2017-12-13 22:45:47

by Gustavo A. R. Silva

[permalink] [raw]
Subject: [PATCH] ACPI / CPPC: Fix negative array index read in cppc_set_perf

If pcc_ss_id is less than 0, there is a negative array index read
before verifying pcc_ss_id is not a negative value.

Fix this by removing the code that triggers this issue.

Notice that this code is already properly placed after the check
on pcc_ss_id at line 1182: pcc_ss_data = pcc_data[pcc_ss_id];

Addresses-Coverity-ID: 1426090 ("Negative array index read")
Fixes: 1ecbd7170d65 ("ACPI / CPPC: Fix KASAN global out of bounds warning")
Signed-off-by: Gustavo A. R. Silva <[email protected]>
---
drivers/acpi/cppc_acpi.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/acpi/cppc_acpi.c b/drivers/acpi/cppc_acpi.c
index 30e84cc..06ea474 100644
--- a/drivers/acpi/cppc_acpi.c
+++ b/drivers/acpi/cppc_acpi.c
@@ -1171,7 +1171,7 @@ int cppc_set_perf(int cpu, struct cppc_perf_ctrls *perf_ctrls)
struct cpc_desc *cpc_desc = per_cpu(cpc_desc_ptr, cpu);
struct cpc_register_resource *desired_reg;
int pcc_ss_id = per_cpu(cpu_pcc_subspace_idx, cpu);
- struct cppc_pcc_data *pcc_ss_data = pcc_data[pcc_ss_id];
+ struct cppc_pcc_data *pcc_ss_data;
int ret = 0;

if (!cpc_desc || pcc_ss_id < 0) {
--
2.7.4


2017-12-15 18:18:40

by George Cherian

[permalink] [raw]
Subject: Re: [PATCH] ACPI / CPPC: Fix negative array index read in cppc_set_perf



On Thursday 14 December 2017 04:15 AM, Gustavo A. R. Silva wrote:
> If pcc_ss_id is less than 0, there is a negative array index read
> before verifying pcc_ss_id is not a negative value.
>
> Fix this by removing the code that triggers this issue.
>
> Notice that this code is already properly placed after the check
> on pcc_ss_id at line 1182: pcc_ss_data = pcc_data[pcc_ss_id];
>
> Addresses-Coverity-ID: 1426090 ("Negative array index read")
> Fixes: 1ecbd7170d65 ("ACPI / CPPC: Fix KASAN global out of bounds warning")

Reviewed-by: George Cherian <[email protected]>


> Signed-off-by: Gustavo A. R. Silva <[email protected]>
> ---
> drivers/acpi/cppc_acpi.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/acpi/cppc_acpi.c b/drivers/acpi/cppc_acpi.c
> index 30e84cc..06ea474 100644
> --- a/drivers/acpi/cppc_acpi.c
> +++ b/drivers/acpi/cppc_acpi.c
> @@ -1171,7 +1171,7 @@ int cppc_set_perf(int cpu, struct cppc_perf_ctrls *perf_ctrls)
> struct cpc_desc *cpc_desc = per_cpu(cpc_desc_ptr, cpu);
> struct cpc_register_resource *desired_reg;
> int pcc_ss_id = per_cpu(cpu_pcc_subspace_idx, cpu);
> - struct cppc_pcc_data *pcc_ss_data = pcc_data[pcc_ss_id];
> + struct cppc_pcc_data *pcc_ss_data;
> int ret = 0;
>
> if (!cpc_desc || pcc_ss_id < 0) {

2017-12-16 01:48:19

by Rafael J. Wysocki

[permalink] [raw]
Subject: Re: [PATCH] ACPI / CPPC: Fix negative array index read in cppc_set_perf

On Wednesday, December 13, 2017 11:45:41 PM CET Gustavo A. R. Silva wrote:
> If pcc_ss_id is less than 0, there is a negative array index read
> before verifying pcc_ss_id is not a negative value.
>
> Fix this by removing the code that triggers this issue.
>
> Notice that this code is already properly placed after the check
> on pcc_ss_id at line 1182: pcc_ss_data = pcc_data[pcc_ss_id];
>
> Addresses-Coverity-ID: 1426090 ("Negative array index read")
> Fixes: 1ecbd7170d65 ("ACPI / CPPC: Fix KASAN global out of bounds warning")
> Signed-off-by: Gustavo A. R. Silva <[email protected]>
> ---
> drivers/acpi/cppc_acpi.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/acpi/cppc_acpi.c b/drivers/acpi/cppc_acpi.c
> index 30e84cc..06ea474 100644
> --- a/drivers/acpi/cppc_acpi.c
> +++ b/drivers/acpi/cppc_acpi.c
> @@ -1171,7 +1171,7 @@ int cppc_set_perf(int cpu, struct cppc_perf_ctrls *perf_ctrls)
> struct cpc_desc *cpc_desc = per_cpu(cpc_desc_ptr, cpu);
> struct cpc_register_resource *desired_reg;
> int pcc_ss_id = per_cpu(cpu_pcc_subspace_idx, cpu);
> - struct cppc_pcc_data *pcc_ss_data = pcc_data[pcc_ss_id];
> + struct cppc_pcc_data *pcc_ss_data;
> int ret = 0;
>
> if (!cpc_desc || pcc_ss_id < 0) {
>

I applied a patch from Coling Ian King that did the same thing and was
posted eaelier.

Thanks,
Rafael