From: "Daniel P. Berrangé" <berrange@redhat.com> To: Stefan Hajnoczi <stefanha@redhat.com> Cc: qemu-devel@nongnu.org, "Philippe Mathieu-Daudé" <philmd@redhat.com>, "Peter Maydell" <peter.maydell@linaro.org>, "Markus Armbruster" <armbru@redhat.com>, "Paolo Bonzini" <pbonzini@redhat.com>, "Eduardo Otubo" <otubo@redhat.com> Subject: Re: [Qemu-devel] [PATCH v2] security.rst: add Security Guide to developer docs Date: Fri, 3 May 2019 11:19:21 +0100 [thread overview] Message-ID: <20190503101921.GC17905@redhat.com> (raw) In-Reply-To: <20190425133503.30847-1-stefanha@redhat.com> On Thu, Apr 25, 2019 at 02:35:03PM +0100, Stefan Hajnoczi wrote: > At KVM Forum 2018 I gave a presentation on security in QEMU: > https://www.youtube.com/watch?v=YAdRf_hwxU8 (video) > https://vmsplice.net/~stefan/stefanha-kvm-forum-2018.pdf (slides) > > This patch adds a security guide to the developer docs. This document > covers things that developers should know about security in QEMU. It is > just a starting point that we can expand on later. I hope it will be > useful as a resource for new contributors and will save code reviewers > from explaining the same concepts many times. I'm wondering if we should split this doc in two parts. The first 50% of it is actually relevant to both QEMU developers and downstream QEMU developers of mgmt apps and/or end users. The latter half is purely of interest to QEMU developers. > Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> > --- > v2: > * Added mention of passthrough USB and PCI devices [philmd] > * Reworded resource limits [philmd] > * Added qemu_log_mask(LOG_GUEST_ERROR) [philmd] > --- > docs/devel/index.rst | 1 + > docs/devel/security.rst | 225 ++++++++++++++++++++++++++++++++++++++++ > 2 files changed, 226 insertions(+) > create mode 100644 docs/devel/security.rst > > diff --git a/docs/devel/index.rst b/docs/devel/index.rst > index ebbab636ce..fd0b5fa387 100644 > --- a/docs/devel/index.rst > +++ b/docs/devel/index.rst > @@ -20,3 +20,4 @@ Contents: > stable-process > testing > decodetree > + security > diff --git a/docs/devel/security.rst b/docs/devel/security.rst > new file mode 100644 > index 0000000000..83c6fb2231 > --- /dev/null > +++ b/docs/devel/security.rst > @@ -0,0 +1,225 @@ > +============== > +Security Guide > +============== > +Overview > +-------- > +This guide covers security topics relevant to developers working on QEMU. It > +includes an explanation of the security requirements that QEMU gives its users, > +the architecture of the code, and secure coding practices. > + > +Security Requirements > +--------------------- > +QEMU supports many different use cases, some of which have stricter security > +requirements than others. The community has agreed on the overall security > +requirements that users may depend on. These requirements define what is > +considered supported from a security perspective. > + > +Virtualization Use Case > +~~~~~~~~~~~~~~~~~~~~~~~ > +The virtualization use case covers cloud and virtual private server (VPS) > +hosting, as well as traditional data center and desktop virtualization. These > +use cases rely on hardware virtualization extensions to execute guest code > +safely on the physical CPU at close-to-native speed. > + > +The following entities are **untrusted**, meaning that they may be buggy or > +malicious: > + > +* Guest > +* User-facing interfaces (e.g. VNC, SPICE, WebSocket) > +* Network protocols (e.g. NBD, live migration) > +* User-supplied files (e.g. disk images, kernels, device trees) > +* Passthrough devices (e.g. PCI, USB) > + > +Bugs affecting these entities are evaluated on whether they can cause damage in > +real-world use cases and treated as security bugs if this is the case. > + > +Non-virtualization Use Case > +~~~~~~~~~~~~~~~~~~~~~~~~~~~ > +The non-virtualization use case covers emulation using the Tiny Code Generator > +(TCG). In principle the TCG and device emulation code used in conjunction with > +the non-virtualization use case should meet the same security requirements as > +the virtualization use case. However, for historical reasons much of the > +non-virtualization use case code was not written with these security > +requirements in mind. > + > +Bugs affecting the non-virtualization use case are not considered security > +bugs at this time. Users with non-virtualization use cases must not rely on > +QEMU to provide guest isolation or any security guarantees. > + > +Architecture > +------------ > +This section describes the design principles that ensure the security > +requirements are met. > + > +Guest Isolation > +~~~~~~~~~~~~~~~ > +Guest isolation is the confinement of guest code to the virtual machine. When > +guest code gains control of execution on the host this is called escaping the > +virtual machine. Isolation also includes resource limits such as throttling of > +CPU, memory, disk, or network. Guests must be unable to exceed their resource > +limits. > + > +QEMU presents an attack surface to the guest in the form of emulated devices. > +The guest must not be able to gain control of QEMU. Bugs in emulated devices > +could allow malicious guests to gain code execution in QEMU. At this point the > +guest has escaped the virtual machine and is able to act in the context of the > +QEMU process on the host. > + > +Guests often interact with other guests and share resources with them. A > +malicious guest must not gain control of other guests or access their data. > +Disk image files and network traffic must be protected from other guests unless > +explicitly shared between them by the user. > + > +Principle of Least Privilege > +~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > +The principle of least privilege states that each component only has access to > +the privileges necessary for its function. In the case of QEMU this means that > +each process only has access to resources belonging to the guest. > + > +The QEMU process should not have access to any resources that are inaccessible > +to the guest. This way the guest does not gain anything by escaping into the > +QEMU process since it already has access to those same resources from within > +the guest. > + > +Following the principle of least privilege immediately fulfills guest isolation > +requirements. For example, guest A only has access to its own disk image file > +``a.img`` and not guest B's disk image file ``b.img``. > + > +In reality certain resources are inaccessible to the guest but must be > +available to QEMU to perform its function. For example, host system calls are > +necessary for QEMU but are not exposed to guests. A guest that escapes into > +the QEMU process can then begin invoking host system calls. > + > +New features must be designed to follow the principle of least privilege. > +Should this not be possible for technical reasons, the security risk must be > +clearly documented so users are aware of the trade-off of enabling the feature. > + > +Isolation mechanisms > +~~~~~~~~~~~~~~~~~~~~ > +Several isolation mechanisms are available to realize this architecture of > +guest isolation and the principle of least privilege. With the exception of > +Linux seccomp, these mechanisms are all deployed by management tools that > +launch QEMU, such as libvirt. They are also platform-specific so they are only > +described briefly for Linux here. > + > +The fundamental isolation mechanism is that QEMU processes must run as > +**unprivileged users**. Sometimes it seems more convenient to launch QEMU as > +root to give it access to host devices (e.g. ``/dev/net/tun``) but this poses a > +huge security risk. File descriptor passing can be used to give an otherwise > +unprivileged QEMU process access to host devices without running QEMU as root. > + > +**SELinux** and **AppArmor** make it possible to confine processes beyond the > +traditional UNIX process and file permissions model. They restrict the QEMU > +process from accessing processes and files on the host system that are not > +needed by QEMU. > + > +**Resource limits** and **cgroup controllers** provide throughput and utilization > +limits on key resources such as CPU time, memory, and I/O bandwidth. > + > +**Linux namespaces** can be used to make process, file system, and other system > +resources unavailable to QEMU. A namespaced QEMU process is restricted to only > +those resources that were granted to it. > + > +**Linux seccomp** is available via the QEMU ``--sandbox`` option. It disables > +system calls that are not needed by QEMU, thereby reducing the host kernel > +attack surface. Break here. Everything above here is useful to QEMU devs, app devs & end users and should be made part of the main QEMU doc - convert it to texi and @include it from qemu-doc.texi, as we do for other stuff under docs/ Everything below here could just be renamed to "secure-coding-practices.rst" and solely target qemu devs. > + > +Secure coding practices > +----------------------- > +At the source code level there are several points to keep in mind. Both > +developers and security researchers must be aware of them so that they can > +develop safe code and audit existing code properly. > + > +General Secure C Coding Practices > +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > +Most CVEs (security bugs) reported against QEMU are not specific to > +virtualization or emulation. They are simply C programming bugs. Therefore > +it's critical to be aware of common classes of security bugs. > + > +There is a wide selection of resources available covering secure C coding. For > +example, the `CERT C Coding Standard > +<https://wiki.sei.cmu.edu/confluence/display/c/SEI+CERT+C+Coding+Standard>`_ > +covers the most important classes of security bugs. > + > +Instead of describing them in detail here, only the names of the most important > +classes of security bugs are mentioned: > + > +* Buffer overflows > +* Use-after-free and double-free > +* Integer overflows > +* Format string vulnerabilities > + > +Some of these classes of bugs can be detected by analyzers. Static analysis is > +performed regularly by Coverity and the most obvious of these bugs are even > +reported by compilers. Dynamic analysis is possible with valgrind, tsan, and > +asan. > + > +Input Validation > +~~~~~~~~~~~~~~~~ > +Inputs from the guest or external sources (e.g. network, files) cannot be > +trusted and may be invalid. Inputs must be checked before using them in a way > +that could crash the program, expose host memory to the guest, or otherwise be > +exploitable by an attacker. > + > +The most sensitive attack surface is device emulation. All hardware register > +accesses and data read from guest memory must be validated. A typical example > +is a device that contains multiple units that are selectable by the guest via > +an index register:: > + > + typedef struct { > + ProcessingUnit unit[2]; > + ... > + } MyDeviceState; > + > + static void mydev_writel(void *opaque, uint32_t addr, uint32_t val) > + { > + MyDeviceState *mydev = opaque; > + ProcessingUnit *unit; > + > + switch (addr) { > + case MYDEV_SELECT_UNIT: > + unit = &mydev->unit[val]; <-- this input wasn't validated! > + ... > + } > + } > + > +If ``val`` is not in range [0, 1] then an out-of-bounds memory access will take > +place when ``unit`` is dereferenced. The code must check that ``val`` is 0 or > +1 and handle the case where it is invalid. > + > +Unexpected Device Accesses > +~~~~~~~~~~~~~~~~~~~~~~~~~~ > +The guest may access device registers in unusual orders or at unexpected > +moments. Device emulation code must not assume that the guest follows the > +typical "theory of operation" presented in driver writer manuals. The guest > +may make nonsense accesses to device registers such as starting operations > +before the device has been fully initialized. > + > +A related issue is that device emulation code must be prepared for unexpected > +device register accesses while asynchronous operations are in progress. A > +well-behaved guest might wait for a completion interrupt before accessing > +certain device registers. Device emulation code must handle the case where the > +guest overwrites registers or submits further requests before an ongoing > +request completes. Unexpected accesses must not cause memory corruption or > +leaks in QEMU. > + > +Invalid device register accesses can be reported with > +``qemu_log_mask(LOG_GUEST_ERROR, ...)``. The ``-d guest_errors`` command-line > +option enables these log messages. > + > +Live migration > +~~~~~~~~~~~~~~ > +Device state can be saved to disk image files and shared with other users. > +Live migration code must validate inputs when loading device state so an > +attacker cannot gain control by crafting invalid device states. Device state > +is therefore considered untrusted even though it is typically generated by QEMU > +itself. > + > +Guest Memory Access Races > +~~~~~~~~~~~~~~~~~~~~~~~~~ > +Guests with multiple vCPUs may modify guest RAM while device emulation code is > +running. Device emulation code must copy in descriptors and other guest RAM > +structures and only process the local copy. This prevents > +time-of-check-to-time-of-use (TOCTOU) race conditions that could cause QEMU to > +crash when a vCPU thread modifies guest RAM while device emulation is > +processing it. > -- > 2.20.1 > Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
WARNING: multiple messages have this Message-ID (diff)
From: "Daniel P. Berrangé" <berrange@redhat.com> To: Stefan Hajnoczi <stefanha@redhat.com> Cc: "Eduardo Otubo" <otubo@redhat.com>, "Peter Maydell" <peter.maydell@linaro.org>, qemu-devel@nongnu.org, "Markus Armbruster" <armbru@redhat.com>, "Paolo Bonzini" <pbonzini@redhat.com>, "Philippe Mathieu-Daudé" <philmd@redhat.com> Subject: Re: [Qemu-devel] [PATCH v2] security.rst: add Security Guide to developer docs Date: Fri, 3 May 2019 11:19:21 +0100 [thread overview] Message-ID: <20190503101921.GC17905@redhat.com> (raw) Message-ID: <20190503101921.kqIjUvPug4QlZV6mLfN4JeVkG_vOv4RTFUGXdUSW4PU@z> (raw) In-Reply-To: <20190425133503.30847-1-stefanha@redhat.com> On Thu, Apr 25, 2019 at 02:35:03PM +0100, Stefan Hajnoczi wrote: > At KVM Forum 2018 I gave a presentation on security in QEMU: > https://www.youtube.com/watch?v=YAdRf_hwxU8 (video) > https://vmsplice.net/~stefan/stefanha-kvm-forum-2018.pdf (slides) > > This patch adds a security guide to the developer docs. This document > covers things that developers should know about security in QEMU. It is > just a starting point that we can expand on later. I hope it will be > useful as a resource for new contributors and will save code reviewers > from explaining the same concepts many times. I'm wondering if we should split this doc in two parts. The first 50% of it is actually relevant to both QEMU developers and downstream QEMU developers of mgmt apps and/or end users. The latter half is purely of interest to QEMU developers. > Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> > --- > v2: > * Added mention of passthrough USB and PCI devices [philmd] > * Reworded resource limits [philmd] > * Added qemu_log_mask(LOG_GUEST_ERROR) [philmd] > --- > docs/devel/index.rst | 1 + > docs/devel/security.rst | 225 ++++++++++++++++++++++++++++++++++++++++ > 2 files changed, 226 insertions(+) > create mode 100644 docs/devel/security.rst > > diff --git a/docs/devel/index.rst b/docs/devel/index.rst > index ebbab636ce..fd0b5fa387 100644 > --- a/docs/devel/index.rst > +++ b/docs/devel/index.rst > @@ -20,3 +20,4 @@ Contents: > stable-process > testing > decodetree > + security > diff --git a/docs/devel/security.rst b/docs/devel/security.rst > new file mode 100644 > index 0000000000..83c6fb2231 > --- /dev/null > +++ b/docs/devel/security.rst > @@ -0,0 +1,225 @@ > +============== > +Security Guide > +============== > +Overview > +-------- > +This guide covers security topics relevant to developers working on QEMU. It > +includes an explanation of the security requirements that QEMU gives its users, > +the architecture of the code, and secure coding practices. > + > +Security Requirements > +--------------------- > +QEMU supports many different use cases, some of which have stricter security > +requirements than others. The community has agreed on the overall security > +requirements that users may depend on. These requirements define what is > +considered supported from a security perspective. > + > +Virtualization Use Case > +~~~~~~~~~~~~~~~~~~~~~~~ > +The virtualization use case covers cloud and virtual private server (VPS) > +hosting, as well as traditional data center and desktop virtualization. These > +use cases rely on hardware virtualization extensions to execute guest code > +safely on the physical CPU at close-to-native speed. > + > +The following entities are **untrusted**, meaning that they may be buggy or > +malicious: > + > +* Guest > +* User-facing interfaces (e.g. VNC, SPICE, WebSocket) > +* Network protocols (e.g. NBD, live migration) > +* User-supplied files (e.g. disk images, kernels, device trees) > +* Passthrough devices (e.g. PCI, USB) > + > +Bugs affecting these entities are evaluated on whether they can cause damage in > +real-world use cases and treated as security bugs if this is the case. > + > +Non-virtualization Use Case > +~~~~~~~~~~~~~~~~~~~~~~~~~~~ > +The non-virtualization use case covers emulation using the Tiny Code Generator > +(TCG). In principle the TCG and device emulation code used in conjunction with > +the non-virtualization use case should meet the same security requirements as > +the virtualization use case. However, for historical reasons much of the > +non-virtualization use case code was not written with these security > +requirements in mind. > + > +Bugs affecting the non-virtualization use case are not considered security > +bugs at this time. Users with non-virtualization use cases must not rely on > +QEMU to provide guest isolation or any security guarantees. > + > +Architecture > +------------ > +This section describes the design principles that ensure the security > +requirements are met. > + > +Guest Isolation > +~~~~~~~~~~~~~~~ > +Guest isolation is the confinement of guest code to the virtual machine. When > +guest code gains control of execution on the host this is called escaping the > +virtual machine. Isolation also includes resource limits such as throttling of > +CPU, memory, disk, or network. Guests must be unable to exceed their resource > +limits. > + > +QEMU presents an attack surface to the guest in the form of emulated devices. > +The guest must not be able to gain control of QEMU. Bugs in emulated devices > +could allow malicious guests to gain code execution in QEMU. At this point the > +guest has escaped the virtual machine and is able to act in the context of the > +QEMU process on the host. > + > +Guests often interact with other guests and share resources with them. A > +malicious guest must not gain control of other guests or access their data. > +Disk image files and network traffic must be protected from other guests unless > +explicitly shared between them by the user. > + > +Principle of Least Privilege > +~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > +The principle of least privilege states that each component only has access to > +the privileges necessary for its function. In the case of QEMU this means that > +each process only has access to resources belonging to the guest. > + > +The QEMU process should not have access to any resources that are inaccessible > +to the guest. This way the guest does not gain anything by escaping into the > +QEMU process since it already has access to those same resources from within > +the guest. > + > +Following the principle of least privilege immediately fulfills guest isolation > +requirements. For example, guest A only has access to its own disk image file > +``a.img`` and not guest B's disk image file ``b.img``. > + > +In reality certain resources are inaccessible to the guest but must be > +available to QEMU to perform its function. For example, host system calls are > +necessary for QEMU but are not exposed to guests. A guest that escapes into > +the QEMU process can then begin invoking host system calls. > + > +New features must be designed to follow the principle of least privilege. > +Should this not be possible for technical reasons, the security risk must be > +clearly documented so users are aware of the trade-off of enabling the feature. > + > +Isolation mechanisms > +~~~~~~~~~~~~~~~~~~~~ > +Several isolation mechanisms are available to realize this architecture of > +guest isolation and the principle of least privilege. With the exception of > +Linux seccomp, these mechanisms are all deployed by management tools that > +launch QEMU, such as libvirt. They are also platform-specific so they are only > +described briefly for Linux here. > + > +The fundamental isolation mechanism is that QEMU processes must run as > +**unprivileged users**. Sometimes it seems more convenient to launch QEMU as > +root to give it access to host devices (e.g. ``/dev/net/tun``) but this poses a > +huge security risk. File descriptor passing can be used to give an otherwise > +unprivileged QEMU process access to host devices without running QEMU as root. > + > +**SELinux** and **AppArmor** make it possible to confine processes beyond the > +traditional UNIX process and file permissions model. They restrict the QEMU > +process from accessing processes and files on the host system that are not > +needed by QEMU. > + > +**Resource limits** and **cgroup controllers** provide throughput and utilization > +limits on key resources such as CPU time, memory, and I/O bandwidth. > + > +**Linux namespaces** can be used to make process, file system, and other system > +resources unavailable to QEMU. A namespaced QEMU process is restricted to only > +those resources that were granted to it. > + > +**Linux seccomp** is available via the QEMU ``--sandbox`` option. It disables > +system calls that are not needed by QEMU, thereby reducing the host kernel > +attack surface. Break here. Everything above here is useful to QEMU devs, app devs & end users and should be made part of the main QEMU doc - convert it to texi and @include it from qemu-doc.texi, as we do for other stuff under docs/ Everything below here could just be renamed to "secure-coding-practices.rst" and solely target qemu devs. > + > +Secure coding practices > +----------------------- > +At the source code level there are several points to keep in mind. Both > +developers and security researchers must be aware of them so that they can > +develop safe code and audit existing code properly. > + > +General Secure C Coding Practices > +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > +Most CVEs (security bugs) reported against QEMU are not specific to > +virtualization or emulation. They are simply C programming bugs. Therefore > +it's critical to be aware of common classes of security bugs. > + > +There is a wide selection of resources available covering secure C coding. For > +example, the `CERT C Coding Standard > +<https://wiki.sei.cmu.edu/confluence/display/c/SEI+CERT+C+Coding+Standard>`_ > +covers the most important classes of security bugs. > + > +Instead of describing them in detail here, only the names of the most important > +classes of security bugs are mentioned: > + > +* Buffer overflows > +* Use-after-free and double-free > +* Integer overflows > +* Format string vulnerabilities > + > +Some of these classes of bugs can be detected by analyzers. Static analysis is > +performed regularly by Coverity and the most obvious of these bugs are even > +reported by compilers. Dynamic analysis is possible with valgrind, tsan, and > +asan. > + > +Input Validation > +~~~~~~~~~~~~~~~~ > +Inputs from the guest or external sources (e.g. network, files) cannot be > +trusted and may be invalid. Inputs must be checked before using them in a way > +that could crash the program, expose host memory to the guest, or otherwise be > +exploitable by an attacker. > + > +The most sensitive attack surface is device emulation. All hardware register > +accesses and data read from guest memory must be validated. A typical example > +is a device that contains multiple units that are selectable by the guest via > +an index register:: > + > + typedef struct { > + ProcessingUnit unit[2]; > + ... > + } MyDeviceState; > + > + static void mydev_writel(void *opaque, uint32_t addr, uint32_t val) > + { > + MyDeviceState *mydev = opaque; > + ProcessingUnit *unit; > + > + switch (addr) { > + case MYDEV_SELECT_UNIT: > + unit = &mydev->unit[val]; <-- this input wasn't validated! > + ... > + } > + } > + > +If ``val`` is not in range [0, 1] then an out-of-bounds memory access will take > +place when ``unit`` is dereferenced. The code must check that ``val`` is 0 or > +1 and handle the case where it is invalid. > + > +Unexpected Device Accesses > +~~~~~~~~~~~~~~~~~~~~~~~~~~ > +The guest may access device registers in unusual orders or at unexpected > +moments. Device emulation code must not assume that the guest follows the > +typical "theory of operation" presented in driver writer manuals. The guest > +may make nonsense accesses to device registers such as starting operations > +before the device has been fully initialized. > + > +A related issue is that device emulation code must be prepared for unexpected > +device register accesses while asynchronous operations are in progress. A > +well-behaved guest might wait for a completion interrupt before accessing > +certain device registers. Device emulation code must handle the case where the > +guest overwrites registers or submits further requests before an ongoing > +request completes. Unexpected accesses must not cause memory corruption or > +leaks in QEMU. > + > +Invalid device register accesses can be reported with > +``qemu_log_mask(LOG_GUEST_ERROR, ...)``. The ``-d guest_errors`` command-line > +option enables these log messages. > + > +Live migration > +~~~~~~~~~~~~~~ > +Device state can be saved to disk image files and shared with other users. > +Live migration code must validate inputs when loading device state so an > +attacker cannot gain control by crafting invalid device states. Device state > +is therefore considered untrusted even though it is typically generated by QEMU > +itself. > + > +Guest Memory Access Races > +~~~~~~~~~~~~~~~~~~~~~~~~~ > +Guests with multiple vCPUs may modify guest RAM while device emulation code is > +running. Device emulation code must copy in descriptors and other guest RAM > +structures and only process the local copy. This prevents > +time-of-check-to-time-of-use (TOCTOU) race conditions that could cause QEMU to > +crash when a vCPU thread modifies guest RAM while device emulation is > +processing it. > -- > 2.20.1 > Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
next prev parent reply other threads:[~2019-05-03 10:19 UTC|newest] Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top 2019-04-25 13:35 [Qemu-devel] [PATCH v2] security.rst: add Security Guide to developer docs Stefan Hajnoczi 2019-04-25 13:35 ` Stefan Hajnoczi 2019-05-01 16:20 ` Stefan Hajnoczi 2019-05-01 16:20 ` Stefan Hajnoczi 2019-05-03 8:14 ` Stefano Garzarella 2019-05-03 8:14 ` Stefano Garzarella 2019-05-03 9:04 ` Alex Bennée 2019-05-03 9:04 ` Alex Bennée 2019-05-03 10:10 ` Philippe Mathieu-Daudé 2019-05-03 10:10 ` Philippe Mathieu-Daudé 2019-05-03 17:32 ` Stefan Hajnoczi 2019-05-03 17:32 ` Stefan Hajnoczi 2019-05-03 10:19 ` Daniel P. Berrangé [this message] 2019-05-03 10:19 ` Daniel P. Berrangé 2019-05-03 10:28 ` Peter Maydell 2019-05-03 10:28 ` Peter Maydell 2019-05-03 10:35 ` Daniel P. Berrangé 2019-05-03 10:35 ` Daniel P. Berrangé 2019-05-03 17:30 ` Stefan Hajnoczi 2019-05-03 17:30 ` Stefan Hajnoczi
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20190503101921.GC17905@redhat.com \ --to=berrange@redhat.com \ --cc=armbru@redhat.com \ --cc=otubo@redhat.com \ --cc=pbonzini@redhat.com \ --cc=peter.maydell@linaro.org \ --cc=philmd@redhat.com \ --cc=qemu-devel@nongnu.org \ --cc=stefanha@redhat.com \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).