On Tue, Jan 12, 2021 at 10:10:45PM +0800, kernel test robot wrote: > FYI, we noticed a -1.2% regression of will-it-scale.per_process_ops due to > commit: > > commit: dc83615370e7ebcb181a21a8ad13a77c278ab81c ("PCI: Fix PREL32 > relocations for LTO") > https://git.kernel.org/cgit/linux/kernel/git/kees/linux.git for-next/kspp
O_o Well that's surprising. This only creates stubs. > will-it-scale.workload > > > > 2.14e+07 +----------------------------------------------------------------+ > > |.+. .+ : + +.+. +.+. .+ +.+. .+ : | > > 2.13e+07 |-+ + : +.+ + + + + : | > > | : : + .+.+..+. .+ +. | > > 2.12e+07 |-+ : : + + +.+.| > > | +.+ | > > 2.11e+07 |-+ | > > | | > > 2.1e+07 |-+ O O O O O O O | > > | O O O O O O O O O | > > 2.09e+07 |-+ O | > > | O O | > > 2.08e+07 |-O O O O O | > > | | > > 2.07e+07 +----------------------------------------------------------------+ > > > > > > [*] bisect-good sample > [O] bisect-bad sample But it's pretty clear _something_ has happened. We'll investigate... -- Kees Cook