[dpdk-dev] cryptodev: fix segmentation fault

  This patch fix problem in function rte_cryptodev_devices_get().
Program received signal SIGSEGV, Segmentation fault.
It also rework the function to use correct types and clean up visibility.

Fixes: 38227c0e3ad2 ("cryptodev: retrieve device info")

Signed-off-by: Slawomir Mrozowicz <slawomirx.mrozowicz@intel.com>
---
 lib/librte_cryptodev/rte_cryptodev.c | 35 +++++++++++++++--------------------
 lib/librte_cryptodev/rte_cryptodev.h |  4 ++--
 2 files changed, 17 insertions(+), 22 deletions(-)
  

Hi Slawomir,

> -----Original Message-----
> From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Slawomir
> Mrozowicz
> Sent: Friday, February 03, 2017 9:18 AM
> To: Doherty, Declan
> Cc: dev@dpdk.org; Mrozowicz, SlawomirX
> Subject: [dpdk-dev] [PATCH] cryptodev: fix segmentation fault
> 
> This patch fix problem in function rte_cryptodev_devices_get().
> Program received signal SIGSEGV, Segmentation fault.
> It also rework the function to use correct types and clean up visibility.

Are you fixing any Coverity issues with this patch?
If so, state it here.

> 
> Fixes: 38227c0e3ad2 ("cryptodev: retrieve device info")
> 
> Signed-off-by: Slawomir Mrozowicz <slawomirx.mrozowicz@intel.com>
> ---
>  lib/librte_cryptodev/rte_cryptodev.c | 35 +++++++++++++++-------------------
> -
>  lib/librte_cryptodev/rte_cryptodev.h |  4 ++--
>  2 files changed, 17 insertions(+), 22 deletions(-)
> 
> diff --git a/lib/librte_cryptodev/rte_cryptodev.c
> b/lib/librte_cryptodev/rte_cryptodev.c
> index e557e77..10a59ba 100644
> --- a/lib/librte_cryptodev/rte_cryptodev.c
> +++ b/lib/librte_cryptodev/rte_cryptodev.c
> @@ -482,34 +482,29 @@ rte_cryptodev_count_devtype(enum
> rte_cryptodev_type type)
>  	return dev_count;
>  }
> 
> -int
> +uint8_t
>  rte_cryptodev_devices_get(const char *dev_name, uint8_t *devices,
> -	uint8_t nb_devices)
> +	const uint8_t nb_devices)

I don't think it is necessary to use "const" here. Even if you change it in the function,
the value wouldn't be changed outside.

>  {
> -	uint8_t i, cmp, count = 0;
> -	struct rte_cryptodev **devs = &rte_cryptodev_globals->devs;
> -	struct rte_device *dev;
> -
> -	for (i = 0; i < rte_cryptodev_globals->max_devs && count <
> nb_devices;
> -			i++) {
> +	uint8_t i, count = 0;
> +	struct rte_cryptodev *devs = rte_cryptodev_globals->devs;
> +	uint8_t max_devs = rte_cryptodev_globals->max_devs;
> 
> -		if ((*devs + i)
> -				&& (*devs + i)->attached ==
> -						RTE_CRYPTODEV_ATTACHED)
> {
> +	for (i = 0; i < max_devs && count < nb_devices;	i++) {
> 
> -			dev = (*devs + i)->device;
> +		if (devs[i].attached == RTE_CRYPTODEV_ATTACHED) {
> +			const struct rte_cryptodev_driver *drv =
> devs[i].driver;
> +			int cmp = -1;

Is it necessary to initialize this value to -1? cmp is going to be overwritten always.
> 
> -			if (dev)
> -				cmp = strncmp(dev->driver->name,
> -						dev_name,
> -						strlen(dev_name));
> +			if (drv)
> +				cmp = strncmp(drv->pci_drv.driver.name,
> +						dev_name,
> strlen(dev_name));
>  			else
> -				cmp = strncmp((*devs + i)->data->name,
> -						dev_name,
> -						strlen(dev_name));
> +				cmp = strncmp(devs[i].data->name,
> +						dev_name,
> strlen(dev_name));
> 
>  			if (cmp == 0)
> -				devices[count++] = (*devs + i)->data-
> >dev_id;
> +				devices[count++] = devs[i].data->dev_id;
>  		}
>  	}
>