Return-path: Received: from esa5.microchip.iphmx.com ([216.71.150.166]:24958 "EHLO esa5.microchip.iphmx.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726995AbeHYDuI (ORCPT ); Fri, 24 Aug 2018 23:50:08 -0400 Date: Fri, 24 Aug 2018 17:13:14 -0700 From: Adham Abozaeid To: Claudiu Beznea CC: Ajay Singh , , , , , , Subject: Re: [PATCH 09/24] staging: wilc1000: move static variable clients_count to 'wilc' structure Message-ID: <20180824171314.466fd48d@edcwlt13-vm> (sfid-20180825_021332_197394_C435204A) In-Reply-To: References: <1534229416-13254-1-git-send-email-ajay.kathat@microchip.com> <1534229416-13254-10-git-send-email-ajay.kathat@microchip.com> MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Sender: linux-wireless-owner@vger.kernel.org List-ID: On Thu, 23 Aug 2018 11:09:56 +0300 Claudiu Beznea wrote: > > > On 14.08.2018 09:50, Ajay Singh wrote: > > Avoid use of static variable 'clients_count' and move it part of 'wilc' > > structure. > > > > Signed-off-by: Ajay Singh > > --- > > drivers/staging/wilc1000/host_interface.c | 9 ++++----- > > drivers/staging/wilc1000/wilc_wfi_netdevice.h | 1 + > > 2 files changed, 5 insertions(+), 5 deletions(-) > > > > diff --git a/drivers/staging/wilc1000/host_interface.c b/drivers/staging/wilc1000/host_interface.c > > index 6225e67..d930f06 100644 > > --- a/drivers/staging/wilc1000/host_interface.c > > +++ b/drivers/staging/wilc1000/host_interface.c > > @@ -199,7 +199,6 @@ static u8 rcv_assoc_resp[MAX_ASSOC_RESP_FRAME_SIZE]; > > > > static u8 set_ip[2][4]; > > static u8 get_ip[2][4]; > > -static u32 clients_count; > > > > static int host_int_get_ipaddress(struct wilc_vif *vif, u8 *ip_addr, u8 idx); > > > > @@ -3456,7 +3455,7 @@ int wilc_init(struct net_device *dev, struct host_if_drv **hif_drv_handler) > > > > vif->obtaining_ip = false; > > > > - if (clients_count == 0) { > > + if (wilc->clients_count == 0) { > > The check of this should be mutex protected or a hif init/deinit should be > done in probe/remove code not not in open (I'm for the second approach). > Imagine this scenario: > > vif 1: is down > vif 2: is down. "ifconfig vif2 up" is executed and the execution reach this > point and then suspends after "if (wilc->clients_count == 0)" check > vif 1: "ifconfig vif1 up" and execution reach the end of this function and > then the execution for vif 2 resumes. This will init twice the hif > mutex, completion etc. > > > init_completion(&hif_driver_comp); > > mutex_init(&hif_deinit_lock); > > Instead of using mutex protection, it would be better to move this block to wilc_netdev_init() where wilc structure is initialized > > @@ -3490,7 +3489,7 @@ int wilc_init(struct net_device *dev, struct host_if_drv **hif_drv_handler) > > > > mutex_unlock(&hif_drv->cfg_values_lock); > > > > - clients_count++; > > + wilc->clients_count++; > > > > return 0; > > } > > @@ -3526,7 +3525,7 @@ int wilc_deinit(struct wilc_vif *vif) > > > > hif_drv->hif_state = HOST_IF_IDLE; > > > > - if (clients_count == 1) { > > + if (vif->wilc->clients_count == 1) { > > struct host_if_msg *msg; > > > > msg = wilc_alloc_work(vif, handle_hif_exit_work, true); > > @@ -3544,7 +3543,7 @@ int wilc_deinit(struct wilc_vif *vif) > > > > kfree(hif_drv); > > > > - clients_count--; > > + vif->wilc->clients_count--; > > terminated_handle = NULL; > > mutex_unlock(&hif_deinit_lock); > > return result; > > diff --git a/drivers/staging/wilc1000/wilc_wfi_netdevice.h b/drivers/staging/wilc1000/wilc_wfi_netdevice.h > > index 8e56a28..8cccbbc 100644 > > --- a/drivers/staging/wilc1000/wilc_wfi_netdevice.h > > +++ b/drivers/staging/wilc1000/wilc_wfi_netdevice.h > > @@ -171,6 +171,7 @@ struct wilc { > > > > struct rf_info dummy_statistics; > > bool enable_ps; > > + int clients_count; > > }; > > > > struct wilc_wfi_mon_priv { > >