rtc-s5m.c
@@ -38,6 +38,9 @@
*/
#define UDR_READ_RETRY_CNT 5
+/* Maximum number of registers for setting time/alarm0/alarm1 */
+#define MAX_NUM_TIME_REGS 8
+
/*
* Registers used by the driver which are different between chipsets.
*
@@ -367,7 +370,7 @@ static void s5m8763_tm_to_data(struct rtc_time *tm, u8 *data)
static int s5m_rtc_read_time(struct device *dev, struct rtc_time *tm)
{
struct s5m_rtc_info *info = dev_get_drvdata(dev);
- u8 data[info->regs->regs_count];
+ u8 data[MAX_NUM_TIME_REGS];
int ret;
if (info->regs->read_time_udr_mask) {
@@ -413,7 +416,7 @@ static int s5m_rtc_read_time(struct device *dev, struct rtc_time *tm)
static int s5m_rtc_set_time(struct device *dev, struct rtc_time *tm)
{
struct s5m_rtc_info *info = dev_get_drvdata(dev);
- u8 data[info->regs->regs_count];
+ u8 data[MAX_NUM_TIME_REGS];
int ret = 0;
switch (info->device_type) {
@@ -450,7 +453,7 @@ static int s5m_rtc_set_time(struct device *dev, struct rtc_time *tm)
static int s5m_rtc_read_alarm(struct device *dev, struct rtc_wkalrm *alrm)
{
struct s5m_rtc_info *info = dev_get_drvdata(dev);
- u8 data[info->regs->regs_count];
+ u8 data[MAX_NUM_TIME_REGS];
unsigned int val;
int ret, i;
@@ -500,7 +503,7 @@ static int s5m_rtc_read_alarm(struct device *dev, struct rtc_wkalrm *alrm)
static int s5m_rtc_stop_alarm(struct s5m_rtc_info *info)
{
- u8 data[info->regs->regs_count];
+ u8 data[MAX_NUM_TIME_REGS];
int ret, i;
struct rtc_time tm;
@@ -545,7 +548,7 @@ static int s5m_rtc_stop_alarm(struct s5m_rtc_info *info)
static int s5m_rtc_start_alarm(struct s5m_rtc_info *info)
{
int ret;
- u8 data[info->regs->regs_count];
+ u8 data[MAX_NUM_TIME_REGS];
u8 alarm0_conf;
struct rtc_time tm;
@@ -598,7 +601,7 @@ static int s5m_rtc_start_alarm(struct s5m_rtc_info *info)
static int s5m_rtc_set_alarm(struct device *dev, struct rtc_wkalrm *alrm)
{
struct s5m_rtc_info *info = dev_get_drvdata(dev);
- u8 data[info->regs->regs_count];
+ u8 data[MAX_NUM_TIME_REGS];
int ret;
switch (info->device_type) {
--
2.7.4
^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [PATCH] rtc: s5m: Remove VLA usage
2018-03-08 17:20 [PATCH] rtc: s5m: Remove VLA usage Gustavo A. R. Silva
@ 2018-03-08 17:58 ` Kees Cook
2018-03-08 18:03 ` Gustavo A. R. Silva
0 siblings, 1 reply; 5+ messages in thread
From: Kees Cook @ 2018-03-08 17:58 UTC (permalink / raw)
To: Gustavo A. R. Silva
Cc: Sangbeom Kim, Krzysztof Kozlowski, Bartlomiej Zolnierkiewicz,
Alessandro Zummo, Alexandre Belloni, LKML, linux-samsung-soc,
linux-rtc, Kernel Hardening, Gustavo A. R. Silva
On Thu, Mar 8, 2018 at 9:20 AM, Gustavo A. R. Silva
<gustavo@embeddedor.com> wrote:
> In preparation to enabling -Wvla, remove VLAs and replace them
> with fixed-length arrays instead.
>
> From a security viewpoint, the use of Variable Length Arrays can be
> a vector for stack overflow attacks. Also, in general, as the code
> evolves it is easy to lose track of how big a VLA can get. Thus, we
> can end up having segfaults that are hard to debug.
>
> Also, fixed as part of the directive to remove all VLAs from
> the kernel: https://lkml.org/lkml/2018/3/7/621
>
> Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
> ---
> drivers/rtc/rtc-s5m.c | 15 +++++++++------
> 1 file changed, 9 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/rtc/rtc-s5m.c b/drivers/rtc/rtc-s5m.c
> index 6deae10..2b5f4f7 100644
> --- a/drivers/rtc/rtc-s5m.c
> +++ b/drivers/rtc/rtc-s5m.c
> @@ -38,6 +38,9 @@
> */
> #define UDR_READ_RETRY_CNT 5
>
> +/* Maximum number of registers for setting time/alarm0/alarm1 */
> +#define MAX_NUM_TIME_REGS 8
I would adjust the various const struct s5m_rtc_reg_config's
.regs_count to be represented by this new define, so the stack and the
structures stay in sync. Something like:
static const struct s5m_rtc_reg_config s2mps13_rtc_regs = {
.regs_count = MAX_NUM_TIME_REGS - 1,
?
-Kees
> +
> /*
> * Registers used by the driver which are different between chipsets.
> *
> @@ -367,7 +370,7 @@ static void s5m8763_tm_to_data(struct rtc_time *tm, u8 *data)
> static int s5m_rtc_read_time(struct device *dev, struct rtc_time *tm)
> {
> struct s5m_rtc_info *info = dev_get_drvdata(dev);
> - u8 data[info->regs->regs_count];
> + u8 data[MAX_NUM_TIME_REGS];
> int ret;
>
> if (info->regs->read_time_udr_mask) {
> @@ -413,7 +416,7 @@ static int s5m_rtc_read_time(struct device *dev, struct rtc_time *tm)
> static int s5m_rtc_set_time(struct device *dev, struct rtc_time *tm)
> {
> struct s5m_rtc_info *info = dev_get_drvdata(dev);
> - u8 data[info->regs->regs_count];
> + u8 data[MAX_NUM_TIME_REGS];
> int ret = 0;
>
> switch (info->device_type) {
> @@ -450,7 +453,7 @@ static int s5m_rtc_set_time(struct device *dev, struct rtc_time *tm)
> static int s5m_rtc_read_alarm(struct device *dev, struct rtc_wkalrm *alrm)
> {
> struct s5m_rtc_info *info = dev_get_drvdata(dev);
> - u8 data[info->regs->regs_count];
> + u8 data[MAX_NUM_TIME_REGS];
> unsigned int val;
> int ret, i;
>
> @@ -500,7 +503,7 @@ static int s5m_rtc_read_alarm(struct device *dev, struct rtc_wkalrm *alrm)
>
> static int s5m_rtc_stop_alarm(struct s5m_rtc_info *info)
> {
> - u8 data[info->regs->regs_count];
> + u8 data[MAX_NUM_TIME_REGS];
> int ret, i;
> struct rtc_time tm;
>
> @@ -545,7 +548,7 @@ static int s5m_rtc_stop_alarm(struct s5m_rtc_info *info)
> static int s5m_rtc_start_alarm(struct s5m_rtc_info *info)
> {
> int ret;
> - u8 data[info->regs->regs_count];
> + u8 data[MAX_NUM_TIME_REGS];
> u8 alarm0_conf;
> struct rtc_time tm;
>
> @@ -598,7 +601,7 @@ static int s5m_rtc_start_alarm(struct s5m_rtc_info *info)
> static int s5m_rtc_set_alarm(struct device *dev, struct rtc_wkalrm *alrm)
> {
> struct s5m_rtc_info *info = dev_get_drvdata(dev);
> - u8 data[info->regs->regs_count];
> + u8 data[MAX_NUM_TIME_REGS];
> int ret;
>
> switch (info->device_type) {
> --
> 2.7.4
>
--
Kees Cook
Pixel Security
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] rtc: s5m: Remove VLA usage
2018-03-08 17:58 ` Kees Cook
@ 2018-03-08 18:03 ` Gustavo A. R. Silva
2018-03-09 13:04 ` Krzysztof Kozlowski
0 siblings, 1 reply; 5+ messages in thread
From: Gustavo A. R. Silva @ 2018-03-08 18:03 UTC (permalink / raw)
To: Kees Cook, Gustavo A. R. Silva
Cc: Sangbeom Kim, Krzysztof Kozlowski, Bartlomiej Zolnierkiewicz,
Alessandro Zummo, Alexandre Belloni, LKML, linux-samsung-soc,
linux-rtc, Kernel Hardening
On 03/08/2018 11:58 AM, Kees Cook wrote:
> On Thu, Mar 8, 2018 at 9:20 AM, Gustavo A. R. Silva
> <gustavo@embeddedor.com> wrote:
>> In preparation to enabling -Wvla, remove VLAs and replace them
>> with fixed-length arrays instead.
>>
>> From a security viewpoint, the use of Variable Length Arrays can be
>> a vector for stack overflow attacks. Also, in general, as the code
>> evolves it is easy to lose track of how big a VLA can get. Thus, we
>> can end up having segfaults that are hard to debug.
>>
>> Also, fixed as part of the directive to remove all VLAs from
>> the kernel: https://lkml.org/lkml/2018/3/7/621
>>
>> Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
>> ---
>> drivers/rtc/rtc-s5m.c | 15 +++++++++------
>> 1 file changed, 9 insertions(+), 6 deletions(-)
>>
>> diff --git a/drivers/rtc/rtc-s5m.c b/drivers/rtc/rtc-s5m.c
>> index 6deae10..2b5f4f7 100644
>> --- a/drivers/rtc/rtc-s5m.c
>> +++ b/drivers/rtc/rtc-s5m.c
>> @@ -38,6 +38,9 @@
>> */
>> #define UDR_READ_RETRY_CNT 5
>>
>> +/* Maximum number of registers for setting time/alarm0/alarm1 */
>> +#define MAX_NUM_TIME_REGS 8
>
> I would adjust the various const struct s5m_rtc_reg_config's
> .regs_count to be represented by this new define, so the stack and the
> structures stay in sync. Something like:
>
> static const struct s5m_rtc_reg_config s2mps13_rtc_regs = {
> .regs_count = MAX_NUM_TIME_REGS - 1,
>
> ?
>
Yep. I thought about that and decided to wait for some feedback first.
But yeah, I think is that'd be a good change.
--
Gustavo
> -Kees
>
>> +
>> /*
>> * Registers used by the driver which are different between chipsets.
>> *
>> @@ -367,7 +370,7 @@ static void s5m8763_tm_to_data(struct rtc_time *tm, u8 *data)
>> static int s5m_rtc_read_time(struct device *dev, struct rtc_time *tm)
>> {
>> struct s5m_rtc_info *info = dev_get_drvdata(dev);
>> - u8 data[info->regs->regs_count];
>> + u8 data[MAX_NUM_TIME_REGS];
>> int ret;
>>
>> if (info->regs->read_time_udr_mask) {
>> @@ -413,7 +416,7 @@ static int s5m_rtc_read_time(struct device *dev, struct rtc_time *tm)
>> static int s5m_rtc_set_time(struct device *dev, struct rtc_time *tm)
>> {
>> struct s5m_rtc_info *info = dev_get_drvdata(dev);
>> - u8 data[info->regs->regs_count];
>> + u8 data[MAX_NUM_TIME_REGS];
>> int ret = 0;
>>
>> switch (info->device_type) {
>> @@ -450,7 +453,7 @@ static int s5m_rtc_set_time(struct device *dev, struct rtc_time *tm)
>> static int s5m_rtc_read_alarm(struct device *dev, struct rtc_wkalrm *alrm)
>> {
>> struct s5m_rtc_info *info = dev_get_drvdata(dev);
>> - u8 data[info->regs->regs_count];
>> + u8 data[MAX_NUM_TIME_REGS];
>> unsigned int val;
>> int ret, i;
>>
>> @@ -500,7 +503,7 @@ static int s5m_rtc_read_alarm(struct device *dev, struct rtc_wkalrm *alrm)
>>
>> static int s5m_rtc_stop_alarm(struct s5m_rtc_info *info)
>> {
>> - u8 data[info->regs->regs_count];
>> + u8 data[MAX_NUM_TIME_REGS];
>> int ret, i;
>> struct rtc_time tm;
>>
>> @@ -545,7 +548,7 @@ static int s5m_rtc_stop_alarm(struct s5m_rtc_info *info)
>> static int s5m_rtc_start_alarm(struct s5m_rtc_info *info)
>> {
>> int ret;
>> - u8 data[info->regs->regs_count];
>> + u8 data[MAX_NUM_TIME_REGS];
>> u8 alarm0_conf;
>> struct rtc_time tm;
>>
>> @@ -598,7 +601,7 @@ static int s5m_rtc_start_alarm(struct s5m_rtc_info *info)
>> static int s5m_rtc_set_alarm(struct device *dev, struct rtc_wkalrm *alrm)
>> {
>> struct s5m_rtc_info *info = dev_get_drvdata(dev);
>> - u8 data[info->regs->regs_count];
>> + u8 data[MAX_NUM_TIME_REGS];
>> int ret;
>>
>> switch (info->device_type) {
>> --
>> 2.7.4
>>
>
>
>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] rtc: s5m: Remove VLA usage
2018-03-08 18:03 ` Gustavo A. R. Silva
@ 2018-03-09 13:04 ` Krzysztof Kozlowski
2018-03-10 4:34 ` Gustavo A. R. Silva
0 siblings, 1 reply; 5+ messages in thread
From: Krzysztof Kozlowski @ 2018-03-09 13:04 UTC (permalink / raw)
To: Gustavo A. R. Silva
Cc: Kees Cook, Gustavo A. R. Silva, Sangbeom Kim,
Bartlomiej Zolnierkiewicz, Alessandro Zummo, Alexandre Belloni,
LKML, linux-samsung-soc, linux-rtc, Kernel Hardening
On Thu, Mar 8, 2018 at 7:03 PM, Gustavo A. R. Silva
<garsilva@embeddedor.com> wrote:
>
>
> On 03/08/2018 11:58 AM, Kees Cook wrote:
>>
>> On Thu, Mar 8, 2018 at 9:20 AM, Gustavo A. R. Silva
>> <gustavo@embeddedor.com> wrote:
>>>
>>> In preparation to enabling -Wvla, remove VLAs and replace them
>>> with fixed-length arrays instead.
>>>
>>> From a security viewpoint, the use of Variable Length Arrays can be
>>> a vector for stack overflow attacks. Also, in general, as the code
>>> evolves it is easy to lose track of how big a VLA can get. Thus, we
>>> can end up having segfaults that are hard to debug.
>>>
>>> Also, fixed as part of the directive to remove all VLAs from
>>> the kernel: https://lkml.org/lkml/2018/3/7/621
>>>
>>> Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
>>> ---
>>> drivers/rtc/rtc-s5m.c | 15 +++++++++------
>>> 1 file changed, 9 insertions(+), 6 deletions(-)
>>>
>>> diff --git a/drivers/rtc/rtc-s5m.c b/drivers/rtc/rtc-s5m.c
>>> index 6deae10..2b5f4f7 100644
>>> --- a/drivers/rtc/rtc-s5m.c
>>> +++ b/drivers/rtc/rtc-s5m.c
>>> @@ -38,6 +38,9 @@
>>> */
>>> #define UDR_READ_RETRY_CNT 5
>>>
>>> +/* Maximum number of registers for setting time/alarm0/alarm1 */
>>> +#define MAX_NUM_TIME_REGS 8
>>
>>
>> I would adjust the various const struct s5m_rtc_reg_config's
>> .regs_count to be represented by this new define, so the stack and the
>> structures stay in sync. Something like:
>>
>> static const struct s5m_rtc_reg_config s2mps13_rtc_regs = {
>> .regs_count = MAX_NUM_TIME_REGS - 1,
>>
>> ?
>>
>
> Yep. I thought about that and decided to wait for some feedback first. But
> yeah, I think is that'd be a good change.
Define and these assignments should be somehow connected with enum
defining the offsets for data[] (from
include/linux/mfd/samsung/rtc.h). Otherwise we define the same in two
places. The enum could be itself (in separate patch) moved to the
driver because it is meaningless for others.
Best regards,
Krzysztof
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] rtc: s5m: Remove VLA usage
2018-03-09 13:04 ` Krzysztof Kozlowski
@ 2018-03-10 4:34 ` Gustavo A. R. Silva
0 siblings, 0 replies; 5+ messages in thread
From: Gustavo A. R. Silva @ 2018-03-10 4:34 UTC (permalink / raw)
To: Krzysztof Kozlowski
Cc: Kees Cook, Gustavo A. R. Silva, Sangbeom Kim,
Bartlomiej Zolnierkiewicz, Alessandro Zummo, Alexandre Belloni,
LKML, linux-samsung-soc, linux-rtc, Kernel Hardening
Hi Krzysztof,
On 03/09/2018 07:04 AM, Krzysztof Kozlowski wrote:
> On Thu, Mar 8, 2018 at 7:03 PM, Gustavo A. R. Silva
> <garsilva@embeddedor.com> wrote:
>>
>>
>> On 03/08/2018 11:58 AM, Kees Cook wrote:
>>>
>>> On Thu, Mar 8, 2018 at 9:20 AM, Gustavo A. R. Silva
>>> <gustavo@embeddedor.com> wrote:
>>>>
>>>> In preparation to enabling -Wvla, remove VLAs and replace them
>>>> with fixed-length arrays instead.
>>>>
>>>> From a security viewpoint, the use of Variable Length Arrays can be
>>>> a vector for stack overflow attacks. Also, in general, as the code
>>>> evolves it is easy to lose track of how big a VLA can get. Thus, we
>>>> can end up having segfaults that are hard to debug.
>>>>
>>>> Also, fixed as part of the directive to remove all VLAs from
>>>> the kernel: https://lkml.org/lkml/2018/3/7/621
>>>>
>>>> Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
>>>> ---
>>>> drivers/rtc/rtc-s5m.c | 15 +++++++++------
>>>> 1 file changed, 9 insertions(+), 6 deletions(-)
>>>>
>>>> diff --git a/drivers/rtc/rtc-s5m.c b/drivers/rtc/rtc-s5m.c
>>>> index 6deae10..2b5f4f7 100644
>>>> --- a/drivers/rtc/rtc-s5m.c
>>>> +++ b/drivers/rtc/rtc-s5m.c
>>>> @@ -38,6 +38,9 @@
>>>> */
>>>> #define UDR_READ_RETRY_CNT 5
>>>>
>>>> +/* Maximum number of registers for setting time/alarm0/alarm1 */
>>>> +#define MAX_NUM_TIME_REGS 8
>>>
>>>
>>> I would adjust the various const struct s5m_rtc_reg_config's
>>> .regs_count to be represented by this new define, so the stack and the
>>> structures stay in sync. Something like:
>>>
>>> static const struct s5m_rtc_reg_config s2mps13_rtc_regs = {
>>> .regs_count = MAX_NUM_TIME_REGS - 1,
>>>
>>> ?
>>>
>>
>> Yep. I thought about that and decided to wait for some feedback first. But
>> yeah, I think is that'd be a good change.
>
> Define and these assignments should be somehow connected with enum
> defining the offsets for data[] (from
> include/linux/mfd/samsung/rtc.h). Otherwise we define the same in two
> places. The enum could be itself (in separate patch) moved to the
> driver because it is meaningless for others.
>
I got it.
I'll move the enum to rtc-s5m.c and add RTC_MAX_NUM_TIME_REGS at the end
of it. I'll send a patch series for this.
Thanks for the feedback.
--
Gustavo
> Best regards,
> Krzysztof
>
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2018-03-10 4:34 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-03-08 17:20 [PATCH] rtc: s5m: Remove VLA usage Gustavo A. R. Silva
2018-03-08 17:58 ` Kees Cook
2018-03-08 18:03 ` Gustavo A. R. Silva
2018-03-09 13:04 ` Krzysztof Kozlowski
2018-03-10 4:34 ` Gustavo A. R. Silva
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).