Re: [PATCH v4 13/15] firmware: qcom: tzmem: enable SHM Bridge support

[Andrew Halaney]

On Fri, Oct 13, 2023 at 01:48:41PM +0200, Bartosz Golaszewski wrote:
> From: Bartosz Golaszewski <bartosz.golaszewski@xxxxxxxxxx>
> 
> Add a new Kconfig option for selecting the SHM Bridge mode of operation
> for the TrustZone memory allocator.
> 
> If enabled at build-time, it will still be checked for availability at
> run-time. If the architecture doesn't support SHM Bridge, the allocator
> will work just like in the default mode.
> 
> Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@xxxxxxxxxx>
> ---
>  drivers/firmware/qcom/Kconfig      | 10 +++++
>  drivers/firmware/qcom/qcom_tzmem.c | 65 +++++++++++++++++++++++++++++-
>  2 files changed, 74 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/firmware/qcom/Kconfig b/drivers/firmware/qcom/Kconfig
> index 237da40de832..e01407e31ae4 100644
> --- a/drivers/firmware/qcom/Kconfig
> +++ b/drivers/firmware/qcom/Kconfig
> @@ -27,6 +27,16 @@ config QCOM_TZMEM_MODE_DEFAULT
>  	  Use the default allocator mode. The memory is page-aligned, non-cachable
>  	  and contiguous.
>  
> +config QCOM_TZMEM_MODE_SHMBRIDGE
> +	bool "SHM Bridge"
> +	help
> +	  Use Qualcomm Shared Memory Bridge. The memory has the same alignment as
> +	  in the 'Default' allocator but is also explicitly marked as an SHM Bridge
> +	  buffer.
> +
> +	  With this selected, all buffers passed to the TrustZone must be allocated
> +	  using the TZMem allocator or else the TrustZone will refuse to use them.
> +
>  endchoice
>  
>  config QCOM_SCM_DOWNLOAD_MODE_DEFAULT
> diff --git a/drivers/firmware/qcom/qcom_tzmem.c b/drivers/firmware/qcom/qcom_tzmem.c
> index 68ca59c5598e..6c4f29a00f26 100644
> --- a/drivers/firmware/qcom/qcom_tzmem.c
> +++ b/drivers/firmware/qcom/qcom_tzmem.c
> @@ -55,7 +55,70 @@ static void qcom_tzmem_cleanup_pool(struct qcom_tzmem_pool *pool)
>  
>  }
>  
> -#endif /* CONFIG_QCOM_TZMEM_MODE_DEFAULT */
> +#elif IS_ENABLED(CONFIG_QCOM_TZMEM_MODE_SHMBRIDGE)
> +
> +#include <linux/firmware/qcom/qcom_scm.h>
> +
> +#define QCOM_SHM_BRIDGE_NUM_VM_SHIFT 9
> +
> +static bool qcom_tzmem_using_shm_bridge;
> +
> +static int qcom_tzmem_init(void)
> +{
> +	int ret;
> +
> +	ret = qcom_scm_shm_bridge_enable();
> +	if (ret == -EOPNOTSUPP) {
> +		dev_info(qcom_tzmem_dev, "SHM Bridge not supported\n");
> +		ret = 0;
> +	}
> +
> +	if (!ret)
> +		qcom_tzmem_using_shm_bridge = true;
> +
> +	return ret;
> +}

Restating the comment we just concluded on in v3 (just to make it
obvious for other reviewers), the -EOPNOTSUPP
case here should probably return 0, not ret = 0, otherwise we'll try to
use SHM bridge on memory later even though it is not supported:

https://lore.kernel.org/linux-arm-msm/20231009153427.20951-1-brgl@xxxxxxxx/T/#m175bd35782cb5824820977bb3f799b32e32fb2d7

> +
> +static int qcom_tzmem_init_pool(struct qcom_tzmem_pool *pool)
> +{
> +	u64 pfn_and_ns_perm, ipfn_and_s_perm, size_and_flags, ns_perms;
> +	int ret;
> +
> +	if (!qcom_tzmem_using_shm_bridge)
> +		return 0;
> +
> +	ns_perms = (QCOM_SCM_PERM_WRITE | QCOM_SCM_PERM_READ);
> +	pfn_and_ns_perm = (u64)pool->pbase | ns_perms;
> +	ipfn_and_s_perm = (u64)pool->pbase | ns_perms;
> +	size_and_flags = pool->size | (1 << QCOM_SHM_BRIDGE_NUM_VM_SHIFT);
> +
> +	u64 *handle __free(kfree) = kzalloc(sizeof(*handle), GFP_KERNEL);
> +	if (!handle)
> +		return -ENOMEM;
> +
> +	ret = qcom_scm_shm_bridge_create(qcom_tzmem_dev, pfn_and_ns_perm,
> +					 ipfn_and_s_perm, size_and_flags,
> +					 QCOM_SCM_VMID_HLOS, handle);
> +	if (ret)
> +		return ret;
> +
> +	pool->priv = no_free_ptr(handle);
> +
> +	return 0;
> +}
> +
> +static void qcom_tzmem_cleanup_pool(struct qcom_tzmem_pool *pool)
> +{
> +	u64 *handle = pool->priv;
> +
> +	if (!qcom_tzmem_using_shm_bridge)
> +		return;
> +
> +	qcom_scm_shm_bridge_delete(qcom_tzmem_dev, *handle);
> +	kfree(handle);
> +}
> +
> +#endif /* CONFIG_QCOM_TZMEM_MODE_SHMBRIDGE */
>  
>  /**
>   * qcom_tzmem_pool_new() - Create a new TZ memory pool.
> -- 
> 2.39.2
>