Re: [PATCH 05/11] KVM: Add fields for MSI device assignment
On Tue, Nov 25, 2008 at 04:49:49PM +0200, Avi Kivity wrote: > Avi Kivity wrote: >>> (In fact, PCI spec defined message data length is u64, but as you >>> see, now msi_msg for whole Linux only have u32...) >>> >> >> Well in that case please define data as __u64, so we don't have >> surprises later on. >> > > btw, if you agree with this, don't resend; I'll edit the patch. Sorry, I found I made a mistake. Intel SDM 3A misleaded me... The PCI spec defined message data as u16(!!), but Intel SDM 3A got a figure(Figure 8-25) that reserved all bits up to bit 63, so I thought message data got u64 long...(At last moment, I remembered the figure on PCI spec and felt something strange...) Now I think the latest patchset(v6) is OK to apply. -- regards Yang, Sheng > > -- > error compiling committee.c: too many arguments to function > > -- > To unsubscribe from this list: send the line "unsubscribe kvm" in > the body of a message to [EMAIL PROTECTED] > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 05/11] KVM: Add fields for MSI device assignment
Avi Kivity wrote: (In fact, PCI spec defined message data length is u64, but as you see, now msi_msg for whole Linux only have u32...) Well in that case please define data as __u64, so we don't have surprises later on. btw, if you agree with this, don't resend; I'll edit the patch. -- error compiling committee.c: too many arguments to function -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 05/11] KVM: Add fields for MSI device assignment
Sheng Yang wrote: union { + struct { + __u32 addr_lo; + __u32 addr_hi; __u64 addr; ? Here I followed the spec that distinguish the "Message Address" and "Message Upper address". And the native Linux structure: struct msi_msg { u32 address_lo; /* low 32 bits of msi message address */ u32 address_hi; /* high 32 bits of msi message address */ u32 data; /* 16 bits of msi message data */ }; For now, we needn't care about address_hi. I can only see address_hi used in hypertransport part... So I think keep it independence here is OK. Fair enough, documentation wins. (In fact, PCI spec defined message data length is u64, but as you see, now msi_msg for whole Linux only have u32...) Well in that case please define data as __u64, so we don't have surprises later on. -- error compiling committee.c: too many arguments to function -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 05/11] KVM: Add fields for MSI device assignment
On Sunday 23 November 2008 18:10:47 Avi Kivity wrote: > Sheng Yang wrote: > > Prepared for kvm_arch_assigned_device_msi_dispatch(). > > > > @@ -507,10 +507,17 @@ struct kvm_assigned_irq { > > __u32 guest_irq; > > __u32 flags; > > union { > > + struct { > > + __u32 addr_lo; > > + __u32 addr_hi; > > __u64 addr; > > ? Here I followed the spec that distinguish the "Message Address" and "Message Upper address". And the native Linux structure: struct msi_msg { u32 address_lo; /* low 32 bits of msi message address */ u32 address_hi; /* high 32 bits of msi message address */ u32 data; /* 16 bits of msi message data */ }; For now, we needn't care about address_hi. I can only see address_hi used in hypertransport part... So I think keep it independence here is OK. (In fact, PCI spec defined message data length is u64, but as you see, now msi_msg for whole Linux only have u32...) -- regards Yang, Sheng > > > + __u32 data; > > > > > > > > @@ -307,8 +308,11 @@ struct kvm_assigned_dev_kernel { > > int host_devfn; > > int host_irq; > > int guest_irq; > > + struct msi_msg guest_msi; > > #define KVM_ASSIGNED_DEV_GUEST_INTX(1 << 0) > > +#define KVM_ASSIGNED_DEV_GUEST_MSI (1 << 1) > > #define KVM_ASSIGNED_DEV_HOST_INTX (1 << 8) > > +#define KVM_ASSIGNED_DEV_HOST_MSI (1 << 9) > > Okay, I see the reason for non sequential assignment. -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 05/11] KVM: Add fields for MSI device assignment
Sheng Yang wrote: Prepared for kvm_arch_assigned_device_msi_dispatch(). @@ -507,10 +507,17 @@ struct kvm_assigned_irq { __u32 guest_irq; __u32 flags; union { + struct { + __u32 addr_lo; + __u32 addr_hi; __u64 addr; ? + __u32 data; @@ -307,8 +308,11 @@ struct kvm_assigned_dev_kernel { int host_devfn; int host_irq; int guest_irq; + struct msi_msg guest_msi; #define KVM_ASSIGNED_DEV_GUEST_INTX(1 << 0) +#define KVM_ASSIGNED_DEV_GUEST_MSI (1 << 1) #define KVM_ASSIGNED_DEV_HOST_INTX (1 << 8) +#define KVM_ASSIGNED_DEV_HOST_MSI (1 << 9) Okay, I see the reason for non sequential assignment. -- error compiling committee.c: too many arguments to function -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html