2016-04-19 13:44:30

by Krzysztof Kozlowski

[permalink] [raw]
Subject: [PATCH 1/2] crypto: s5p-sss - Fix use after free of copied input buffer in error path

The driver makes copies of memory (input or output scatterlists) if they
are not aligned. In s5p_aes_crypt_start() error path (on unsuccessful
initialization of output scatterlist), if input scatterlist was not
aligned, the driver first freed copied input memory and then unmapped it
from the device, instead of doing otherwise (unmap and then free).

This was wrong in two ways:
1. Freed pages were still mapped to the device.
2. The dma_unmap_sg() iterated over freed scatterlist structure.

The call to s5p_free_sg_cpy() in this error path is not needed because
the copied scatterlists will be freed by s5p_aes_complete().

Fixes: 9e4a1100a445 ("crypto: s5p-sss - Handle unaligned buffers")
Signed-off-by: Krzysztof Kozlowski <[email protected]>
---
drivers/crypto/s5p-sss.c | 1 -
1 file changed, 1 deletion(-)

diff --git a/drivers/crypto/s5p-sss.c b/drivers/crypto/s5p-sss.c
index 4f6d5b3ec418..b0484d4d68d9 100644
--- a/drivers/crypto/s5p-sss.c
+++ b/drivers/crypto/s5p-sss.c
@@ -577,7 +577,6 @@ static void s5p_aes_crypt_start(struct s5p_aes_dev *dev, unsigned long mode)
return;

outdata_error:
- s5p_free_sg_cpy(dev, &dev->sg_src_cpy);
s5p_unset_indata(dev);

indata_error:
--
1.9.1


2016-04-19 13:44:12

by Krzysztof Kozlowski

[permalink] [raw]
Subject: [PATCH 2/2] crypto: s5p-sss - Remove useless hash interrupt handler

Beside regular feed control interrupt, the driver requires also hash
interrupt for older SoCs (samsung,s5pv210-secss). However after
requesting it, the interrupt handler isn't doing anything with it, not
even clearing the hash interrupt bit.

Driver does not provide hash functions so it is safe to remove the hash
interrupt related code and to not require the interrupt in Device Tree.

Signed-off-by: Krzysztof Kozlowski <[email protected]>
---
.../devicetree/bindings/crypto/samsung-sss.txt | 6 ++--
drivers/crypto/s5p-sss.c | 34 ++++------------------
2 files changed, 8 insertions(+), 32 deletions(-)

diff --git a/Documentation/devicetree/bindings/crypto/samsung-sss.txt b/Documentation/devicetree/bindings/crypto/samsung-sss.txt
index a6dafa83c6df..7a5ca56683cc 100644
--- a/Documentation/devicetree/bindings/crypto/samsung-sss.txt
+++ b/Documentation/devicetree/bindings/crypto/samsung-sss.txt
@@ -23,10 +23,8 @@ Required properties:
- "samsung,exynos4210-secss" for Exynos4210, Exynos4212, Exynos4412, Exynos5250,
Exynos5260 and Exynos5420 SoCs.
- reg : Offset and length of the register set for the module
-- interrupts : interrupt specifiers of SSS module interrupts, should contain
- following entries:
- - first : feed control interrupt (required for all variants),
- - second : hash interrupt (required only for samsung,s5pv210-secss).
+- interrupts : interrupt specifiers of SSS module interrupts (one feed
+ control interrupt).

- clocks : list of clock phandle and specifier pairs for all clocks listed in
clock-names property.
diff --git a/drivers/crypto/s5p-sss.c b/drivers/crypto/s5p-sss.c
index b0484d4d68d9..71ca6a5d636d 100644
--- a/drivers/crypto/s5p-sss.c
+++ b/drivers/crypto/s5p-sss.c
@@ -149,7 +149,6 @@

/**
* struct samsung_aes_variant - platform specific SSS driver data
- * @has_hash_irq: true if SSS module uses hash interrupt, false otherwise
* @aes_offset: AES register offset from SSS module's base.
*
* Specifies platform specific configuration of SSS module.
@@ -157,7 +156,6 @@
* expansion of its usage.
*/
struct samsung_aes_variant {
- bool has_hash_irq;
unsigned int aes_offset;
};

@@ -178,7 +176,6 @@ struct s5p_aes_dev {
struct clk *clk;
void __iomem *ioaddr;
void __iomem *aes_ioaddr;
- int irq_hash;
int irq_fc;

struct ablkcipher_request *req;
@@ -201,12 +198,10 @@ struct s5p_aes_dev {
static struct s5p_aes_dev *s5p_dev;

static const struct samsung_aes_variant s5p_aes_data = {
- .has_hash_irq = true,
.aes_offset = 0x4000,
};

static const struct samsung_aes_variant exynos_aes_data = {
- .has_hash_irq = false,
.aes_offset = 0x200,
};

@@ -421,15 +416,13 @@ static irqreturn_t s5p_aes_interrupt(int irq, void *dev_id)

spin_lock_irqsave(&dev->lock, flags);

- if (irq == dev->irq_fc) {
- status = SSS_READ(dev, FCINTSTAT);
- if (status & SSS_FCINTSTAT_BRDMAINT)
- s5p_aes_rx(dev);
- if (status & SSS_FCINTSTAT_BTDMAINT)
- s5p_aes_tx(dev);
+ status = SSS_READ(dev, FCINTSTAT);
+ if (status & SSS_FCINTSTAT_BRDMAINT)
+ s5p_aes_rx(dev);
+ if (status & SSS_FCINTSTAT_BTDMAINT)
+ s5p_aes_tx(dev);

- SSS_WRITE(dev, FCINTPEND, status);
- }
+ SSS_WRITE(dev, FCINTPEND, status);

spin_unlock_irqrestore(&dev->lock, flags);

@@ -795,21 +788,6 @@ static int s5p_aes_probe(struct platform_device *pdev)
goto err_irq;
}

- if (variant->has_hash_irq) {
- pdata->irq_hash = platform_get_irq(pdev, 1);
- if (pdata->irq_hash < 0) {
- err = pdata->irq_hash;
- dev_warn(dev, "hash interrupt is not available.\n");
- goto err_irq;
- }
- err = devm_request_irq(dev, pdata->irq_hash, s5p_aes_interrupt,
- IRQF_SHARED, pdev->name, pdev);
- if (err < 0) {
- dev_warn(dev, "hash interrupt is not available.\n");
- goto err_irq;
- }
- }
-
pdata->busy = false;
pdata->variant = variant;
pdata->dev = dev;
--
1.9.1

2016-04-20 09:59:29

by Herbert Xu

[permalink] [raw]
Subject: Re: [PATCH 1/2] crypto: s5p-sss - Fix use after free of copied input buffer in error path

On Tue, Apr 19, 2016 at 03:44:11PM +0200, Krzysztof Kozlowski wrote:
> The driver makes copies of memory (input or output scatterlists) if they
> are not aligned. In s5p_aes_crypt_start() error path (on unsuccessful
> initialization of output scatterlist), if input scatterlist was not
> aligned, the driver first freed copied input memory and then unmapped it
> from the device, instead of doing otherwise (unmap and then free).
>
> This was wrong in two ways:
> 1. Freed pages were still mapped to the device.
> 2. The dma_unmap_sg() iterated over freed scatterlist structure.
>
> The call to s5p_free_sg_cpy() in this error path is not needed because
> the copied scatterlists will be freed by s5p_aes_complete().
>
> Fixes: 9e4a1100a445 ("crypto: s5p-sss - Handle unaligned buffers")
> Signed-off-by: Krzysztof Kozlowski <[email protected]>

Both applied.
--
Email: Herbert Xu <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

2016-04-20 10:03:48

by Vladimir Zapolskiy

[permalink] [raw]
Subject: Re: [PATCH 1/2] crypto: s5p-sss - Fix use after free of copied input buffer in error path

Hi Krzysztof,

On 19.04.2016 16:44, Krzysztof Kozlowski wrote:
> The driver makes copies of memory (input or output scatterlists) if they
> are not aligned. In s5p_aes_crypt_start() error path (on unsuccessful
> initialization of output scatterlist), if input scatterlist was not
> aligned, the driver first freed copied input memory and then unmapped it
> from the device, instead of doing otherwise (unmap and then free).
>
> This was wrong in two ways:
> 1. Freed pages were still mapped to the device.
> 2. The dma_unmap_sg() iterated over freed scatterlist structure.
>
> The call to s5p_free_sg_cpy() in this error path is not needed because
> the copied scatterlists will be freed by s5p_aes_complete().
>
> Fixes: 9e4a1100a445 ("crypto: s5p-sss - Handle unaligned buffers")
> Signed-off-by: Krzysztof Kozlowski <[email protected]>

I see that Herbert have just applied the changes, but anyway I reviewed
them and they are good in my opinion.

Acked-by: Vladimir Zapolskiy <[email protected]>

> ---
> drivers/crypto/s5p-sss.c | 1 -
> 1 file changed, 1 deletion(-)
>
> diff --git a/drivers/crypto/s5p-sss.c b/drivers/crypto/s5p-sss.c
> index 4f6d5b3ec418..b0484d4d68d9 100644
> --- a/drivers/crypto/s5p-sss.c
> +++ b/drivers/crypto/s5p-sss.c
> @@ -577,7 +577,6 @@ static void s5p_aes_crypt_start(struct s5p_aes_dev *dev, unsigned long mode)
> return;
>
> outdata_error:
> - s5p_free_sg_cpy(dev, &dev->sg_src_cpy);
> s5p_unset_indata(dev);
>
> indata_error:
>

--
With best wishes,
Vladimir

2016-04-20 10:04:02

by Vladimir Zapolskiy

[permalink] [raw]
Subject: Re: [PATCH 2/2] crypto: s5p-sss - Remove useless hash interrupt handler

On 19.04.2016 16:44, Krzysztof Kozlowski wrote:
> Beside regular feed control interrupt, the driver requires also hash
> interrupt for older SoCs (samsung,s5pv210-secss). However after
> requesting it, the interrupt handler isn't doing anything with it, not
> even clearing the hash interrupt bit.
>
> Driver does not provide hash functions so it is safe to remove the hash
> interrupt related code and to not require the interrupt in Device Tree.
>
> Signed-off-by: Krzysztof Kozlowski <[email protected]>

Thank you for the change.

Acked-by: Vladimir Zapolskiy <[email protected]>

> ---
> .../devicetree/bindings/crypto/samsung-sss.txt | 6 ++--
> drivers/crypto/s5p-sss.c | 34 ++++------------------
> 2 files changed, 8 insertions(+), 32 deletions(-)
>

---
With best wishes,
Vladimir

2016-04-21 15:24:56

by Rob Herring

[permalink] [raw]
Subject: Re: [PATCH 2/2] crypto: s5p-sss - Remove useless hash interrupt handler

On Tue, Apr 19, 2016 at 03:44:12PM +0200, Krzysztof Kozlowski wrote:
> Beside regular feed control interrupt, the driver requires also hash
> interrupt for older SoCs (samsung,s5pv210-secss). However after
> requesting it, the interrupt handler isn't doing anything with it, not
> even clearing the hash interrupt bit.
>
> Driver does not provide hash functions so it is safe to remove the hash
> interrupt related code and to not require the interrupt in Device Tree.

NAK for the DT change. If the h/w has a 2nd interrupt, then it should be
defined. Whether the driver uses it or not is a driver problem.

Rob