FYI, we noticed a +1209 bytes kernel size regression due to commit: commit: 17e5888e4e180b45af7bafe7f3a86440d42717f3 (x86: Select HARDIRQS_SW_RESEND on x86) https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master Details as below (size data is obtained by `nm --size-sort vmlinux`): 87f2d1c6: genirq/irqdomain: Check pointer in irq_domain_alloc_irqs_hierarchy() 17e5888e: x86: Select HARDIRQS_SW_RESEND on x86 +----------------------------+----------+----------+-------+ | symbol | 87f2d1c6 | 17e5888e | delta | +----------------------------+----------+----------+-------+ | nm.b.irqs_resend | 0 | 1028 | 1028 | | bzImage | 440336 | 440496 | 160 | | nm.t.resend_irqs | 0 | 64 | 64 | | nm.T.check_irq_resend | 84 | 139 | 55 | | nm.T.irq_set_parent | 0 | 48 | 48 | | nm.d.resend_tasklet | 0 | 20 | 20 | | nm.t.irq_put_desc_unlock | 0 | 7 | 7 | | nm.R.linux_banner | 116 | 118 | 2 | | nm.T.disable_percpu_irq | 44 | 42 | -2 | | nm.T.teardown_percpu_nmi | 48 | 46 | -2 | | nm.T.can_request_irq | 73 | 71 | -2 | | nm.T.prepare_percpu_nmi | 77 | 75 | -2 | | nm.T.irq_set_vcpu_affinity | 78 | 76 | -2 | | nm.T.enable_percpu_irq | 80 | 78 | -2 | | nm.T.irq_percpu_is_enabled | 61 | 58 | -3 | +----------------------------+----------+----------+-------+ Thanks, Kbuild test robot