On 03/08/2017 01:36 PM, Jeff Moyer wrote: > Hi, > > With a device dax alignment of 4KB or 2MB, I get sigbus when running the > attached fio job file for the current kernel (4.11.0-rc1+). If I > specify an alignment of 1GB, it works. > > I turned on debug output, and saw that it was failing in the huge fault > code. > > [ 4614.138357] dax dax1.0: dax_open > [ 4614.154838] dax dax1.0: dax_mmap > [ 4614.171898] dax dax1.0: dax_dev_huge_fault: fio: write (0x7f08f0a00000 - > 0x7f0ce0800000) > [ 4614.211720] dax dax1.0: __dax_dev_pud_fault: > phys_to_pgoff(0xffffffffcf600) failed > [ 4614.568911] dax dax1.0: dax_release > > Note that the printk is backwards: it's not phys_to_pgoff but rather > pgoff_to_phys that fails. > > I don't have time to dig into this right now, so I figured I'd report it > in case others do. > > This does make me wonder how much testing device dax receives. I tried > 3 different kernels (a 4.10-rc, this one, and then a version just before > the huge page support went in) and they all broke in different ways. > This test is pretty simple. Dan, can we add it to whatever regression > tests you guys run for each of 4KB, 2MB and 1GB alignments?
I'll take a look. I know I ran fio tests around the time I submitted the dev-dax ioengine and did not run into that. But you are saying it's failing even before the 1G support went in? > > Cheers, > Jeff > > [global] > ioengine=dev-dax > direct=0 > filename=/dev/dax0.0 > bs=2m > > [write] > rw=write > > [read] > stonewall > rw=read > _______________________________________________ Linux-nvdimm mailing list [email protected] https://lists.01.org/mailman/listinfo/linux-nvdimm
