On 12/02/2024 10:54, Nilay Shroff wrote:
Shall we ensure here that we don't cross max limit of atomic write supported by device? It seems that if the boundary size is not advertized by the device (in fact, I have one NVMe drive which has boundary size zero i.e. nabo/nabspf/ nawupf are all zero but awupf is non-zero) then we (unconditionally) allow merging.
BTW, if you don't mind, can you share awupf value and device model? I have been on the search for NVMe devices which support atomic writes (with non-zero PF reported value). All I have is a M.2 card which has a 4KB PF atomic write value.
But if this is private info, then ok. Thanks, John
