CVE-2023-52910

In the Linux kernel, the following vulnerability has been resolved:

iommu/iova: Fix alloc iova overflows issue

In __allocandinsertiovarange, there is an issue that retrypfn overflows. The value of iovad->anchor.pfnhi is ~0UL, then when iovad->cachednode is iovad->anchor, curriova->pfnhi + 1 will overflow. As a result, if the retry logic is executed, lowpfn is updated to 0, and then newpfn < lowpfn returns false to make the allocation successful.

This issue occurs in the following two situations: 1. The first iova size exceeds the domain size. When initializing iova domain, iovad->cachednode is assigned as iovad->anchor. For example, the iova domain size is 10M, startpfn is 0x1F0000000, and the iova size allocated for the first time is 11M. The following is the log information, new->pfnlo is smaller than iovad->cachednode.

Example log as follows: [ 223.798112][T1705487] sh: [name:iova&]__allocandinsertiovarange startpfn:0x1f0000,retrypfn:0x0,size:0xb00,limitpfn:0x1f0a00 [ 223.799590][T1705487] sh: [name:iova&]allocandinsertiovarange success startpfn:0x1f0000,new->pfnlo:0x1efe00,new->pfnhi:0x1f08ff

2. The node with the largest iova->pfnlo value in the iova domain is deleted, iovad->cachednode will be updated to iovad->anchor, and then the alloc iova size exceeds the maximum iova size that can be allocated in the domain.

After judging that retrypfn is less than limitpfn, call retry_pfn+1 to fix the overflow issue.