On 2019/8/15 下午10:13, Greg Kroah-Hartman wrote:
On Wed, Aug 14, 2019 at 05:34:25PM +0800, Zhangfei Gao wrote:+int uacce_register(struct uacce *uacce) +{ + int ret; + + if (!uacce->pdev) { + pr_debug("uacce parent device not set\n"); + return -ENODEV; + } + + if (uacce->flags & UACCE_DEV_NOIOMMU) { + add_taint(TAINT_CRAP, LOCKDEP_STILL_OK); + dev_warn(uacce->pdev, + "Register to noiommu mode, which export kernel data to user space and may vulnerable to attack"); + }THat is odd, why even offer this feature then if it is a major issue?
UACCE_DEV_NOIOMMU maybe confusing here. In this mode, app use ioctl to get dma_handle from dma_alloc_coherent. It does not matter iommu is enabled or not. In case iommu is disabled, it maybe dangerous to kernel, so we added warning here, is it required? We support two modes. 1. support sva, (va = iova) a. If pasid is supported, multi-process can be supported. b. If pasid is not supported, multi-process can NOT be supported. We borrowed va from user space to achieve a single virtual address system. 2. Can not support sva, iova != va, Here user app get dma_handle from dma_alloc_coherent via ioctl. We need this mode for two reasons: 1. This mode can support multi-process, to support openssl etc. 2. Some accelerators (crypto like zip, etc) can work without sva, just prepare data and kick start. Thanks
