Re: [PATCH v4 5/8] scsi: ufs: UFS crypto API

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Wed, Aug 21, 2019 at 12:57:11AM -0700, Satya Tangirala wrote:
> +static int ufshcd_crypto_cap_find(void *hba_p,
> +			   enum blk_crypto_mode_num crypto_mode,
> +			   unsigned int data_unit_size)
> +{
> +	struct ufs_hba *hba = hba_p;
> +	enum ufs_crypto_alg ufs_alg;
> +	u8 data_unit_mask;
> +	int cap_idx;
> +	enum ufs_crypto_key_size ufs_key_size;
> +	union ufs_crypto_cap_entry *ccap_array = hba->crypto_cap_array;
> +
> +	if (!ufshcd_hba_is_crypto_supported(hba))
> +		return -EINVAL;
> +
> +	switch (crypto_mode) {
> +	case BLK_ENCRYPTION_MODE_AES_256_XTS:
> +		ufs_alg = UFS_CRYPTO_ALG_AES_XTS;
> +		ufs_key_size = UFS_CRYPTO_KEY_SIZE_256;
> +		break;
> +	/*
> +	 * case BLK_CRYPTO_ALG_BITLOCKER_AES_CBC:
> +	 *	ufs_alg = UFS_CRYPTO_ALG_BITLOCKER_AES_CBC;
> +	 *	break;
> +	 * case BLK_CRYPTO_ALG_AES_ECB:
> +	 *	ufs_alg = UFS_CRYPTO_ALG_AES_ECB;
> +	 *	break;
> +	 * case BLK_CRYPTO_ALG_ESSIV_AES_CBC:
> +	 *	ufs_alg = UFS_CRYPTO_ALG_ESSIV_AES_CBC;
> +	 *	break;
> +	 */

Perhaps just delete this comment... the constants are already outdated.

> +	hba->crypto_cfgs =
> +		devm_kcalloc(hba->dev,
> +			     hba->crypto_capabilities.config_count + 1,
> +			     sizeof(hba->crypto_cfgs[0]),
> +			     GFP_KERNEL);

Can use NUM_KEYSLOTS(hba) here, to avoid hardcoding the awkward '+ 1' again.

> +void ufshcd_crypto_setup_rq_keyslot_manager(struct ufs_hba *hba,
> +					    struct request_queue *q)
> +{
> +	if (!ufshcd_hba_is_crypto_supported(hba))
> +		return;
> +
> +	if (q) {
> +		mutex_lock(&hba->ksm_lock);
> +		if (!hba->ksm) {
> +			hba->ksm = keyslot_manager_create(
> +				hba->crypto_capabilities.config_count + 1,
> +				&ufshcd_ksm_ops, hba);
> +			hba->ksm_num_refs = 0;

Same here.

> +		}
> +		hba->ksm_num_refs++;
> +		mutex_unlock(&hba->ksm_lock);
> +		q->ksm = hba->ksm;
> +	}
> +	/*
> +	 * If we fail we make it look like
> +	 * crypto is not supported, which will avoid issues
> +	 * with reset
> +	 */
> +	if (!q || !q->ksm) {
> +		ufshcd_crypto_disable(hba);
> +		hba->crypto_capabilities.reg_val = 0;
> +		devm_kfree(hba->dev, hba->crypto_cap_array);
> +		devm_kfree(hba->dev, hba->crypto_cfgs);
> +	}
> +}
> +
> +void ufshcd_crypto_destroy_rq_keyslot_manager(struct ufs_hba *hba,
> +					      struct request_queue *q)
> +{
> +	if (q && q->ksm) {
> +		q->ksm = NULL;
> +		mutex_lock(&hba->ksm_lock);
> +		hba->ksm_num_refs--;
> +		if (hba->ksm_num_refs == 0) {
> +			keyslot_manager_destroy(hba->ksm);
> +			hba->ksm = NULL;
> +		}
> +		mutex_unlock(&hba->ksm_lock);
> +	}
> +}

Why is the keyslot_manager reference counted?  Doesn't it live as long as the
individual devices do?  So, can't we just create the keyslot manager when the
ufs_hba is created, and destroy it when the ufs_hba is destroyed?  Then for each
device we'd just set 'q->ksm = hba->ksm;', with no refcounting needed.

- Eric



[Index of Archives]     [Linux RAID]     [Linux SCSI]     [Linux ATA RAID]     [IDE]     [Linux Wireless]     [Linux Kernel]     [ATH6KL]     [Linux Bluetooth]     [Linux Netdev]     [Kernel Newbies]     [Security]     [Git]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Device Mapper]

  Powered by Linux