Wei Liu
2011-May-26 03:09 UTC
[Xen-devel] [PATCH 2/2] libxc: enabling emulated MSI injection
commit 67038b5b18096b3bdcda7fe393d1be6a91dc75a6 Author: Wei Liu <liuw@liuw.name> Date: Thu May 26 10:27:41 2011 +0800 libxc: add new operation in HVMOP to deliver emulated MSI. Signed-off-by: Wei Liu <liuw@liuw.name> diff --git a/tools/libxc/xc_misc.c b/tools/libxc/xc_misc.c index b87bbc0..38af397 100644 --- a/tools/libxc/xc_misc.c +++ b/tools/libxc/xc_misc.c @@ -417,6 +417,35 @@ int xc_hvm_set_pci_link_route( return rc; } +int xc_hvm_inj_msi( + xc_interface *xch, domid_t dom, uint64_t addr, uint32_t data) +{ + DECLARE_HYPERCALL; + DECLARE_HYPERCALL_BUFFER(struct xen_hvm_inj_msi, arg); + int rc; + + arg = xc_hypercall_buffer_alloc(xch, arg, sizeof(*arg)); + if ( arg == NULL ) + { + PERROR("Could not allocate memory for xc_hvm_inj_msi hypercall"); + return -1; + } + + hypercall.op = __HYPERVISOR_hvm_op; + hypercall.arg[0] = HVMOP_inj_msi; + hypercall.arg[1] = HYPERCALL_BUFFER_AS_ARG(arg); + + arg->domid = dom; + arg->addr = addr; + arg->data = data; + + rc = do_xen_hypercall(xch, &hypercall); + + xc_hypercall_buffer_free(xch, arg); + + return rc; +} + int xc_hvm_track_dirty_vram( xc_interface *xch, domid_t dom, uint64_t first_pfn, uint64_t nr, diff --git a/tools/libxc/xenctrl.h b/tools/libxc/xenctrl.h index 9a4355f..2442c84 100644 --- a/tools/libxc/xenctrl.h +++ b/tools/libxc/xenctrl.h @@ -1387,6 +1387,8 @@ int xc_hvm_set_isa_irq_level( int xc_hvm_set_pci_link_route( xc_interface *xch, domid_t dom, uint8_t link, uint8_t isa_irq); +int xc_hvm_inj_msi( + xc_interface *xch, domid_t dom, uint64_t addr, uint32_t data); /* * Track dirty bit changes in the VRAM area -- Best regards Wei Liu Twitter: @iliuw Site: http://liuw.name _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Konrad Rzeszutek Wilk
2011-May-26 12:34 UTC
Re: [Xen-devel] [PATCH 2/2] libxc: enabling emulated MSI injection
On Thu, May 26, 2011 at 11:09:43AM +0800, Wei Liu wrote:> commit 67038b5b18096b3bdcda7fe393d1be6a91dc75a6 > Author: Wei Liu <liuw@liuw.name> > Date: Thu May 26 10:27:41 2011 +0800 > > libxc: add new operation in HVMOP to deliver emulated MSI.What is the purpose of this call?> > Signed-off-by: Wei Liu <liuw@liuw.name> > > diff --git a/tools/libxc/xc_misc.c b/tools/libxc/xc_misc.c > index b87bbc0..38af397 100644 > --- a/tools/libxc/xc_misc.c > +++ b/tools/libxc/xc_misc.c > @@ -417,6 +417,35 @@ int xc_hvm_set_pci_link_route( > return rc; > } > > +int xc_hvm_inj_msi( > + xc_interface *xch, domid_t dom, uint64_t addr, uint32_t data) > +{ > + DECLARE_HYPERCALL; > + DECLARE_HYPERCALL_BUFFER(struct xen_hvm_inj_msi, arg); > + int rc; > + > + arg = xc_hypercall_buffer_alloc(xch, arg, sizeof(*arg)); > + if ( arg == NULL ) > + { > + PERROR("Could not allocate memory for xc_hvm_inj_msi hypercall"); > + return -1; > + } > + > + hypercall.op = __HYPERVISOR_hvm_op; > + hypercall.arg[0] = HVMOP_inj_msi; > + hypercall.arg[1] = HYPERCALL_BUFFER_AS_ARG(arg); > + > + arg->domid = dom; > + arg->addr = addr; > + arg->data = data;Shouldn''t those be done before you make the HYPERCALL_BUFFER_AS_ARG in case they get bounced?> + > + rc = do_xen_hypercall(xch, &hypercall); > + > + xc_hypercall_buffer_free(xch, arg); > + > + return rc; > +} > + > int xc_hvm_track_dirty_vram( > xc_interface *xch, domid_t dom, > uint64_t first_pfn, uint64_t nr, > diff --git a/tools/libxc/xenctrl.h b/tools/libxc/xenctrl.h > index 9a4355f..2442c84 100644 > --- a/tools/libxc/xenctrl.h > +++ b/tools/libxc/xenctrl.h > @@ -1387,6 +1387,8 @@ int xc_hvm_set_isa_irq_level( > int xc_hvm_set_pci_link_route( > xc_interface *xch, domid_t dom, uint8_t link, uint8_t isa_irq); > > +int xc_hvm_inj_msi( > + xc_interface *xch, domid_t dom, uint64_t addr, uint32_t data); > > /* > * Track dirty bit changes in the VRAM area > > -- > Best regards > Wei Liu > Twitter: @iliuw > Site: http://liuw.name > > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xensource.com > http://lists.xensource.com/xen-devel_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Wei Liu
2011-May-26 12:56 UTC
Re: [Xen-devel] [PATCH 2/2] libxc: enabling emulated MSI injection
On Thu, May 26, 2011 at 8:34 PM, Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> wrote:> On Thu, May 26, 2011 at 11:09:43AM +0800, Wei Liu wrote: >> commit 67038b5b18096b3bdcda7fe393d1be6a91dc75a6 >> Author: Wei Liu <liuw@liuw.name> >> Date: Thu May 26 10:27:41 2011 +0800 >> >> libxc: add new operation in HVMOP to deliver emulated MSI. > > What is the purpose of this call? >> >> Signed-off-by: Wei Liu <liuw@liuw.name> >> >> diff --git a/tools/libxc/xc_misc.c b/tools/libxc/xc_misc.c >> index b87bbc0..38af397 100644 >> --- a/tools/libxc/xc_misc.c >> +++ b/tools/libxc/xc_misc.c >> @@ -417,6 +417,35 @@ int xc_hvm_set_pci_link_route( >> return rc; >> } >> >> +int xc_hvm_inj_msi( >> + xc_interface *xch, domid_t dom, uint64_t addr, uint32_t data) >> +{ >> + DECLARE_HYPERCALL; >> + DECLARE_HYPERCALL_BUFFER(struct xen_hvm_inj_msi, arg); >> + int rc; >> + >> + arg = xc_hypercall_buffer_alloc(xch, arg, sizeof(*arg)); >> + if ( arg == NULL ) >> + { >> + PERROR("Could not allocate memory for xc_hvm_inj_msi hypercall"); >> + return -1; >> + } >> + >> + hypercall.op = __HYPERVISOR_hvm_op; >> + hypercall.arg[0] = HVMOP_inj_msi; >> + hypercall.arg[1] = HYPERCALL_BUFFER_AS_ARG(arg); >> + >> + arg->domid = dom; >> + arg->addr = addr; >> + arg->data = data; > > Shouldn''t those be done before you make the HYPERCALL_BUFFER_AS_ARG > in case they get bounced? >Hmm... I didn''t think that far. I just copy-and-edit from a previous hypercall. I think that should be OK then.>> + >> + rc = do_xen_hypercall(xch, &hypercall); >> + >> + xc_hypercall_buffer_free(xch, arg); >> + >> + return rc; >> +} >> + >> int xc_hvm_track_dirty_vram( >> xc_interface *xch, domid_t dom, >> uint64_t first_pfn, uint64_t nr, >> diff --git a/tools/libxc/xenctrl.h b/tools/libxc/xenctrl.h >> index 9a4355f..2442c84 100644 >> --- a/tools/libxc/xenctrl.h >> +++ b/tools/libxc/xenctrl.h >> @@ -1387,6 +1387,8 @@ int xc_hvm_set_isa_irq_level( >> int xc_hvm_set_pci_link_route( >> xc_interface *xch, domid_t dom, uint8_t link, uint8_t isa_irq); >> >> +int xc_hvm_inj_msi( >> + xc_interface *xch, domid_t dom, uint64_t addr, uint32_t data); >> >> /* >> * Track dirty bit changes in the VRAM area >> >> -- >> Best regards >> Wei Liu >> Twitter: @iliuw >> Site: http://liuw.name >> >> _______________________________________________ >> Xen-devel mailing list >> Xen-devel@lists.xensource.com >> http://lists.xensource.com/xen-devel >-- Best regards Wei Liu Twitter: @iliuw Site: http://liuw.name _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Wei Liu
2011-May-26 12:59 UTC
Re: [Xen-devel] [PATCH 2/2] libxc: enabling emulated MSI injection
On Thu, May 26, 2011 at 8:34 PM, Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> wrote:> On Thu, May 26, 2011 at 11:09:43AM +0800, Wei Liu wrote: >> commit 67038b5b18096b3bdcda7fe393d1be6a91dc75a6 >> Author: Wei Liu <liuw@liuw.name> >> Date: Thu May 26 10:27:41 2011 +0800 >> >> libxc: add new operation in HVMOP to deliver emulated MSI. > > What is the purpose of this call?Sorry I missed this one. Currently QEMU cannot inject MSI for emulated devices, which limits virtio to use irq to deliver interrupts. This call is used to inject MSI from QEMU to HVM, suggested by Stefano.>> >> Signed-off-by: Wei Liu <liuw@liuw.name> >> >> diff --git a/tools/libxc/xc_misc.c b/tools/libxc/xc_misc.c >> index b87bbc0..38af397 100644 >> --- a/tools/libxc/xc_misc.c >> +++ b/tools/libxc/xc_misc.c >> @@ -417,6 +417,35 @@ int xc_hvm_set_pci_link_route( >> return rc; >> } >> >> +int xc_hvm_inj_msi( >> + xc_interface *xch, domid_t dom, uint64_t addr, uint32_t data) >> +{ >> + DECLARE_HYPERCALL; >> + DECLARE_HYPERCALL_BUFFER(struct xen_hvm_inj_msi, arg); >> + int rc; >> + >> + arg = xc_hypercall_buffer_alloc(xch, arg, sizeof(*arg)); >> + if ( arg == NULL ) >> + { >> + PERROR("Could not allocate memory for xc_hvm_inj_msi hypercall"); >> + return -1; >> + } >> + >> + hypercall.op = __HYPERVISOR_hvm_op; >> + hypercall.arg[0] = HVMOP_inj_msi; >> + hypercall.arg[1] = HYPERCALL_BUFFER_AS_ARG(arg); >> + >> + arg->domid = dom; >> + arg->addr = addr; >> + arg->data = data; > > Shouldn''t those be done before you make the HYPERCALL_BUFFER_AS_ARG > in case they get bounced? > >> + >> + rc = do_xen_hypercall(xch, &hypercall); >> + >> + xc_hypercall_buffer_free(xch, arg); >> + >> + return rc; >> +} >> + >> int xc_hvm_track_dirty_vram( >> xc_interface *xch, domid_t dom, >> uint64_t first_pfn, uint64_t nr, >> diff --git a/tools/libxc/xenctrl.h b/tools/libxc/xenctrl.h >> index 9a4355f..2442c84 100644 >> --- a/tools/libxc/xenctrl.h >> +++ b/tools/libxc/xenctrl.h >> @@ -1387,6 +1387,8 @@ int xc_hvm_set_isa_irq_level( >> int xc_hvm_set_pci_link_route( >> xc_interface *xch, domid_t dom, uint8_t link, uint8_t isa_irq); >> >> +int xc_hvm_inj_msi( >> + xc_interface *xch, domid_t dom, uint64_t addr, uint32_t data); >> >> /* >> * Track dirty bit changes in the VRAM area >> >> -- >> Best regards >> Wei Liu >> Twitter: @iliuw >> Site: http://liuw.name >> >> _______________________________________________ >> Xen-devel mailing list >> Xen-devel@lists.xensource.com >> http://lists.xensource.com/xen-devel >-- Best regards Wei Liu Twitter: @iliuw Site: http://liuw.name _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Konrad Rzeszutek Wilk
2011-May-26 13:05 UTC
Re: [Xen-devel] [PATCH 2/2] libxc: enabling emulated MSI injection
> >> + hypercall.op = __HYPERVISOR_hvm_op; > >> + hypercall.arg[0] = HVMOP_inj_msi; > >> + hypercall.arg[1] = HYPERCALL_BUFFER_AS_ARG(arg); > >> + > >> + arg->domid = dom; > >> + arg->addr = addr; > >> + arg->data = data; > > > > Shouldn''t those be done before you make the HYPERCALL_BUFFER_AS_ARG > > in case they get bounced? > > > > Hmm... I didn''t think that far. > > I just copy-and-edit from a previous hypercall. I think that should be OK then.On a second look at the code, I believe you are right. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel