From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753275AbdCPV70 (ORCPT ); Thu, 16 Mar 2017 17:59:26 -0400 Received: from mx1.redhat.com ([209.132.183.28]:47774 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752122AbdCPV7Y (ORCPT ); Thu, 16 Mar 2017 17:59:24 -0400 DMARC-Filter: OpenDMARC Filter v1.3.2 mx1.redhat.com 0867464A67 Authentication-Results: ext-mx02.extmail.prod.ext.phx2.redhat.com; dmarc=none (p=none dis=none) header.from=redhat.com Authentication-Results: ext-mx02.extmail.prod.ext.phx2.redhat.com; spf=pass smtp.mailfrom=eric.auger@redhat.com DKIM-Filter: OpenDKIM Filter v2.11.0 mx1.redhat.com 0867464A67 Subject: Re: [RFC PATCH 27/33] irqchip/gic-v3-its: Add VPE interrupt masking To: Marc Zyngier , linux-kernel@vger.kernel.org, linux-arm-kernel@lists.infradead.org, kvmarm@lists.cs.columbia.edu References: <1484648454-21216-1-git-send-email-marc.zyngier@arm.com> <1484648454-21216-28-git-send-email-marc.zyngier@arm.com> Cc: Thomas Gleixner , Jason Cooper , Christoffer Dall From: Auger Eric Message-ID: <971d3d21-d811-7259-3d0c-7cac532743cf@redhat.com> Date: Thu, 16 Mar 2017 22:58:03 +0100 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Thunderbird/45.4.0 MIME-Version: 1.0 In-Reply-To: <1484648454-21216-28-git-send-email-marc.zyngier@arm.com> Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.26]); Thu, 16 Mar 2017 21:58:08 +0000 (UTC) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi Marc, On 17/01/2017 11:20, Marc Zyngier wrote: > When masking/unmasking a doorbell interrupt, it is necessary > to issue an invalidation to the corresponding redistributor. > We use the DirectLPI feature by writting directly to the corresponding s/writting/writing is DirectLPI always supported? I can see the GICR_TYPER advertises (or not) the modality. > redistributor. > > Signed-off-by: Marc Zyngier > --- > drivers/irqchip/irq-gic-v3-its.c | 23 +++++++++++++++++++++-- > 1 file changed, 21 insertions(+), 2 deletions(-) > > diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c > index 06e959f..008fb71 100644 > --- a/drivers/irqchip/irq-gic-v3-its.c > +++ b/drivers/irqchip/irq-gic-v3-its.c > @@ -950,14 +950,19 @@ static inline u32 its_get_event_id(struct irq_data *d) > return d->hwirq - its_dev->event_map.lpi_base; > } > > +static inline bool its_is_vpe_lpi(struct irq_data *d) > +{ > + return !irq_desc_get_msi_desc(irq_data_to_desc(d)); > +} > + > static void lpi_update_config(struct irq_data *d, u8 clr, u8 set) > { > - struct its_device *its_dev = irq_data_get_irq_chip_data(d); > irq_hw_number_t hwirq; > struct page *prop_page; > u8 *cfg; > > if (irqd_is_forwarded_to_vcpu(d)) { > + struct its_device *its_dev = irq_data_get_irq_chip_data(d); nit: leave blank line (sorry I need to find something ;-) > u32 event = its_get_event_id(d); > prop_page = its_dev->event_map.vlpi_map->vpes[0]->its_vm->vprop_page; > hwirq = its_dev->event_map.vlpi_map->vlpis[event].vintid; > @@ -979,7 +984,18 @@ static void lpi_update_config(struct irq_data *d, u8 clr, u8 set) > gic_flush_dcache_to_poc(cfg, sizeof(*cfg)); > else > dsb(ishst); > - its_send_inv(its_dev, its_get_event_id(d)); > + > + if (!its_is_vpe_lpi(d)) { > + struct its_device *its_dev = irq_data_get_irq_chip_data(d); > + > + its_send_inv(its_dev, its_get_event_id(d)); > + } else { > + struct its_vpe *vpe = irq_data_get_irq_chip_data(d); > + void __iomem *rdbase; > + > + rdbase = per_cpu_ptr(gic_rdists->rdist, vpe->col_idx)->rd_base; > + writeq_relaxed(d->hwirq, rdbase + GICR_INVLPIR); > + } > } > > static void its_vlpi_set_doorbell(struct irq_data *d, bool enable) > @@ -2196,6 +2212,9 @@ static int its_vpe_set_vcpu_affinity(struct irq_data *d, void *vcpu_info) > > static struct irq_chip its_vpe_irq_chip = { > .name = "GICv4-vpe", > + .irq_mask = its_mask_irq, > + .irq_unmask = its_unmask_irq, > + .irq_eoi = irq_chip_eoi_parent, not documented in the commit msg? Eric > .irq_set_affinity = its_vpe_set_affinity, > .irq_set_vcpu_affinity = its_vpe_set_vcpu_affinity, > }; >