> -----Original Message----- > From: Simon Horman <[email protected]> > Sent: Monday, September 2, 2024 1:49 AM > To: Zaki, Ahmed <[email protected]> > Cc: [email protected]; [email protected]; Linga, Pavan > Kumar <[email protected]>; Kitszel, Przemyslaw > <[email protected]> > Subject: Re: [PATCH iwl-net] idpf: fix VF dynamic interrupt ctl register > initialization > > On Wed, Aug 28, 2024 at 04:38:25PM -0600, Ahmed Zaki wrote: > > The VF's dynamic interrupt ctl "dyn_ctl_intrvl_s" is not initialized > > in idpf_vf_intr_reg_init(). This resulted in the following UBSAN error > > whenever a VF is created: > > > > [ 564.345655] UBSAN: shift-out-of-bounds in > drivers/net/ethernet/intel/idpf/idpf_txrx.c:3654:10 > > [ 564.345663] shift exponent 4294967295 is too large for 32-bit type 'int' > > [ 564.345671] CPU: 33 UID: 0 PID: 2458 Comm: NetworkManager Not tainted > 6.11.0-rc4+ #1 > > [ 564.345678] Hardware name: Intel Corporation > M50CYP2SBSTD/M50CYP2SBSTD, BIOS SE5C6200.86B.0027.P10.2201070222 > 01/07/2022 > > [ 564.345683] Call Trace: > > [ 564.345688] <TASK> > > [ 564.345693] dump_stack_lvl+0x91/0xb0 > > [ 564.345708] __ubsan_handle_shift_out_of_bounds+0x16b/0x320 > > [ 564.345730] idpf_vport_intr_update_itr_ena_irq.cold+0x13/0x39 [idpf] > > [ 564.345755] ? __pfx_idpf_vport_intr_update_itr_ena_irq+0x10/0x10 [idpf] > > [ 564.345771] ? static_obj+0x95/0xd0 > > [ 564.345782] ? lockdep_init_map_type+0x1a5/0x800 > > [ 564.345794] idpf_vport_intr_ena+0x5ef/0x9f0 [idpf] > > [ 564.345814] idpf_vport_open+0x2cc/0x1240 [idpf] > > [ 564.345837] idpf_open+0x6d/0xc0 [idpf] > > [ 564.345850] __dev_open+0x241/0x420 > > > > Fixes: d4d558718266 ("idpf: initialize interrupts and enable vport") > > Reviewed-by: Przemek Kitszel <[email protected]> > > Signed-off-by: Ahmed Zaki <[email protected]> > > Thanks, > > I see that this now matches the implementation of idpf_intr_reg_init(). > > Reviewed-by: Simon Horman <[email protected]>
Tested-by: Krishneil Singh <[email protected]>
