From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:40020) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1auepf-0000fw-Fm for qemu-devel@nongnu.org; Mon, 25 Apr 2016 07:37:40 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1auepe-0003Vy-Iz for qemu-devel@nongnu.org; Mon, 25 Apr 2016 07:37:39 -0400 Message-ID: <571E0169.5030705@huawei.com> Date: Mon, 25 Apr 2016 19:37:13 +0800 From: Shannon Zhao MIME-Version: 1.0 References: <1461568306-14624-1-git-send-email-zhaoshenglong@huawei.com> <1461568306-14624-3-git-send-email-zhaoshenglong@huawei.com> <20160425092239.qhmyqjlrvms3auim@hawk.localdomain> In-Reply-To: <20160425092239.qhmyqjlrvms3auim@hawk.localdomain> Content-Type: text/plain; charset="ISO-8859-1" Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH v3 2/3] hw/arm/virt: Add PMU node for virt machine List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Andrew Jones Cc: qemu-arm@nongnu.org, peter.maydell@linaro.org, shannon.zhao@linaro.org, qemu-devel@nongnu.org, peter.huangpeng@huawei.com On 2016/4/25 17:22, Andrew Jones wrote: > On Mon, Apr 25, 2016 at 03:11:45PM +0800, Shannon Zhao wrote: >> > From: Shannon Zhao >> > >> > Add a virtual PMU device for virt machine while use PPI 7 for PMU >> > overflow interrupt number. >> > >> > Signed-off-by: Shannon Zhao >> > --- >> > hw/arm/virt.c | 34 ++++++++++++++++++++++++++++++++++ >> > include/hw/arm/virt.h | 4 ++++ >> > include/sysemu/kvm.h | 1 + >> > stubs/kvm.c | 5 +++++ >> > target-arm/kvm64.c | 39 +++++++++++++++++++++++++++++++++++++++ >> > 5 files changed, 83 insertions(+) >> > >> > diff --git a/hw/arm/virt.c b/hw/arm/virt.c >> > index 56d35c7..c3632d8 100644 >> > --- a/hw/arm/virt.c >> > +++ b/hw/arm/virt.c >> > @@ -428,6 +428,38 @@ static void fdt_add_gic_node(VirtBoardInfo *vbi, int type) >> > qemu_fdt_setprop_cell(vbi->fdt, "/intc", "phandle", vbi->gic_phandle); >> > } >> > >> > +static void fdt_add_pmu_nodes(const VirtBoardInfo *vbi, int gictype) >> > +{ >> > + CPUState *cpu; >> > + ARMCPU *armcpu; >> > + uint32_t irqflags = GIC_FDT_IRQ_FLAGS_LEVEL_HI; >> > + >> > + CPU_FOREACH(cpu) { >> > + armcpu = ARM_CPU(cpu); >> > + if (!armcpu->has_pmu) { >> > + return; >> > + } >> > + >> > + kvm_arm_pmu_create(cpu, PPI(VIRTUAL_PMU_IRQ)); > I think we need to return a failure code from kvm_arm_pmu_create, and > then do > > if (!kvm_arm_pmu_create(...)) { > return; > } > > Otherwise we create a /pmu node for the guest that won't work. > Ok, will update. >> > + } >> > + >> > + if (gictype == 2) { >> > + irqflags = deposit32(irqflags, GIC_FDT_IRQ_PPI_CPU_START, >> > + GIC_FDT_IRQ_PPI_CPU_WIDTH, >> > + (1 << vbi->smp_cpus) - 1); > So, if we're using gicv3, then we're level triggered and these cpu > mask bits aren't defined by the arm,gic-v3.txt bindings spec. Good. > > If we're using gicv2, then we're edge triggered and this mask is > defined. Assuming the GIC implementation requires using edge > triggered interrupts (as the comment in fdt_add_timer_nodes says), > then OK. > IIRC the comments in fdt_add_timer_nodes are not correct now since KVM makes PPI level triggered. /* Note that on A15 h/w these interrupts are level-triggered, * but for the GIC implementation provided by both QEMU and KVM * they are edge-triggered. */ Thanks, -- Shannon