From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:42802) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1cqGnf-00052d-JE for qemu-devel@nongnu.org; Tue, 21 Mar 2017 06:14:01 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1cqGnc-0005O8-9K for qemu-devel@nongnu.org; Tue, 21 Mar 2017 06:13:59 -0400 Received: from smtp.eu.citrix.com ([185.25.65.24]:19219) by eggs.gnu.org with esmtps (TLS1.0:RSA_ARCFOUR_SHA1:16) (Exim 4.71) (envelope-from ) id 1cqGnb-0005Nq-U4 for qemu-devel@nongnu.org; Tue, 21 Mar 2017 06:13:56 -0400 From: Paul Durrant Date: Tue, 21 Mar 2017 10:13:53 +0000 Message-ID: <072db16bacfd432b8eb7df09cb1eab7d@AMSPEX02CL03.citrite.net> References: <1490028748-3780-1-git-send-email-paul.durrant@citrix.com> In-Reply-To: Content-Language: en-US Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Subject: Re: [Qemu-devel] [PATCH v2] xen: use libxendevice model to restrict operations List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: 'Stefano Stabellini' Cc: "qemu-devel@nongnu.org" , Anthony Perard , Paolo Bonzini , Ian Jackson Fixing list address... > -----Original Message----- > From: Stefano Stabellini [mailto:sstabellini@kernel.org] > Sent: 20 March 2017 22:05 > To: Paul Durrant > Cc: qemu-devel@lists.xenproject.org; Stefano Stabellini > ; Anthony Perard ; > Paolo Bonzini > Subject: Re: [PATCH v2] xen: use libxendevice model to restrict operation= s >=20 > On Mon, 20 Mar 2017, Paul Durrant wrote: > > This patch adds a command-line option (-xen-domid-restrict) which will > > use the new libxendevicemodel API to restrict devicemodel operations to > > the specified domid. > > > > This patch also adds a tracepoint to allow successful enabling of the > > restriction to be monitored. > > > > Signed-off-by: Paul Durrant > > --- > > Cc: Stefano Stabellini > > Cc: Anthony Perard > > Cc: Paolo Bonzini > > > > NOTE: This is already re-based on Juergen Gross's patch "xen: use 5 dig= it > > xen versions" and so should not be applied until after that patch > > has been applied. > > > > v2: > > - Log errno in tracepoint > > --- > > hw/xen/trace-events | 1 + > > include/hw/xen/xen.h | 1 + > > include/hw/xen/xen_common.h | 20 ++++++++++++++++++++ > > qemu-options.hx | 6 ++++++ > > vl.c | 8 ++++++++ > > xen-hvm.c | 8 ++++++++ > > 6 files changed, 44 insertions(+) > > > > diff --git a/hw/xen/trace-events b/hw/xen/trace-events > > index c4fb6f1..5615dce 100644 > > --- a/hw/xen/trace-events > > +++ b/hw/xen/trace-events > > @@ -11,3 +11,4 @@ xen_map_portio_range(uint32_t id, uint64_t > start_addr, uint64_t end_addr) "id: % > > xen_unmap_portio_range(uint32_t id, uint64_t start_addr, uint64_t > end_addr) "id: %u start: %#"PRIx64" end: %#"PRIx64 > > xen_map_pcidev(uint32_t id, uint8_t bus, uint8_t dev, uint8_t func) "i= d: > %u bdf: %02x.%02x.%02x" > > xen_unmap_pcidev(uint32_t id, uint8_t bus, uint8_t dev, uint8_t func) = "id: > %u bdf: %02x.%02x.%02x" > > +xen_domid_restrict(int err) "err: %u" > > diff --git a/include/hw/xen/xen.h b/include/hw/xen/xen.h > > index 2b1733b..7efcdaa 100644 > > --- a/include/hw/xen/xen.h > > +++ b/include/hw/xen/xen.h > > @@ -21,6 +21,7 @@ enum xen_mode { > > > > extern uint32_t xen_domid; > > extern enum xen_mode xen_mode; > > +extern bool xen_domid_restrict; > > > > extern bool xen_allowed; > > > > diff --git a/include/hw/xen/xen_common.h > b/include/hw/xen/xen_common.h > > index df098c7..4f3bd35 100644 > > --- a/include/hw/xen/xen_common.h > > +++ b/include/hw/xen/xen_common.h > > @@ -152,6 +152,13 @@ static inline int xendevicemodel_set_mem_type( > > return xc_hvm_set_mem_type(dmod, domid, mem_type, first_pfn, nr); > > } > > > > +static inline int xendevicemodel_restrict( > > + xendevicemodel_handle *dmod, domid_t domid) > > +{ > > + errno =3D ENOTTY; > > + return -1; > > +} > > + > > #else /* CONFIG_XEN_CTRL_INTERFACE_VERSION >=3D 40900 */ > > > > #include > > @@ -206,6 +213,19 @@ static inline int xen_modified_memory(domid_t > domid, uint64_t first_pfn, > > return xendevicemodel_modified_memory(xen_dmod, domid, > first_pfn, nr); > > } > > > > +static inline int xen_restrict(domid_t domid) > > +{ > > + int rc =3D xendevicemodel_restrict(xen_dmod, domid); > > + > > + trace_xen_domid_restrict(errno); > > + > > + if (errno =3D=3D ENOTTY) { > > + return 0; > > + } > > + > > + return rc; > > +} > > + > > /* Xen 4.2 through 4.6 */ > > #if CONFIG_XEN_CTRL_INTERFACE_VERSION < 40701 > > > > diff --git a/qemu-options.hx b/qemu-options.hx > > index 99af8ed..4aab077 100644 > > --- a/qemu-options.hx > > +++ b/qemu-options.hx > > @@ -3354,6 +3354,10 @@ DEF("xen-attach", 0, > QEMU_OPTION_xen_attach, > > "-xen-attach attach to existing xen domain\n" > > " xend will use this when starting QEMU\n", > > QEMU_ARCH_ALL) > > +DEF("xen-domid-restrict", 0, QEMU_OPTION_xen_domid_restrict, > > + "-xen-domid-restrict restrict set of available xen operations\= n" > > + " to specified domain id\n", > > + QEMU_ARCH_ALL) >=20 > Please make a note that the QEMU xenpv machine (backend provider for PV > guests) is completely unaffected. Sure. >=20 > Also, does this xendevicemodel_restrict operation restrict any of the > xenstore accesses? >=20 Not yet. The intention is that it will (which is why I am not specific in t= he usage comment) but that facility is not available as yet. The other rest= riction that is also not yet implemented is foreign memory mapping, which w= ill require a new call to be added to libxenforeignmemory. Paul >=20 > > STEXI > > @item -xen-domid @var{id} > > @findex -xen-domid > > @@ -3366,6 +3370,8 @@ Warning: should not be used when xend is in use > (XEN only). > > @findex -xen-attach > > Attach to existing xen domain. > > xend will use this when starting QEMU (XEN only). > > +@findex -xen-domid-restrict > > +Restrict set of available xen operations to specified domain id (XEN o= nly). > > ETEXI > > DEF("no-reboot", 0, QEMU_OPTION_no_reboot, \ > > diff --git a/vl.c b/vl.c > > index 0b4ed52..f46e070 100644 > > --- a/vl.c > > +++ b/vl.c > > @@ -205,6 +205,7 @@ static NotifierList machine_init_done_notifiers =3D > > bool xen_allowed; > > uint32_t xen_domid; > > enum xen_mode xen_mode =3D XEN_EMULATE; > > +bool xen_domid_restrict; > > > > static int has_defaults =3D 1; > > static int default_serial =3D 1; > > @@ -3933,6 +3934,13 @@ int main(int argc, char **argv, char **envp) > > } > > xen_mode =3D XEN_ATTACH; > > break; > > + case QEMU_OPTION_xen_domid_restrict: > > + if (!(xen_available())) { > > + error_report("Option not supported for this target= "); > > + exit(1); > > + } > > + xen_domid_restrict =3D true; > > + break; > > case QEMU_OPTION_trace: > > g_free(trace_file); > > trace_file =3D trace_opt_parse(optarg); > > diff --git a/xen-hvm.c b/xen-hvm.c > > index 4b928cf..335e263 100644 > > --- a/xen-hvm.c > > +++ b/xen-hvm.c > > @@ -1226,6 +1226,14 @@ void xen_hvm_init(PCMachineState *pcms, > MemoryRegion **ram_memory) > > goto err; > > } > > > > + if (xen_domid_restrict) { > > + rc =3D xen_restrict(xen_domid); > > + if (rc < 0) { > > + error_report("failed to restrict: error %d", errno); > > + goto err; > > + } > > + } > > + > > xen_create_ioreq_server(xen_domid, &state->ioservid); > > > > state->exit.notify =3D xen_exit_notifier; > > -- > > 2.1.4 > >