docs/qemu-deprivilege: Revise and update with status and future plans
PATCH v4 1/6 docs/qemu-deprivilege: Revise and update with status and future plans
docs/qemu-deprivilege.txt had some basic instructions for using dm_restrict, but it was incomplete, misleading, and stale.
Update the docs in a number of ways.
First, separate user-facing documentation and technical description into docs/features and docs/design, respectively.
In the feature doc:
* Introduce a section mentioning minimim versions of Linux, Xen, and qemu required (TBD)
* Fix the discussion of qemu userid. Mention xen-qemuuser-range-base, and provide example shell code that actually has some hope of working (instead of failing out after creating 900 userids).
* Describe how to enable restrictions, as well as features which probably don't or definitely don't work.
In the design doc, introduce a "Technical Details" section which describes specifically what restrictions are currently done, and also what restrictions we are looking at doing in the future.
The idea here is that as we implement the various items for the future, we move them from "Restrictions still to do" to "Restrictions done". This can also act as a design document -- a place for public discussion of what can or should be done and how.
Also add an entry to SUPPORT.md.
---
Changes since v3:
- Fix typo (32->16)
- Use an example value not close to the nobody uids, but still a multiple of 2^16.
- Mention that using a multiple of 2^16 may have advantages.
- Have the example create a group as well
- Reorganize two comments on the "range-base" method for clarity
Changes since v2:
- Extraneous privcmd / evtchn instances aren't closed
- Expand description of how to test fd deprivileging
- Rework and clarify two namespace sections, give reference for QEMU NAK
- Add more information about migration technical challenges
- In UID section, mention possibility of container ID collisions.
- Fix name of design document.
- Add SUPPORT.md statement. Specify Linux, to make sure that FreeBSD is evaluated separately.
- Mention that -sandbox is a blacklist and why
Changes since v1:
- Break into two, and move into appropriate directories (rather than 'misc')
- Updated version requirements
- Distinguish between features which "don't yet work" and features which we never expect to work
- Update description of xen-restrict functionality
- Reorder and expand further restrictions
- Make it more clear which restrictions are available on Linux only
- Include detailed description of how to kill a process
- Add RLIMIT_NPROC as something we can do without further changes to qemu
- Document the need to check for the sandbox feature before using it
Thank you to Ross Lagerwall, whose description of what XenServer is doing formed much of the basis for the text here.
--- docs/designs/qemu-deprivilege.md | 322 ++++++++++++++++++++++++++ docs/features/qemu-deprivilege.pandoc | 101 ++++++++ docs/misc/qemu-deprivilege.txt | 36 --- 3 files changed, 423 insertions(+), 36 deletions(-) create mode 100644 docs/designs/qemu-deprivilege.md create mode 100644 docs/features/qemu-deprivilege.pandoc delete mode 100644 docs/misc/qemu-deprivilege.txt
docs/designs/qemu-deprivilege.md new file +322 -0
@@ -0,0 +1,322 @@ +# Introduction + +The goal of deprilvileging qemu is this: Even if there is a bug (for +example in qemu) which permits a domain to gain control of the device +model, the compromised device model process is prevented from +violating the system's overall security properties. Ie, a guest +cannot "escape" from the virtualisation by using a qemu bug. + +This document lists the various technical measures which we either +have taken, or plan to take to effect this goal. Some of them are +required to be considered secure (that is, there are known attack +vectors which they close); others are "just in case" (that is, there +are no known attack vectors, but we perform the restrictions to reduce +the possibility of unknown attack vectors). + +# Restrictions done + +The following restrictions are currently implemented. + +## Having qemu switch user + +'''Description''': As mentioned above, having QEMU switch to a +non-root user, one per domain id. Not being the root user limits what +a compromised QEMU process can do to the system, and having one user +per domain id limits what a comprimised QEMU process can do to the +QEMU processes of other VMs. + +'''Implementation''': The toolstack adds the following to the qemu command-line: + + -runas <uid>:<gid> + +'''How to test''': + + grep /proc/<qpid>/status [UG]id + +'''Testing Status''': Not tested + +## Xen library / file-descriptor restrictions + +'''Description''': Close and restrict Xen-related file descriptors. +Specifically: + * Close all xenstore-related file descriptors + * Make sure that all open instances of `privcmd` and `evtchn` file +descriptors have had `IOCTL_PRIVCMD_RESTRICT` and +`IOCTL_EVTCHN_RESTRICT_DOMID` ioctls called on them, respectively. + +FIXME: Double-check the correctness of the above + +'''Implementation''': Toolstack adds the following to the qemu command-line: + + -xen-domid-restrict + +'''How to test''': + +Use `fishdescriptor` to pull a file descriptor from a running QEMU, +then use `depriv-fd-checker` to check that it has the desired +properties, and that hypercalls which are meant to fail do fail. (In +Debian `fishdescriptor` can be found in the binary package +`chiark-scripts`; the `depriv-fd-checker` is included in the Xen +source tree.) + +'''Testing status''': Tested + +# Restrictions / improvements still to do + +This lists potential restrictions still to do. It is meant to be +listed in order of ease of implementation, with low-hanging fruit +first. + +## Chroot + +'''Description''': Qemu runs in its own chroot, such that even if it +could call an 'open' command of some sort, there would be nothing for +it to see. + +'''Implementation''': The toolstack creates a directory in the libxl "run-dir"; e.g. +`/var/run/xen/qemu-root-<domid>` + +Then adds the following to the qemu command-line: + + -chroot /var/run/xen/qemu-root-<domid> + +'''How to test''': Check `/proc/<qpid>/root` + +'''Tested''': Not tested + +## Namespaces for unused functionality (Linux only) + +'''Description''': QEMU doesn't use the functionality associated with +mount and IPC namespaces. (IPC namespaces contol non-file-based IPC +mechanisms within the kernel; unix and network sockets are not +affected by this.) Making separate namespaces for these for QEMU +won't affect normal operation, but it does mean that even if other +restrictions fail, the process won't be able to even name system mount +points or existing non-file-based IPC descriptors to attempt to attack +them. + +'''Implementation''': + +In theory this could be done in QEMU (similar to -sandbox, -runas, +-chroot, and so on), but a patch doing this in QEMU was NAKed upstream +(see [qemu-namespaces]). They preferred that this was done as a setup step by +whatever executes QEMU; i.e., have the process which exec's QEMU first +call: + + unshare(CLONE_NEWNS | CLONE_NEWIPC) + +'''How to test''': Check `/proc/<qpid>/ns/[ipc,mnt]` + +'''Tested''': Not tested + +[qemu-namespaces]: https://lists.gnu.org/archive/html/qemu-devel/2017-10/msg04723.html + +### Basic RLIMITs + +'''Description''': A number of limits on the resources that a given +process / userid is allowed to consume. These can limit the ability +of a compromised QEMU process to DoS domain 0 by exhausting various +resources available to it. + +'''Implementation''' + +Limits that can be implemented immediately without much effort: + - RLIMIT_FSIZE` (file size) to 256KiB. + - RLIMIT_NPROC (after uid changes to a unique uid) + +Probably not necessary but why not: + - RLIMIT_CORE: 0 + - RLIMIT_MSGQUEUE: 0 + - RLIMIT_LOCKS: 0 + - RLIMIT_MEMLOCK: 0 + +Note: mlock() is used by QEMU only when both "realtime" and "mlock" +are specified; this does not apply to QEMU running as a Xen DM. + +'''How to test''': Check `/proc/<qpid>/limits` + +'''Tested''': Not tested + +### Further RLIMITs + +RLIMIT_AS limits the total amount of memory; but this includes the +virtual memory which QEMU uses as a mapcache. xen-mapcache.c already +fiddles with this; it would be straightforward to make it *set* the +rlimit to what it thinks a sensible limit is. + +Other things that would take some cleverness / changes to QEMU to +utilize due to ordering constrants: + - RLIMIT_NOFILES (after all necessary files are opened) + +### libxl UID cleanup + +'''Description''': Domain IDs are reused, and thus restricted UIDs are +reused. If a compromised QEMU can fork (due to seccomp or +RLIMIT_NPROC limits being ineffective for some reason), it may avoid +being killed when its domain dies, then wait until the domain ID is +reused again, at which point it will have control over the domain in +question (which probably belongs to someone else). + +libxl should kill all UIDs associated with a domain both when the VM +is destroyed, and before starting a VM with the same UID. + +'''Implementation''': This is unnecessarily tricky. + +The kill() system call can have three kinds of targets: + - A single pid + - A process group + - "Every process except me to which I am allowed to send a signal" (-1) + +Targeting a single pid is racy and likely to be beaten by the +following loop: + + while(1) { + if(fork()) + _exit(0); + } + +That is, by the time you've read the process list and found the +process id you want to kill, that process has exited and there is a +new process whose pid you don't know about. + +Targeting a process group will be ineffective, as unprivileged +processes are allowed to make their own process groups. + +kill(-1) can be used but must be done with care. Consider the +following code, for example: + + setuid(target_uid); + kill(-1, 9); + +This looks like it will do the trick; but by setting all of the user +ids (effective, real, and saved), it opens the 'killing' process up to +being killed by the target process: + + while(1) { + if(fork()) + _exit(0); + else + kill(-1, 9); + } + +Fortunately there is an assymetry we can take advantage of. From the +POSIX spec: + +> For a process to have permission to send a signal to a process +> designated by pid, unless the sending process has appropriate +> privileges, the real or effective user ID of the sending process shall +> match the real or saved set-user-ID of the receiving process. + +The solution is to allocate a second "reaper" uid that is only used to kill +target processes. We set the euid of the killing process to the `target_uid`, +but the ruid of the killing process to `reaper_uid`, leaving the suid of the +killing process as 0: + + setresuid(reaper_uid, target_uid, 0); + kill(-1, 9); + +NOTE: We cannot use `setreuid(reaper_uid, target_uid)` here, as that +will set *both* euid *and* suid to `target_uid`, making the killing +process vulnerable to the target process again. + +Since this will kill all other `reaper_uid` processes as well, we must +either allocate a separate `reaper_uid` per domain, or use locking to +ensure that only one killing process is active at a time. + +## libxl: Treat QMP connection as untrusted + +'''Description''': Currently libxl talks with QEMU via QMP; but its +interactions have not historically considered from a security point of +view. For example, qmp_synchronous_send() waits for a response from +QEMU, which a compromised QEMU could simply not send (thus preventing +the toolstack from making forward progress). + +'''Implementation''': Audit toolstack interactions with QEMU which +happen after the guest has started running, and assume QEMU has been +compromised. + +### seccomp filtering (Linux only) + +'''Description''': Turn on seccomp filtering to disable syscalls which +QEMU doesn't need. + +'''Implementation''': Enable from the command-line: + + -sandbox on,obsolete=deny,elevateprivileges=allow,spawn=deny,resourcecontrol=deny + +`elevateprivileges` is currently required to allow `-runas` to work. +Removing this requirement would mean making sure that the uid change +happened before the seccomp2 call, perhaps by changing the uid before +executing QEMU. (But this would then require other changes to create +the QMP socket, VNC socket, and so on). + +It should be noted that `-sandbox` is implemented as a blacklist, not +a whitelist; that is, it disables known-unsed functionality which may +be harmful, rather than disabling all functionality except that known +to be safe and needed. This is unfortunately necessary since qemu +doesn't know what system calls libraries might end up making. (See +[lwn-seccomp] for a more complete discussion.) + +This feature is not on by default and may not be available in all +environments. We therefore need to either: + 1. Require that this feature be enabled to build qemu + 2. Check for `-sandbox` support at runtime before + +[lwn-seccomp]: https://lwn.net/Articles/738694/ + +### Disks + +The chroot (and seccomp?) happens late enough such that QEMU can +initialize itself and open its disks. If you want to add a disk at run +time via or insert a CD, you can't pass a path because QEMU is +chrooted. Instead use the add-fd QMP command and use +/dev/fdset/<fdset-id> as the path. + +A further layer of restriction could be to set RLIMIT_NOFILES to '0', +and hand all disks over QMP. + +## Migration + +When calling xen-save-devices-state, since QEMU is running in a chroot +it is not useful to pass a filename (it doesn't even have write access +inside the chroot). Instead, give it an open fd using the add-fd +mechanism. + +Additionally, all the restrictions need to be applied to the qemu +started up on the post-migration side. One issue that needs to be +solved is how to signal the toolstack on restore that qemu is ready +for the domain to be started (since this is normally done via +xenstore, and at this point the xenstore connections will have been +closed). + +### Network namespacing (Linux only) + +Enter QEMU into its own network namespace (in addition to mount & IPC +namespaces): + + unshare(CLONE_NEWNET); + +QEMU does actually use the network namespace as a Xen DM for two +purposes: 1) To set up network tap devices 2) To open vnc connections. + +#### Network + +If QEMU runs in its own network namespace, it can't open the tap +device itself because the interface won't be visible outside of its +own namespace. So instead, have the toolstack open the device and pass +it as an fd on the command-line: + + -device rtl8139,netdev=tapnet0,mac=... -netdev tap,id=tapnet0,fd=<tapfd> + +#### VNC + +If QEMU runs in its own network namespace, it is not straightforward +to listen on a TCP socket outside of its own network namespace. One +option would be to use VNC over a UNIX socket: + + -vnc unix:/var/run/xen/vnc-<domid> + +However, this would break functionality in the general case; I think +we need to have the toolstack open a socket and pass the fd to QEMU +(which requires changes to QEMU). +
docs/features/qemu-deprivilege.pandoc new file +101 -0
101 lines not on this page · raw message
docs/misc/qemu-deprivilege.txt deleted +0 -36
@@ -1,36 +0,0 @@ -For security reasons, libxl tries to pass a non-root username to QEMU as -argument. During initialization QEMU calls setuid and setgid with the -user ID and the group ID of the user passed as argument. -Libxl looks for the following users in this order: - -1) a user named "xen-qemuuser-domid$domid", -Where $domid is the domid of the domain being created. -This requires the reservation of 65535 uids from xen-qemuuser-domid1 -to xen-qemuuser-domid65535. To use this mechanism, you might want to -create a large number of users at installation time. For example: - -for ((i=1; i<65536; i++)) -do - adduser --no-create-home --system xen-qemuuser-domid$i -done - -You might want to consider passing --group to adduser to create a new -group for each new user. - - -2) a user named "xen-qemuuser-shared" -As a fall back if both 1) fails, libxl will use a single user for -all QEMU instances. The user is named xen-qemuuser-shared. This is -less secure but still better than running QEMU as root. Using this is as -simple as creating just one more user on your host: - -adduser --no-create-home --system xen-qemuuser-shared - - -3) root -As a last resort, libxl will start QEMU as root. - - -Please note that running QEMU as non-root causes several features like -migration and PCI passthrough to not work properly and may prevent the guest -from booting.
Signature · 7 lines
2.19.1 _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
PATCH v4 2/6 SUPPORT.md: Add qemu-depriv section
---
Changes since v3:
- Moved from the qemu-depriv doc patches.
- Reword to include the possibility of having a non-dom0 "devicemodel" domain which may want to be protected
- Specify Linux dom0 as the currently-tech-supported window
--- SUPPORT.md | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+)
SUPPORT.md +20 -0
@@ -525,6 +525,26 @@ Vulnerabilities of a device model stub domain to a hostile driver domain (either compromised or untrusted) are excluded from security support. +### Device Model Deprivileging + + Status, Linux dom0: Tech Preview, with limited support + +This means adding extra restrictions to a device model in order to +prevent a compromised device model from attack the rest of the domain +it's running in (normally dom0). + +"Tech preview with limited support" means we will not issue XSAs for +the _additional_ functionality provided by the feature; but we will +issue XSAs in the event that enabling this feature opens up a security +hole that would not be present without the feature disabled. + +For example, while this is classified as tech preview, a bug in libxl +which failed to change the user ID of QEMU would not receive an XSA, +since without this feature the user ID wouldn't be changed. But a +change which made it possible for a compromised guest to read +arbitrary files on the host filesystem without compromising QEMU would +be issued an XSA, since that does weaken security. + ### KCONFIG Expert Status: Experimental
Signature · 7 lines
2.19.1 _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [PATCH v4 2/6] SUPPORT.md: Add qemu-depriv section
Quoted 41 lines
-----Original Message-----
From: Xen-devel [mailto:[email protected]] On Behalf Of George Dunlap
Sent: 05 November 2018 18:07
To: [email protected]
Cc: Stefano Stabellini <[email protected]>; Wei Liu <[email protected]>; Konrad Wilk <[email protected]>; Andrew Cooper <[email protected]>; Tim (Xen.org) <[email protected]>; George Dunlap <[email protected]>; Ross Lagerwall <[email protected]>; Julien Grall <[email protected]>; Jan Beulich <[email protected]>; Anthony Perard <[email protected]>; Ian Jackson <[email protected]>
Subject: [Xen-devel] [PATCH v4 2/6] SUPPORT.md: Add qemu-depriv sectionSigned-off-by George Dunlap <[email protected]>---
Changes since v3:
- Moved from the qemu-depriv doc patches.
- Reword to include the possibility of having a non-dom0 "devicemodel" domain which may want to be protected
- SpecifyLinux dom0as the currently-tech-supported windowCC Ian Jackson <[email protected]>CC Wei Liu <[email protected]>CC Andrew Cooper <[email protected]>CC Jan Beulich <[email protected]>CC Tim Deegan <[email protected]>CC Konrad Wilk <[email protected]>CC Stefano Stabellini <[email protected]>CC Julien Grall <[email protected]>CC Anthony Perard <[email protected]>CC Ross Lagerwall <[email protected]>--- SUPPORT.md | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+)SUPPORT.md +0
-0
SUPPORT.md Vulnerabilities of a device model stub domain · line 533
Quoted 9 lines
SUPPORT.md +6
-0@@ -525,6 +525,26 @@ Vulnerabilities of a device model stub domain to a hostile driver domain (either compromised or untrusted) are excluded from security support. +### Device Model Deprivileging + + Status, Linux dom0: Tech Preview, with limited support + +This means adding extra restrictions to a device model in order to +prevent a compromised device model from attack the rest of the domain
s/attack/attacking/
Paul
Quoted 21 lines
SUPPORT.md +14
-1+it's running in (normally dom0). + +"Tech preview with limited support" means we will not issue XSAs for +the _additional_ functionality provided by the feature; but we will +issue XSAs in the event that enabling this feature opens up a security +hole that would not be present without the feature disabled. + +For example, while this is classified as tech preview, a bug in libxl +which failed to change the user ID of QEMU would not receive an XSA, +since without this feature the user ID wouldn't be changed. But a +change which made it possible for a compromised guest to read +arbitrary files on the host filesystem without compromising QEMU would +be issued an XSA, since that does weaken security. + ### KCONFIG Expert Status: Experimental --2.19.1
Quoted 4 lines
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [PATCH v4 2/6] SUPPORT.md: Add qemu-depriv section
Quoted 44 lines · Paul Durrant
Quoted 44 lines
-----Original Message-----
From: Xen-devel [mailto:[email protected]] On Behalf Of George Dunlap
Sent: 05 November 2018 18:07
To: [email protected]
Cc: Stefano Stabellini <[email protected]>; Wei Liu <[email protected]>; Konrad Wilk <[email protected]>; Andrew Cooper <[email protected]>; Tim (Xen.org) <[email protected]>; George Dunlap <[email protected]>; Ross Lagerwall <[email protected]>; Julien Grall <[email protected]>; Jan Beulich <[email protected]>; Anthony Perard <[email protected]>; Ian Jackson <[email protected]>
Subject: [Xen-devel] [PATCH v4 2/6] SUPPORT.md: Add qemu-depriv sectionSigned-off-by George Dunlap <[email protected]>---
Changes since v3:
- Moved from the qemu-depriv doc patches.
- Reword to include the possibility of having a non-dom0 "devicemodel" domain which may want to be protected
- SpecifyLinux dom0as the currently-tech-supported windowCC Ian Jackson <[email protected]>CC Wei Liu <[email protected]>CC Andrew Cooper <[email protected]>CC Jan Beulich <[email protected]>CC Tim Deegan <[email protected]>CC Konrad Wilk <[email protected]>CC Stefano Stabellini <[email protected]>CC Julien Grall <[email protected]>CC Anthony Perard <[email protected]>CC Ross Lagerwall <[email protected]>--- SUPPORT.md | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+)SUPPORT.md +0
-0@@ -525,6 +525,26 @@ Vulnerabilities of a device model stub domain to a hostile driver domain (either compromised or untrusted) are excluded from security support.
Quoted 8 lines
SUPPORT.md Vulnerabilities of a device model stub domain · line 533
Quoted 6 lines
SUPPORT.md +6
-0+### Device Model Deprivileging + + Status, Linux dom0: Tech Preview, with limited support + +This means adding extra restrictions to a device model in order to +prevent a compromised device model from attack the rest of the domains/attack/attacking/
Yeah, this paragraph in particular seems to have challenged by ability to form grammatically-correct English sentences... anyway thanks, I'll fix it up on check-in.
-George
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [PATCH v4 2/6] SUPPORT.md: Add qemu-depriv section
In reply to George DunlapGeorge Dunlap writes ("[PATCH v4 2/6] SUPPORT.md: Add qemu-depriv section"):
Quoted 1 line
Signed-off-by George Dunlap <[email protected]>
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
PATCH v4 3/6 tools/dm_restrict: Ask QEMU to chroot
In reply to George DunlapWhen dm_restrict is enabled, ask QEMU to chroot into an empty directory.
* Create /var/run/qemu/root-domid (deleting the old one if it's there)
* Pass the -chroot option to QEMU
Rather than running rm -rf on the directory before creating it (since there is no library function to do this), simply rmdir the directory, relying on the fact that the previous QEMU instance, if properly restricted, shouldn't have been able to write anything anyway.
---
Changes since v2:
- Style fixes
- Testing moved to a different patch
--- docs/designs/qemu-deprivilege.md | 12 +++++----- tools/libxl/libxl_dm.c | 41 +++++++++++++++++++++++++++++++- 2 files changed, 46 insertions(+), 7 deletions(-)
docs/designs/qemu-deprivilege.md +6 -6
@@ -61,12 +61,6 @@ source tree.) '''Testing status''': Tested -# Restrictions / improvements still to do - -This lists potential restrictions still to do. It is meant to be -listed in order of ease of implementation, with low-hanging fruit -first. - ## Chroot '''Description''': Qemu runs in its own chroot, such that even if it
@@ -84,6 +78,12 @@ Then adds the following to the qemu command-line: '''Tested''': Not tested +## Restrictions / improvements still to do + +This lists potential restrictions still to do. It is meant to be +listed in order of ease of implementation, with low-hanging fruit +first. + ## Namespaces for unused functionality (Linux only) '''Description''': QEMU doesn't use the functionality associated with
tools/libxl/libxl_dm.c +40 -1
49 lines not on this page · raw message
Signature · 7 lines
2.19.1 _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [PATCH v4 3/6] tools/dm_restrict: Ask QEMU to chroot
Quoted 10 lines
-----Original Message-----
From: Xen-devel [mailto:[email protected]] On Behalf Of George Dunlap
Sent: 05 November 2018 18:07
To: [email protected]
Cc: Anthony Perard <[email protected]>; Ian Jackson <[email protected]>; Wei Liu <[email protected]>; George Dunlap <[email protected]>
Subject: [Xen-devel] [PATCH v4 3/6]tools/dm_restrict: Ask QEMU to chroot
Quoted 3 lines
When
dm_restrictis enabled, ask QEMU to chroot into an empty directory.* Create
/var/run/qemu/root-domid(deleting the old one if it's there)
This does not appear to match the code: the path should be /var/run/qemu-root-<domid> AFAICT
Quoted 65 lines
* Pass the -chroot option to QEMU
Rather than running
rm -rfon the directory before creating it (since there is no library function to do this), simply rmdir the directory, relying on the fact that the previous QEMU instance, if properly restricted, shouldn't have been able to write anything anyway.Suggested-by Ross Lagerwall <[email protected]>Signed-off-by George Dunlap <[email protected]>Acked-by Ian Jackson <[email protected]>---
Changes since v2:
- Style fixes
- Testing moved to a different patchCC Ian Jackson <[email protected]>CC Wei Liu <[email protected]>CC Anthony Perard <[email protected]>--- docs/designs/qemu-deprivilege.md | 12 +++++----- tools/libxl/libxl_dm.c | 41 +++++++++++++++++++++++++++++++- 2 files changed, 46 insertions(+), 7 deletions(-)docs/designs/qemu- renamed from docs/designs/qemu-deprivilege.md +0
-0deprivilege.md
index787ae1ac7c..0395bbbb40100644docs/designs/qemu-deprivilege.md +6
-6@@ -61,12 +61,6 @@ source tree.) '''Testing status''': Tested -# Restrictions / improvements still to do - -This lists potential restrictions still to do. It is meant to be -listed in order of ease of implementation, with low-hanging fruit -first. - ## Chroot '''Description''': Qemu runs in its own chroot, such that even if it@@ -84,6 +78,12 @@ Then adds the following to the qemu command-line: '''Tested''': Not tested +## Restrictions / improvements still to do + +This lists potential restrictions still to do. It is meant to be +listed in order of ease of implementation, with low-hanging fruit +first. + ## Namespaces for unused functionality (Linux only) '''Description''': QEMU doesn't use the functionality associated withtools/libxl/libxl_dm.c +0
-1@@ -1410,9 +1410,48 @@ static int libxl__build_device_model_args_new(libxl__gc *gc, } } -if (libxl_defbool_val (b_info->dm_restrict))
tools/libxl/libxl_dm.c static int · line 1423
Quoted 10 lines
tools/libxl/libxl_dm.c +7
-0+if (libxl_defbool_val (b_info->dm_restrict)) { +char *chroot_dir =GCSPRINTF ("%s/qemu-root-%d" , +libxl__run_dir_path (), guest_domid); +int r; +flexarray_append (dm_args,"-xen-domid-restrict" ); +/* +* Run QEMU in a chroot at XEN_RUN_DIR/qemu-root-%d
Maybe '<domid>' in the comment rather than '%d'?
tools/libxl/libxl_dm.c static int libxl__build_device_model_args_new(libxl__gc *gc, · line 1427
Quoted 6 lines
tools/libxl/libxl_dm.c +6
-0+* +* There is no library function to do the equivalent of `rm +* -rf`. However deprivileged QEMU in theory shouldn't be +* able to write any files, as the chroot would be owned by +* root, but it would be running as an unprivileged process. +* So in theory, old chroots should always be empty.
How does logging work if QEMU can't write to the chroot? I assume we are relying on stderr? Does using syslog still work?
Paul
Quoted 30 lines
tools/libxl/libxl_dm.c +27
-0+* +* rmdir the directory before attempting to create +* it; if it returns anything other than ENOENT, fail domain +* creation. +*/ + r =rmdir (chroot_dir); +if (r !=0 && errno !=ENOENT ) { +LOGED (ERROR , guest_domid, +"failed to remove existing chroot dir %s" , chroot_dir); +return ERROR_FAIL ; + } + +for (;;) { + r =mkdir (chroot_dir,0000 ); +if (!r) +break ; +if (errno ==EINTR )continue ; +LOGED (ERROR , guest_domid, +"failed to create chroot dir %s" , chroot_dir); +return ERROR_FAIL ; + } + +/* Add "-chroot [dir]" to command-line */ +flexarray_append (dm_args,"-chroot" ); +flexarray_append (dm_args, chroot_dir); + } +if (state->saved_state) {/* This file descriptor is meant to be used by QEMU */ *dm_state_fd = open(state->saved_state, O_RDONLY);
Quoted 8 lines
--
2.19.1_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [PATCH v4 3/6] tools/dm_restrict: Ask QEMU to chroot
Quoted 10 lines · Paul Durrant
Quoted 10 lines
-----Original Message-----
From: Xen-devel [mailto:[email protected]] On Behalf Of George Dunlap
Sent: 05 November 2018 18:07
To: [email protected]
Cc: Anthony Perard <[email protected]>; Ian Jackson <[email protected]>; Wei Liu <[email protected]>; George Dunlap <[email protected]>
Subject: [Xen-devel] [PATCH v4 3/6]tools/dm_restrict: Ask QEMU to chroot
Quoted 5 lines
Quoted 3 lines
When
dm_restrictis enabled, ask QEMU to chroot into an empty directory.* Create
/var/run/qemu/root-domid(deleting the old one if it's there)This does not appear to match the code: the path should be /var/run/qemu-root-<domid> AFAICT
Indeed, I forgot to update this. I can fix this up on check-in.
Quoted 18 lines
Quoted 18 lines
tools/libxl/libxl_dm.c +5
-1@@ -1410,9 +1410,48 @@ static int libxl__build_device_model_args_new(libxl__gc *gc, } } -if (libxl_defbool_val (b_info->dm_restrict)) +if (libxl_defbool_val (b_info->dm_restrict)) { +char *chroot_dir =GCSPRINTF ("%s/qemu-root-%d" , +libxl__run_dir_path (), guest_domid); +int r; +flexarray_append (dm_args,"-xen-domid-restrict" );
Quoted 4 lines
tools/libxl/libxl_dm.c static int libxl__build_device_model_args_new(libxl__gc *gc, · line 1421
Quoted 2 lines
tools/libxl/libxl_dm.c +2
-0+/* +* Run QEMU in a chroot at XEN_RUN_DIR/qemu-root-%d Maybe '<domid>' in the comment rather than '%d'?
Maybe. :-)
Quoted 8 lines
tools/libxl/libxl_dm.c static int libxl__build_device_model_args_new(libxl__gc *gc, · line 1427
Quoted 6 lines
tools/libxl/libxl_dm.c +6
-0+* +* There is no library function to do the equivalent of `rm +* -rf`. However deprivileged QEMU in theory shouldn't be +* able to write any files, as the chroot would be owned by +* root, but it would be running as an unprivileged process. +* So in theory, old chroots should always be empty. How does logging work if QEMU can't write to the chroot? I assume we are relying on stderr? Does using syslog still work?
Everything QEMU needs access to (including vnc sockets, qmp sockets, &c) must either be opened before the chroot happens, or passed to QEMU as an fd via qmp. In the case of logging, this happens through stderr; but if you search for 'chroot' in the design document you'll get a couple of examples of different issues that need to be addressed (including inserting a cd-rom and dealing with migration).
-George
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [PATCH v4 3/6] tools/dm_restrict: Ask QEMU to chroot
Quoted 65 lines
-----Original Message-----
From: George Dunlap [mailto:[email protected]]
Sent: 06 November 2018 10:28
To: Paul Durrant <[email protected]>; [email protected]
Cc: Anthony Perard <[email protected]>; Ian Jackson <[email protected]>; Wei Liu <[email protected]>
Subject: Re: [Xen-devel] [PATCH v4 3/6]tools/dm_restrict: Ask QEMU to chrootQuoted 2 lines · Paul Durrant
Quoted 2 lines
-----Original Message-----
From: Xen-devel [mailto:[email protected]] OnBehalf
Quoted 7 lines
Quoted 7 lines
Of George Dunlap
Sent: 05 November 2018 18:07
To: [email protected]
Cc: Anthony Perard <[email protected]>; Ian Jackson <[email protected]>; Wei Liu <[email protected]>; George Dunlap <[email protected]>
Subject: [Xen-devel] [PATCH v4 3/6]tools/dm_restrict: Ask QEMU tochroot
Quoted 2 lines
Quoted 2 lines
When
dm_restrictis enabled, ask QEMU to chroot into an emptydirectory.
Quoted 4 lines
Quoted 2 lines
* Create
/var/run/qemu/root-domid(deleting the old one if it's there)This does not appear to match the code: the path should be
/var/run/qemu-root-<domid> AFAICT
Indeed, I forgot to update this. I can fix this up on check-in.
Quoted 18 lines
Quoted 18 lines
tools/libxl/libxl_dm.c +5
-1@@ -1410,9 +1410,48 @@ static int libxl__build_device_model_args_new(libxl__gc *gc, } } -if (libxl_defbool_val (b_info->dm_restrict)) +if (libxl_defbool_val (b_info->dm_restrict)) { +char *chroot_dir =GCSPRINTF ("%s/qemu-root-%d" , +libxl__run_dir_path (), guest_domid); +int r; +flexarray_append (dm_args,"-xen-domid-restrict" );Quoted 4 lines
tools/libxl/libxl_dm.c static int libxl__build_device_model_args_new(libxl__gc *gc, · line 1421
Quoted 2 lines
tools/libxl/libxl_dm.c +2
-0+/* +* Run QEMU in a chroot at XEN_RUN_DIR/qemu-root-%d Maybe '<domid>' in the comment rather than '%d'?
Maybe. :-)
Quoted 8 lines
tools/libxl/libxl_dm.c static int libxl__build_device_model_args_new(libxl__gc *gc, · line 1427
Quoted 6 lines
tools/libxl/libxl_dm.c +6
-0+* +* There is no library function to do the equivalent of `rm +* -rf`. However deprivileged QEMU in theory shouldn't be +* able to write any files, as the chroot would be owned by +* root, but it would be running as an unprivileged process. +* So in theory, old chroots should always be empty. How does logging work if QEMU can't write to the chroot? I assume we are
Quoted 8 lines
relying on stderr? Does using syslog still work?
Everything QEMU needs access to (including vnc sockets, qmp sockets, &c) must either be opened before the chroot happens, or passed to QEMU as an fd via qmp. In the case of logging, this happens through stderr; but if you search for 'chroot' in the design document you'll get a couple of examples of different issues that need to be addressed (including inserting a cd-rom and dealing with migration).
Ok. The trace backend is set at build time in tools/Makefile:
if $$source/scripts/tracetool.py --check-backend --backend log ; then \
enable_trace_backend='--enable-trace-backend=log';
elif $$source/scripts/tracetool.py --check-backend --backend stderr ; then \
enable_trace_backend='--enable-trace-backend=stderr'; \
else \
enable_trace_backend='' ; \
fi ; \and log appears to favoured. Is this still going to work with the chroot? We rely on the tracing in xen_platform to get log messages out of PV drivers.
Paul
Quoted 2 lines
-George
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [PATCH v4 3/6] tools/dm_restrict: Ask QEMU to chroot
Quoted 10 lines · Paul Durrant
Ok. The trace backend is set at build time in
tools/Makefile:if $$source/scripts/tracetool.py --check-backend --backend log ; then \ enable_trace_backend='--enable-trace-backend=log'; elif $$source/scripts/tracetool.py --check-backend --backend stderr ; then \ enable_trace_backend='--enable-trace-backend=stderr'; \ else \ enable_trace_backend='' ; \ fi ; \
Quoted 3 lines
and log appears to favoured. Is this still going to work with the chroot? We rely on the tracing in
xen_platformto get log messages out of PV drivers.
log vs stderr are just different names for the same thing. "stderr" doesn't exist anymore in recent version of QEMU and as been renamed "log".
Signature · 6 lines
Anthony PERARD _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [PATCH v4 3/6] tools/dm_restrict: Ask QEMU to chroot
Quoted 23 lines
-----Original Message-----
From: Anthony PERARD [mailto:[email protected]]
Sent: 06 November 2018 11:11
To: Paul Durrant <[email protected]>
Cc: George Dunlap <[email protected]>; xen- [email protected]; Ian Jackson <[email protected]>; Wei Liu <[email protected]>
Subject: Re: [Xen-devel] [PATCH v4 3/6]tools/dm_restrict: Ask QEMU to chrootQuoted 3 lines · Paul Durrant
Ok. The trace backend is set at build time in
tools/Makefile:if $$source/scripts/tracetool.py --check-backend --backend log ;then \
Quoted 2 lines
enable_trace_backend='--enable-trace-backend=log'; elif $$source/scripts/tracetool.py --check-backend --backendstderr ; then \
Quoted 5 lines
enable_trace_backend='--enable-trace-backend=stderr'; \ else \ enable_trace_backend='' ; \ fi ; \
Quoted 7 lines
Quoted 3 lines
and log appears to favoured. Is this still going to work with the chroot? We rely on the tracing in
xen_platformto get log messages out of PV drivers.log vs stderr are just different names for the same thing. "stderr" doesn't exist anymore in recent version of QEMU and as been renamed "log".
Oh yes, I'd forgotten... That's fine then :-)
Paul
Quoted 3 lines
--
Anthony PERARD
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
PATCH v4 4/6 tools/dm_restrict: Unshare mount and IPC namespaces on Linux
In reply to George DunlapQEMU running under Xen doesn't need mount or IPC functionality. Create and enter separate namespaces for each of these before executing QEMU, so that in the event that other restrictions fail, the process won't be able to even name system mount points or exsting non-file-based IPC descriptors to attempt to attack them.
Unsharing is something a process can only do to itself (it would seem); so add an os-specific "dm_preexec_restrict()" hook just before we exec() the device model.
Also add checks to depriv-process-checker.sh to verify that dm is running in a new namespace (or at least, a different one than the caller).
---
Changes since v3:
- Fix some more style issues
Changes since v2:
- Return an error rather than calling exit()
- Use LOGE() and print to the current stderr fd, rather than printing to the new stderr fd via write()
- Use r for external return values rather than rc.
--- docs/designs/qemu-deprivilege.md | 12 ++++++------ tools/libxl/libxl_dm.c | 5 +++++ tools/libxl/libxl_freebsd.c | 5 +++++ tools/libxl/libxl_internal.h | 5 +++++ tools/libxl/libxl_linux.c | 14 ++++++++++++++ tools/libxl/libxl_netbsd.c | 5 +++++ 6 files changed, 40 insertions(+), 6 deletions(-)
docs/designs/qemu-deprivilege.md +6 -6
@@ -78,12 +78,6 @@ Then adds the following to the qemu command-line: '''Tested''': Not tested -## Restrictions / improvements still to do - -This lists potential restrictions still to do. It is meant to be -listed in order of ease of implementation, with low-hanging fruit -first. - ## Namespaces for unused functionality (Linux only) '''Description''': QEMU doesn't use the functionality associated with
@@ -111,6 +105,12 @@ call: [qemu-namespaces]: https://lists.gnu.org/archive/html/qemu-devel/2017-10/msg04723.html +# Restrictions / improvements still to do + +This lists potential restrictions still to do. It is meant to be +listed in order of ease of implementation, with low-hanging fruit +first. + ### Basic RLIMITs '''Description''': A number of limits on the resources that a given
tools/libxl/libxl_dm.c +5 -0
@@ -2393,6 +2393,11 @@ retry_transaction: goto out_close;if (!rc) {/* inner child */ setsid (); +if (libxl_defbool_val (b_info->dm_restrict)) { + rc =libxl__local_dm_preexec_restrict (gc); +if (rc) +_exit (-1 ); + }libxl__exec (gc, null, logfile_w, logfile_w, dm, args, envs); }
tools/libxl/libxl_freebsd.c +5 -0
@@ -245,3 +245,8 @@ int libxl__pci_topology_init(libxl__gc *gc, {return ERROR_NI ; } + +int libxl__local_dm_preexec_restrict (libxl__gc *gc) +{ +return 0 ; +}
tools/libxl/libxl_internal.h +5 -0
@@ -3774,6 +3774,11 @@ struct libxl__dm_spawn_state { _hidden void libxl__spawn_local_dm(libxl__egc *egc, libxl__dm_spawn_state*); +/* +* Called after forking but before executing the local devicemodel. +*/ +_hidden int libxl__local_dm_preexec_restrict(libxl__gc *gc); +/* Stubdom device models. */ typedef struct {
tools/libxl/libxl_linux.c +14 -0
@@ -307,6 +307,20 @@ int libxl__pci_topology_init(libxl__gc *gc,return err; } +int libxl__local_dm_preexec_restrict (libxl__gc *gc) +{ +int r; + +/* Unshare mount and IPC namespaces. These are unused by QEMU. */ + r =unshare (CLONE_NEWNS |CLONE_NEWIPC ); +if (r) { +LOGE (ERROR ,"libxl: Mount and IPC namespace unfailed" ); +return ERROR_FAIL ; + } + +return 0 ; +} +/* * Local variables: * mode: C
tools/libxl/libxl_netbsd.c +5 -0
@@ -124,3 +124,8 @@ int libxl__pci_topology_init(libxl__gc *gc, {return ERROR_NI ; } + +void libxl__local_dm_preexec_restrict (libxl__gc *gc,int stderrfd) +{ +return ; +}
Signature · 7 lines
2.19.1 _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [PATCH v4 4/6] tools/dm_restrict: Unshare mount and IPC namespaces on Linux
Quoted 160 lines
-----Original Message-----
From: Xen-devel [mailto:[email protected]] On Behalf Of George Dunlap
Sent: 05 November 2018 18:07
To: [email protected]
Cc: Anthony Perard <[email protected]>; Ian Jackson <[email protected]>; Wei Liu <[email protected]>; George Dunlap <[email protected]>
Subject: [Xen-devel] [PATCH v4 4/6]tools/dm_restrict: Unshare mount and IPC namespaces on LinuxQEMU running under Xen doesn't need mount or IPC functionality. Create and enter separate namespaces for each of these before executing QEMU, so that in the event that other restrictions fail, the process won't be able to even name system mount points or exsting non-file-based IPC descriptors to attempt to attack them.
Unsharing is something a process can only do to itself (it would seem); so add an os-specific "
dm_preexec_restrict()" hook just before weexec()the device model.Also add checks to
depriv-process-checker.shto verify that dm is running in a new namespace (or at least, a different one than the caller).Suggested-by Ross Lagerwall <[email protected]>Signed-off-by George Dunlap <[email protected]>Acked-by Ian Jackson <[email protected]>---
Changes since v3:
- Fix some more style issuesChanges since v2:
- Return an error rather than callingexit()
- UseLOGE()and print to the current stderr fd, rather than printing to the new stderr fd viawrite()
- Use r for external return values rather than rc.CC Ian Jackson <[email protected]>CC Wei Liu <[email protected]>CC Anthony Perard <[email protected]>--- docs/designs/qemu-deprivilege.md | 12 ++++++------ tools/libxl/libxl_dm.c | 5 +++++ tools/libxl/libxl_freebsd.c | 5 +++++ tools/libxl/libxl_internal.h | 5 +++++ tools/libxl/libxl_linux.c | 14 ++++++++++++++ tools/libxl/libxl_netbsd.c | 5 +++++ 6 files changed, 40 insertions(+), 6 deletions(-)docs/designs/qemu- renamed from docs/designs/qemu-deprivilege.md +0
-0deprivilege.md
index0395bbbb40..a461ebbadd100644docs/designs/qemu-deprivilege.md +0
-6@@ -78,12 +78,6 @@ Then adds the following to the qemu command-line: '''Tested''': Not tested -## Restrictions / improvements still to do - -This lists potential restrictions still to do. It is meant to be -listed in order of ease of implementation, with low-hanging fruit -first. - ## Namespaces for unused functionality (Linux only) '''Description''': QEMU doesn't use the functionality associated with@@ -111,6 +105,12 @@ call: [qemu-namespaces]: https://lists.gnu.org/archive/html/qemu-devel/2017-10/msg04723.html
docs/designs/qemu-deprivilege.md +6
-0+# Restrictions / improvements still to do + +This lists potential restrictions still to do. It is meant to be +listed in order of ease of implementation, with low-hanging fruit +first. + ### Basic RLIMITs '''Description''': A number of limits on the resources that a giventools/libxl/libxl_dm.c +5
-0@@ -2393,6 +2393,11 @@ retry_transaction: goto out_close;if (!rc) {/* inner child */ setsid (); +if (libxl_defbool_val (b_info->dm_restrict)) { + rc =libxl__local_dm_preexec_restrict (gc); +if (rc) +_exit (-1 ); + }libxl__exec (gc, null, logfile_w, logfile_w, dm, args, envs); }tools/libxl/libxl_freebsd.c +5
-0@@ -245,3 +245,8 @@ int libxl__pci_topology_init(libxl__gc *gc, {return ERROR_NI ; } + +int libxl__local_dm_preexec_restrict (libxl__gc *gc) +{ +return 0 ; +}tools/libxl/libxl_internal.h +5
-0@@ -3774,6 +3774,11 @@ struct libxl__dm_spawn_state { _hidden void libxl__spawn_local_dm(libxl__egc *egc,libxl__dm_spawn_state *); +/* +* Called after forking but before executing the local devicemodel. +*/ +_hidden int libxl__local_dm_preexec_restrict(libxl__gc *gc); +/* Stubdom device models. */ typedef struct {
tools/libxl/libxl_linux.c +14
-0@@ -307,6 +307,20 @@ int libxl__pci_topology_init(libxl__gc *gc,return err; } +int libxl__local_dm_preexec_restrict (libxl__gc *gc) +{ +int r; + +/* Unshare mount and IPC namespaces. These are unused by QEMU. */ + r =unshare (CLONE_NEWNS |CLONE_NEWIPC ); +if (r) { +LOGE (ERROR ,"libxl: Mount and IPC namespace unfailed" ); +return ERROR_FAIL ; + } + +return 0 ; +} +/* * Local variables: * mode: C tools/libxl/libxl_netbsd.c +0
-0
tools/libxl/libxl_netbsd.c int libxl__pci_topology_init(libxl__gc *gc, · line 131
Quoted 8 lines
tools/libxl/libxl_netbsd.c +5
-0@@ -124,3 +124,8 @@ int libxl__pci_topology_init(libxl__gc *gc, {return ERROR_NI ; } + +void libxl__local_dm_preexec_restrict (libxl__gc *gc,int stderrfd) +{ +return ; +}
This is a void function whereas the caller always appears to expect an int return value, regardless of OS.
Paul
Quoted 8 lines
--
2.19.1_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [PATCH v4 4/6] tools/dm_restrict: Unshare mount and IPC namespaces on Linux
Quoted 7 lines · Paul Durrant
Quoted 7 lines
tools/libxl/libxl_netbsd.c +0
-0@@ -124,3 +124,8 @@ int libxl__pci_topology_init(libxl__gc *gc, {return ERROR_NI ;
Quoted 8 lines
tools/libxl/libxl_netbsd.c int libxl__pci_topology_init(libxl__gc *gc, · line 131
Quoted 6 lines
tools/libxl/libxl_netbsd.c +5
-0} + +void libxl__local_dm_preexec_restrict (libxl__gc *gc,int stderrfd) +{ +return ; +}This is a void function whereas the caller always appears to expect an int return value, regardless of OS.
Indeed, this would probably break the build on NetBSD. Good catch.
-George
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
PATCH v4 5/6 tools/dm_depriv: Add first cut RLIMITs
In reply to George DunlapLimit the ability of a potentially compromised QEMU to consume system resources. Key limits:
- RLIMIT_FSIZE (file size): 256KiB
- RLIMIT_NPROC (after uid changes to a unique uid)
Probably unnecessary limits but why not:
- RLIMIT_CORE: 0
- RLIMIT_MSGQUEUE: 0
- RLIMIT_LOCKS: 0
- RLIMIT_MEMLOCK: 0
NB that we do not yet set RLIMIT_AS (total virtual memory) or RLIMIT_NOFILES (number of open files), since these require more care and/or more coordination with QEMU to implement.
---
Changes since v3:
- Align RLIMIT_ENTRY list for easier reading
- Fix wrong format string specifier
- Get rid of some trailing whitespace
Changes since v2:
- Use a macro to define rlimit entries
- Use RLIMIT_NLIMITS as an end-of-list marker, rather than -1
- Various style clean-ups
--- docs/designs/qemu-deprivilege.md | 12 ++++----- tools/libxl/libxl_linux.c | 42 ++++++++++++++++++++++++++++++-- 2 files changed, 46 insertions(+), 8 deletions(-)
docs/designs/qemu-deprivilege.md +6 -6
@@ -105,12 +105,6 @@ call: [qemu-namespaces]: https://lists.gnu.org/archive/html/qemu-devel/2017-10/msg04723.html -# Restrictions / improvements still to do - -This lists potential restrictions still to do. It is meant to be -listed in order of ease of implementation, with low-hanging fruit -first. - ### Basic RLIMITs '''Description''': A number of limits on the resources that a given
@@ -137,6 +131,12 @@ are specified; this does not apply to QEMU running as a Xen DM. '''Tested''': Not tested +# Restrictions / improvements still to do + +This lists potential restrictions still to do. It is meant to be +listed in order of ease of implementation, with low-hanging fruit +first. + ### Further RLIMITs RLIMIT_AS limits the total amount of memory; but this includes the
tools/libxl/libxl_linux.c +40 -2
66 lines not on this page · raw message
Signature · 7 lines
2.19.1 _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [PATCH v4 5/6] tools/dm_depriv: Add first cut RLIMITs
Quoted 84 lines
-----Original Message-----
From: Xen-devel [mailto:[email protected]] On Behalf Of George Dunlap
Sent: 05 November 2018 18:07
To: [email protected]
Cc: Anthony Perard <[email protected]>; Ian Jackson <[email protected]>; Wei Liu <[email protected]>; George Dunlap <[email protected]>
Subject: [Xen-devel] [PATCH v4 5/6]tools/dm_depriv: Add first cut RLIMITsLimit the ability of a potentially compromised QEMU to consume system resources. Key limits:
-RLIMIT_FSIZE(file size): 256KiB
-RLIMIT_NPROC(after uid changes to a unique uid)Probably unnecessary limits but why not:
-RLIMIT_CORE: 0
-RLIMIT_MSGQUEUE: 0
-RLIMIT_LOCKS: 0
-RLIMIT_MEMLOCK: 0NB that we do not yet set
RLIMIT_AS(total virtual memory) orRLIMIT_NOFILES(number of open files), since these require more care and/or more coordination with QEMU to implement.Suggested-by Ross Lagerwall <[email protected]>Signed-off-by George Dunlap <[email protected]>---
Changes since v3:
- AlignRLIMIT_ENTRYlist for easier reading
- Fix wrong format string specifier
- Get rid of some trailing whitespaceChanges since v2:
- Use a macro to define rlimit entries
- UseRLIMIT_NLIMITSas an end-of-list marker, rather than -1
- Various style clean-upsCC Ian Jackson <[email protected]>CC Wei Liu <[email protected]>CC Anthony Perard <[email protected]>--- docs/designs/qemu-deprivilege.md | 12 ++++----- tools/libxl/libxl_linux.c | 42 ++++++++++++++++++++++++++++++-- 2 files changed, 46 insertions(+), 8 deletions(-)docs/designs/qemu- renamed from docs/designs/qemu-deprivilege.md +0
-0deprivilege.md
indexa461ebbadd..e984064da6100644docs/designs/qemu-deprivilege.md +0
-0@@ -105,12 +105,6 @@ call: [qemu-namespaces]: https://lists.gnu.org/archive/html/qemu-devel/2017-10/msg04723.html
docs/designs/qemu-deprivilege.md +0
-6-# Restrictions / improvements still to do - -This lists potential restrictions still to do. It is meant to be -listed in order of ease of implementation, with low-hanging fruit -first. - ### Basic RLIMITs '''Description''': A number of limits on the resources that a given@@ -137,6 +131,12 @@ are specified; this does not apply to QEMU running asa Xen DM.
'''Tested''': Not tested
docs/designs/qemu-deprivilege.md +6
-0+# Restrictions / improvements still to do + +This lists potential restrictions still to do. It is meant to be +listed in order of ease of implementation, with low-hanging fruit +first. + ### Further RLIMITs RLIMIT_AS limits the total amount of memory; but this includes thetools/libxl/libxl_linux.c +0
-0
tools/libxl/libxl_linux.c line 15
Quoted 5 lines
tools/libxl/libxl_linux.c +1
-1@@ -12,11 +12,12 @@* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU Lesser General Public License for more details. */ - +
Stray whitespace change?
tools/libxl/libxl_linux.c line 20
Quoted 6 lines
tools/libxl/libxl_linux.c +2
-1#include "libxl_osdeps.h" /* must come before any other headers */ #include "libxl_internal.h" - +#include <sys/resource.h> +
Personally I tend to put local includes after ones from the include path. Is there a reason it needs to come afterwards?
Quoted 17 lines
tools/libxl/libxl_linux.c +10
-0int libxl__try_phy_backend (mode_t st_mode) {if (S_ISBLK (st_mode) ||S_ISREG (st_mode)) {@@ -307,9 +308,31 @@ int libxl__pci_topology_init(libxl__gc *gc,return err; } +static struct { +int resource; +rlim_t limit; +} rlimits[] = { +#define RLIMIT_ENTRY(r, l) \ + { .resource = r, .limit = l } +/* Big enough for log files, not big enough for a DoS */ +RLIMIT_ENTRY (RLIMIT_FSIZE ,256 *1024 ), + +/* Shouldn't need any of these */
tools/libxl/libxl_linux.c int libxl__pci_topology_init(libxl__gc *gc, · line 330
Quoted 10 lines
tools/libxl/libxl_linux.c +10
-0+RLIMIT_ENTRY (RLIMIT_NPROC ,0 ), +RLIMIT_ENTRY (RLIMIT_CORE ,0 ), +RLIMIT_ENTRY (RLIMIT_MSGQUEUE ,0 ), +RLIMIT_ENTRY (RLIMIT_LOCKS ,0 ), +RLIMIT_ENTRY (RLIMIT_MEMLOCK ,0 ), + +/* End-of-list marker */ + RLIMIT_ENTRY(RLIMIT_NLIMITS, 0), +}; +#undef RLIMIT_ENTRY
<pedantic> The undef should come before the brace to get the scoping correct. </pedantic>
Quoted 13 lines
tools/libxl/libxl_linux.c +3
-0+int libxl__local_dm_preexec_restrict (libxl__gc *gc) {int r; +unsigned i;/* Unshare mount and IPC namespaces. These are unused by QEMU. */ r =unshare (CLONE_NEWNS |CLONE_NEWIPC );@@ -318,6 +341,21 @@ int libxl__local_dm_preexec_restrict(libxl__gc *gc)return ERROR_FAIL ; } +/* Set various "easy" rlimits */
tools/libxl/libxl_linux.c int libxl__local_dm_preexec_restrict(libxl__gc *gc) · line 354
Quoted 10 lines
tools/libxl/libxl_linux.c +10
-0+for (i =0 ; rlimits[i].resource !=RLIMIT_NLIMITS ; i++) { +struct rlimit rlim; + + rlim.rlim_cur = rlim.rlim_max = rlimits[i].limit; + + r =setrlimit (rlimits[i].resource, &rlim); +if (r <0 ) { +LOGE (ERROR ,"Setting rlimit %d to %llu failed\n" , + rlimits[i].resource, + (unsigned long long )rlimits[i].limit);
Indentation of the continuation lines looks odd (although libxl's coding style is a mystery to me so they may be correct).
Paul
Quoted 11 lines
tools/libxl/libxl_linux.c +4
-1+return ERROR_FAIL ; + } + } +return 0 ; } --2.19.1
Quoted 4 lines
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [PATCH v4 5/6] tools/dm_depriv: Add first cut RLIMITs
Quoted 78 lines · Paul Durrant
Quoted 78 lines
-----Original Message-----
From: Xen-devel [mailto:[email protected]] On Behalf Of George Dunlap
Sent: 05 November 2018 18:07
To: [email protected]
Cc: Anthony Perard <[email protected]>; Ian Jackson <[email protected]>; Wei Liu <[email protected]>; George Dunlap <[email protected]>
Subject: [Xen-devel] [PATCH v4 5/6]tools/dm_depriv: Add first cut RLIMITsLimit the ability of a potentially compromised QEMU to consume system resources. Key limits:
-RLIMIT_FSIZE(file size): 256KiB
-RLIMIT_NPROC(after uid changes to a unique uid)Probably unnecessary limits but why not:
-RLIMIT_CORE: 0
-RLIMIT_MSGQUEUE: 0
-RLIMIT_LOCKS: 0
-RLIMIT_MEMLOCK: 0NB that we do not yet set
RLIMIT_AS(total virtual memory) orRLIMIT_NOFILES(number of open files), since these require more care and/or more coordination with QEMU to implement.Suggested-by Ross Lagerwall <[email protected]>Signed-off-by George Dunlap <[email protected]>---
Changes since v3:
- AlignRLIMIT_ENTRYlist for easier reading
- Fix wrong format string specifier
- Get rid of some trailing whitespaceChanges since v2:
- Use a macro to define rlimit entries
- UseRLIMIT_NLIMITSas an end-of-list marker, rather than -1
- Various style clean-upsCC Ian Jackson <[email protected]>CC Wei Liu <[email protected]>CC Anthony Perard <[email protected]>--- docs/designs/qemu-deprivilege.md | 12 ++++----- tools/libxl/libxl_linux.c | 42 ++++++++++++++++++++++++++++++-- 2 files changed, 46 insertions(+), 8 deletions(-)docs/designs/qemu- renamed from docs/designs/qemu-deprivilege.md +0
-0deprivilege.md
indexa461ebbadd..e984064da6100644docs/designs/qemu-deprivilege.md +0
-0@@ -105,12 +105,6 @@ call: [qemu-namespaces]: https://lists.gnu.org/archive/html/qemu-devel/2017-10/msg04723.html
docs/designs/qemu-deprivilege.md +0
-6-# Restrictions / improvements still to do - -This lists potential restrictions still to do. It is meant to be -listed in order of ease of implementation, with low-hanging fruit -first. - ### Basic RLIMITs '''Description''': A number of limits on the resources that a given@@ -137,6 +131,12 @@ are specified; this does not apply to QEMU running asa Xen DM.
'''Tested''': Not tested
docs/designs/qemu-deprivilege.md +6
-0+# Restrictions / improvements still to do + +This lists potential restrictions still to do. It is meant to be +listed in order of ease of implementation, with low-hanging fruit +first. + ### Further RLIMITs
Quoted 13 lines
tools/libxl/libxl_linux.c line 15
Quoted 11 lines
RLIMIT_ASlimits the total amount of memory; but this includes thetools/libxl/libxl_linux.c +1
-1@@ -12,11 +12,12 @@* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU Lesser General Public License for more details. */ - +Stray whitespace change?
Got rid of trailing witespace; I mentioned it under "Changes since v3".
Quoted 8 lines
tools/libxl/libxl_linux.c line 20
Quoted 6 lines
tools/libxl/libxl_linux.c +2
-1#include "libxl_osdeps.h" /* must come before any other headers */ #include "libxl_internal.h" - +#include <sys/resource.h> +Personally I tend to put local includes after ones from the include path. Is there a reason it needs to come afterwards?
No reason; just habit to add things to the end. I'll move it earlier unless Ian objects.
Quoted 14 lines
Quoted 14 lines
tools/libxl/libxl_linux.c +14
-0+static struct { +int resource; +rlim_t limit; +} rlimits[] = { +#define RLIMIT_ENTRY(r, l) \ + { .resource = r, .limit = l } +/* Big enough for log files, not big enough for a DoS */ +RLIMIT_ENTRY (RLIMIT_FSIZE ,256 *1024 ), + +/* Shouldn't need any of these */ +RLIMIT_ENTRY (RLIMIT_NPROC ,0 ), +RLIMIT_ENTRY (RLIMIT_CORE ,0 ), +RLIMIT_ENTRY (RLIMIT_MSGQUEUE ,0 ), +RLIMIT_ENTRY (RLIMIT_LOCKS ,0 ),
Quoted 8 lines
tools/libxl/libxl_linux.c int libxl__pci_topology_init(libxl__gc *gc, · line 330
Quoted 6 lines
tools/libxl/libxl_linux.c +6
-0+RLIMIT_ENTRY (RLIMIT_MEMLOCK ,0 ), + +/* End-of-list marker */ + RLIMIT_ENTRY(RLIMIT_NLIMITS, 0), +}; +#undef RLIMIT_ENTRY<pedantic> The undef should come before the brace to get the scoping correct. </pedantic>
Sure.
Quoted 12 lines
Quoted 12 lines
tools/libxl/libxl_linux.c +2
-0+int libxl__local_dm_preexec_restrict (libxl__gc *gc) {int r; +unsigned i;/* Unshare mount and IPC namespaces. These are unused by QEMU. */ r =unshare (CLONE_NEWNS |CLONE_NEWIPC );@@ -318,6 +341,21 @@ int libxl__local_dm_preexec_restrict(libxl__gc *gc)return ERROR_FAIL ; }
Quoted 13 lines
tools/libxl/libxl_linux.c int libxl__local_dm_preexec_restrict(libxl__gc *gc) · line 354
Quoted 11 lines
tools/libxl/libxl_linux.c +11
-0+/* Set various "easy" rlimits */ +for (i =0 ; rlimits[i].resource !=RLIMIT_NLIMITS ; i++) { +struct rlimit rlim; + + rlim.rlim_cur = rlim.rlim_max = rlimits[i].limit; + + r =setrlimit (rlimits[i].resource, &rlim); +if (r <0 ) { +LOGE (ERROR ,"Setting rlimit %d to %llu failed\n" , + rlimits[i].resource, + (unsigned long long )rlimits[i].limit);Indentation of the continuation lines looks odd (although libxl's coding style is a mystery to me so they may be correct).
Don't think it says anything about this case; I find having the arguments indented beyond the format string easier to read.
-George
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [PATCH v4 5/6] tools/dm_depriv: Add first cut RLIMITs
In reply to George DunlapGeorge Dunlap writes ("[PATCH v4 5/6] tools/dm_depriv: Add first cut RLIMITs"):
Quoted 4 lines
Limit the ability of a potentially compromised QEMU to consume system resources. Key limits:
-RLIMIT_FSIZE(file size): 256KiB
-RLIMIT_NPROC(after uid changes to a unique uid)
...
Quoted 2 lines
Suggested-by Ross Lagerwall <[email protected]>Signed-off-by George Dunlap <[email protected]>
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
PATCH v4 6/6 RFC: test/depriv: Add a tool to check process-level depriv
In reply to George DunlapAdd a tool to check whether the various process-level deprivileging operations have actually taken place on the process.
The tool takes a domname or domid, and returns success or failure.
---
Changes since v3:
- Use xen-qemuuser-range-base's gid rather than hard-coding nobody
- Change FIXME about not handling other userid schemes into an NB.
Changes since v2:
- Make grep for Uid line more strict
- Fix Gid grep, make more strict
- Match strictly more than one space
- Look up the group ID for nobody rather than hard-coding it
- Move tests from other patches into one patch
- Remove suffix (in case we change the language)
- Install in the path
NB this patch is included for reference only, while I consider whether to leave this as a stand-alone script, or whether to merge osstest's fd checker functionality into it (perhaps changing the language to perl at the same time). Reviews of the general detection algorithm are welcome, but there's no need for a detailed review of the code until the script is in its final form.
--- tools/tests/depriv/Makefile | 2 +- tools/tests/depriv/depriv-process-checker | 148 ++++++++++++++++++++++ 2 files changed, 149 insertions(+), 1 deletion(-) create mode 100755 tools/tests/depriv/depriv-process-checker
tools/tests/depriv/Makefile +1 -1
@@ -23,7 +23,7 @@ LDLIBS += $(LDLIBS_libxendevicemodel) LDLIBS += $(LDLIBS_libxentoolcore) LDLIBS += $(LDLIBS_libxentoollog) -INSTALL_PRIVBIN-y += depriv-fd-checker +INSTALL_PRIVBIN-y += depriv-fd-checker depriv-process-checker INSTALL_PRIVBIN := $(INSTALL_PRIVBIN-y) TARGETS += $(INSTALL_PRIVBIN)
tools/tests/depriv/depriv-process-checker new file +148 -0
148 lines not on this page · raw message
Signature · 7 lines
2.19.1 _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [PATCH v4 6/6] RFC: test/depriv: Add a tool to check process-level depriv
Quoted 107 lines
-----Original Message-----
From: Xen-devel [mailto:[email protected]] On Behalf Of George Dunlap
Sent: 05 November 2018 18:07
To: [email protected]
Cc: Stefano Stabellini <[email protected]>; Wei Liu <[email protected]>; George Dunlap <[email protected]>; Ross Lagerwall <[email protected]>; Anthony Perard <[email protected]>; Ian Jackson <[email protected]>
Subject: [Xen-devel] [PATCH v4 6/6] RFC: test/depriv: Add a tool to check process-level deprivAdd a tool to check whether the various process-level deprivileging operations have actually taken place on the process.
The tool takes a domname or domid, and returns success or failure.
Signed-off-by George Dunlap <[email protected]>---
Changes since v3:
- Use xen-qemuuser-range-base's gid rather than hard-codingnobody
- Change FIXME about not handling other userid schemes into an NB.Changes since v2:
- Make grep for Uid line more strict
- Fix Gid grep, make more strict
- Match strictly more than one space
- Look up the group ID fornobodyrather than hard-coding it
- Move tests from other patches into one patch
- Remove suffix (in case we change the language)
- Install in the pathNB this patch is included for reference only, while I consider whether to leave this as a stand-alone script, or whether to merge osstest's fd checker functionality into it (perhaps changing the language to perl at the same time). Reviews of the general detection algorithm are welcome, but there's no need for a detailed review of the code until the script is in its final form.
CC Ian Jackson <[email protected]>CC Wei Liu <[email protected]>CC Stefano Stabellini <[email protected]>CC Anthony Perard <[email protected]>CC Ross Lagerwall <[email protected]>--- tools/tests/depriv/Makefile | 2 +- tools/tests/depriv/depriv-process-checker | 148 ++++++++++++++++++++++ 2 files changed, 149 insertions(+), 1 deletion(-) create mode 100755 tools/tests/depriv/depriv-process-checkertools/tests/depriv/Makefile +1
-1@@ -23,7 +23,7 @@ LDLIBS += $(LDLIBS_libxendevicemodel) LDLIBS += $(LDLIBS_libxentoolcore) LDLIBS += $(LDLIBS_libxentoollog) -INSTALL_PRIVBIN-y += depriv-fd-checker +INSTALL_PRIVBIN-y += depriv-fd-checker depriv-process-checker INSTALL_PRIVBIN := $(INSTALL_PRIVBIN-y) TARGETS += $(INSTALL_PRIVBIN)tools/tests/depriv/depriv-process-checker new file +11
-0@@ -0,0 +1,148 @@ +#!/bin/bash + +domain="$1" + +if [[ "$domain" =~ ^[0-9]+$ ]] ; then + domid="$domain" +else + domid=$(xl domid "$domain") +fi + +dmpid=$(xenstore-read /local/domain/$domid/image/device-model-pid2>/dev/null)
tools/tests/depriv/depriv-process-checker new file +24
-0+if [[ -z "$dmpid" ]] ; then + echo "xenstore-read failed" + exit 1 +fi + +failed="false" + +# TEST: Process / group id +# +# Read /proc/<qpid>/status, checking Uid and Gid lines +# +# Uid should be xen-qemuuser-range-base+$domid +# Gid should be gid for xen-qemuuser-range-base +# +# NB this doesn't handle other configurations (e.g., +# xen-qemuuser-shared). +echo -n "Process UID: " +tgt_uid=$(id -u xen-qemuuser-range-base) +tgt_uid=$(( $tgt_uid + $domid )) + +# Example input: +# Uid: 1193 1193 1193 1193 +input=$(grep ^Uid: /proc/$dmpid/status) +if [[ "$input" =~ ^Uid:[[:space:]]+([0-9]+)[[:space:]]+([0-9]+)[[:space:]]+([0-9]+)[[:space:]]+([0-9]+)$ ]] ; then
tools/tests/depriv/depriv-process-checker line 40
Quoted 5 lines
tools/tests/depriv/depriv-process-checker new file +5
-0+ result="PASSED" + for i in {1..4}; do + if [[ "${BASH_REMATCH[$i]}" != "$tgt_uid" ]] ; then + result="FAILED" + failed="true"
This patter occurs many times. Could you not just have a boolean 'failed' value and a supporting 'reason' (which might be empty) then you can just test 'reason' at the end to decide what to echo?
Quoted 29 lines
tools/tests/depriv/depriv-process-checker new file +15
-0+ break + fi + done +else + result="FAILED" + failed="true" +fi +echo $result + +# Example input: +# Gid: 10020 10020 10020 10020 +echo -n "Process GID: " +tgt_gid=$(id -g xen-qemuuser-range-base) +input=$(grep ^Gid: /proc/$dmpid/status) +if [[ "$input" =~ ^Gid:[[:space:]]+([0-9]+)[[:space:]]+([0-9]+)[[:space:]]+([0-9]+)[[:space:]]+([0-9]+)$ ]] ; then
tools/tests/depriv/depriv-process-checker new file +13
-0+ result="PASSED" + for i in {1..4}; do + if [[ "${BASH_REMATCH[$i]}" != "$tgt_gid" ]] ; then + result="FAILED" + failed="true" + break + fi + done +else + result="FAILED" + failed="true" +fi +echo $result
tools/tests/depriv/depriv-process-checker line 78
Quoted 10 lines
tools/tests/depriv/depriv-process-checker new file +10
-0+ +# TEST: chroot +# +# Read /proc/<dmpid>/root to see if it's correct. +echo -n "Chroot: " +if [[ -n "$XEN_RUN_DIR" ]] ; then + tgt_chroot=$XEN_RUN_DIR/qemu-root-$domid + root=$(readlink /proc/$dmpid/root) + if [[ "$root" != "$tgt_chroot" ]] ; then + echo "FAILED"
You are using hard tabs here but 4 spaces in the blocks above.
Quoted 58 lines
tools/tests/depriv/depriv-process-checker new file +45
-0+ failed="true" + else + echo "PASSED" + fi +else + echo "FAILED (XEN_RUN_DIR undefined)" + failed="true" +fi + +# TEST: Namespace unsharing +# +# Read /proc/<dmpid>/ns/<namespace> and make sure it's not equal to +# the current processes' value +for nsname in ipc mnt; do + echo -n "Unshare namespace $nsname: " + dmns=$(readlink /proc/$dmpid/ns/$nsname) + myns=$(readlink /proc/self/ns/$nsname) + + if [[ "$dmns" == "$myns" ]] ; then + echo "FAILED" + failed="true" + else + echo "PASSED" + fi +done + +# TEST: RLIMITs +# +# Read /proc/<dmpid>/limits +function check_rlimit() { + limit_name=$1 + limit_string=$2 + tgt=$3 + + echo -n "rlimit $limit_name: " + input=$(grep "^$limit_string" /proc/$dmpid/limits) + + if [[ -z "$input" ]] ; then + echo "Couldn't find limit $limit" + echo FAILED + failed="true" + return + fi + + if [[ "$input" =~^$limit_string[[:space:]]*([^[:space:]]+)[[:space:]]*([^[:space:]]+)[[:spa ce:]]*[^[:space:]]+ ]] ; then
tools/tests/depriv/depriv-process-checker new file +11
-0+ if [[ "${BASH_REMATCH[1]}" != $tgt || + "${BASH_REMATCH[2]}" != $tgt ]] ; then + echo "FAILED" + failed="true" + else + echo "PASSED" + fi + else + echo "Couldn't parse /proc/<dmpid>/limits" + echo "FAILED" + failed="true"
tools/tests/depriv/depriv-process-checker line 144
Quoted 10 lines
tools/tests/depriv/depriv-process-checker new file +10
-0+ fi +} +check_rlimit FSIZE "Max file size" "262144" +check_rlimit NPROC "Max processes" 0 +check_rlimit CORE "Max core file size" "0" +check_rlimit MSGQUEUE "Max msgqueue size" 0 +check_rlimit LOCKS "Max file locks" 0 +check_rlimit MEMLOCK "Max locked memory" 0 + +if $failed ; then
If you made 'failed' an integer value then you could just 'exit $failed'
Paul
Quoted 8 lines
tools/tests/depriv/depriv-process-checker new file +4
-1+ exit 1 +else + exit 0 +fi --2.19.1
Quoted 4 lines
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [PATCH v4 6/6] RFC: test/depriv: Add a tool to check process-level depriv
Quoted 4 lines · Paul Durrant
Quoted 4 lines
tools/tests/depriv/depriv-process-checker new file +4
-0+# Example input: +# Uid: 1193 1193 1193 1193 +input=$(grep ^Uid: /proc/$dmpid/status) +if [[ "$input" =~ ^Uid:[[:space:]]+([0-9]+)[[:space:]]+([0-
Quoted 8 lines
tools/tests/depriv/depriv-process-checker line 40
Quoted 6 lines
9]+)[[:space:]]+([0-9]+)[[:space:]]+([0-9]+)$ ]] ; then
tools/tests/depriv/depriv-process-checker new file +5
-0+ result="PASSED" + for i in {1..4}; do + if [[ "${BASH_REMATCH[$i]}" != "$tgt_uid" ]] ; then + result="FAILED" + failed="true"This patter occurs many times. Could you not just have a boolean 'failed' value and a supporting 'reason' (which might be empty) then you can just test 'reason' at the end to decide what to echo?
Yes, Ian also suggested that, and if I end up leaving it as a bash script I'll certainly do something like that. However:
Quoted 6 lines
Quoted 6 lines
NB this patch is included for reference only, while I consider whether to leave this as a stand-alone script, or whether to merge osstest's fd checker functionality into it (perhaps changing the language to perl at the same time). Reviews of the general detection algorithm are welcome, but there's no need for a detailed review of the code until the script is in its final form.
Sorry you missed that! If I end up leaving this as a bash script I'll take your comments about tabs/spaces and the value of "failed" into account.
-George
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Quoted 24 lines · George Dunlap
docs/qemu-deprivilege.txthad some basic instructions for usingdm_restrict, but it was incomplete, misleading, and stale.Update the docs in a number of ways.
First, separate user-facing documentation and technical description into docs/features and docs/design, respectively.
In the feature doc:
* Introduce a section mentioning minimim versions of Linux, Xen, and qemu required (TBD)
* Fix the discussion of qemu userid. Mention xen-qemuuser-range-base, and provide example shell code that actually has some hope of working (instead of failing out after creating 900 userids).
* Describe how to enable restrictions, as well as features which probably don't or definitely don't work.
In the design doc, introduce a "Technical Details" section which describes specifically what restrictions are currently done, and also what restrictions we are looking at doing in the future.
Quoted 6 lines
The idea here is that as we implement the various items for the future, we move them from "Restrictions still to do" to "Restrictions done". This can also act as a design document -- a place for public discussion of what can or should be done and how.
Also add an entry to SUPPORT.md.
Sorry, I thought I'd removed this -- the SUPPORT.md changes have been moved to another file. I can fix this up on check-in if necessary.
-George
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Quoted 147 lines
-----Original Message-----
From: Xen-devel [mailto:[email protected]] On Behalf Of George Dunlap
Sent: 05 November 2018 18:07
To: [email protected]
Cc: Stefano Stabellini <[email protected]>; Wei Liu <[email protected]>; Konrad Wilk <[email protected]>; Andrew Cooper <[email protected]>; Tim (Xen.org) <[email protected]>; George Dunlap <[email protected]>; Ross Lagerwall <[email protected]>; Julien Grall <[email protected]>; Jan Beulich <[email protected]>; Anthony Perard <[email protected]>; Ian Jackson <[email protected]>
Subject: [Xen-devel] [PATCH v4 1/6] docs/qemu-deprivilege: Revise and update with status and future plans
docs/qemu-deprivilege.txthad some basic instructions for usingdm_restrict, but it was incomplete, misleading, and stale.Update the docs in a number of ways.
First, separate user-facing documentation and technical description into docs/features and docs/design, respectively.
In the feature doc:
* Introduce a section mentioning minimim versions of Linux, Xen, and qemu required (TBD)
* Fix the discussion of qemu userid. Mention xen-qemuuser-range-base, and provide example shell code that actually has some hope of working (instead of failing out after creating 900 userids).
* Describe how to enable restrictions, as well as features which probably don't or definitely don't work.
In the design doc, introduce a "Technical Details" section which describes specifically what restrictions are currently done, and also what restrictions we are looking at doing in the future.
The idea here is that as we implement the various items for the future, we move them from "Restrictions still to do" to "Restrictions done". This can also act as a design document -- a place for public discussion of what can or should be done and how.
Also add an entry to SUPPORT.md.
Signed-off-by George Dunlap <[email protected]>---
Changes since v3:
- Fix typo (32->16)
- Use an example value not close to thenobodyuids, but still a multiple of 2^16.
- Mention that using a multiple of 2^16 may have advantages.
- Have the example create a group as well
- Reorganize two comments on the "range-base" method for clarityChanges since v2:
- Extraneous privcmd / evtchn instances aren't closed
- Expand description of how to test fd deprivileging
- Rework and clarify two namespace sections, give reference for QEMU NAK
- Add more information about migration technical challenges
- In UID section, mention possibility of container ID collisions.
- Fix name of design document.
- Add SUPPORT.md statement. Specify Linux, to make sure that FreeBSD is evaluated separately.
- Mention that-sandboxis a blacklist and whyChanges since v1:
- Break into two, and move into appropriate directories (rather than 'misc')
- Updated version requirements
- Distinguish between features which "don't yet work" and features which we never expect to work
- Update description of xen-restrict functionality
- Reorder and expand further restrictions
- Make it more clear which restrictions are available on Linux only
- Include detailed description of how to kill a process
- AddRLIMIT_NPROCas something we can do without further changes to qemu
- Document the need to check for the sandbox feature before using itThank you to Ross Lagerwall, whose description of what XenServer is doing formed much of the basis for the text here.
CC Ian Jackson <[email protected]>CC Wei Liu <[email protected]>CC Andrew Cooper <[email protected]>CC Jan Beulich <[email protected]>CC Tim Deegan <[email protected]>CC Konrad Wilk <[email protected]>CC Stefano Stabellini <[email protected]>CC Julien Grall <[email protected]>CC Anthony Perard <[email protected]>CC Ross Lagerwall <[email protected]>--- docs/designs/qemu-deprivilege.md | 322 ++++++++++++++++++++++++++ docs/features/qemu-deprivilege.pandoc | 101 ++++++++ docs/misc/qemu-deprivilege.txt | 36 --- 3 files changed, 423 insertions(+), 36 deletions(-) create mode 100644 docs/designs/qemu-deprivilege.md create mode 100644 docs/features/qemu-deprivilege.pandoc delete mode 100644 docs/misc/qemu-deprivilege.txtdocs/designs/qemu- renamed from docs/designs/qemu-deprivilege.md +0
-0deprivilege.md
new file mode 100644
index 0000000000..787ae1ac7cdocs/designs/qemu-deprivilege.md new file +28
-0@@ -0,0 +1,322 @@ +# Introduction + +The goal of deprilvileging qemu is this: Even if there is a bug (for +example in qemu) which permits a domain to gain control of the device +model, the compromised device model process is prevented from +violating the system's overall security properties. Ie, a guest +cannot "escape" from the virtualisation by using a qemu bug. + +This document lists the various technical measures which we either +have taken, or plan to take to effect this goal. Some of them are +required to be considered secure (that is, there are known attack +vectors which they close); others are "just in case" (that is, there +are no known attack vectors, but we perform the restrictions to reduce +the possibility of unknown attack vectors). + +# Restrictions done + +The following restrictions are currently implemented. + +## Having qemu switch user + +'''Description''': As mentioned above, having QEMU switch to a +non-root user, one per domain id. Not being the root user limits what +a compromised QEMU process can do to the system, and having one user +per domain id limits what a comprimised QEMU process can do to the +QEMU processes of other VMs. + +'''Implementation''': The toolstack adds the following to the qemucommand-line:
docs/designs/qemu-deprivilege.md new file +9
-0+ + -runas <uid>:<gid> + +'''How to test''': + + grep /proc/<qpid>/status [UG]id + +'''Testing Status''': Not tested +
docs/designs/qemu-deprivilege.md line 47
Quoted 10 lines
docs/designs/qemu-deprivilege.md new file +10
-0+## Xen library / file-descriptor restrictions + +'''Description''': Close and restrict Xen-related file descriptors. +Specifically: + * Close all xenstore-related file descriptors + * Make sure that all open instances of `privcmd` and `evtchn` file +descriptors have had `IOCTL_PRIVCMD_RESTRICT` and +`IOCTL_EVTCHN_RESTRICT_DOMID` ioctls called on them, respectively. + +FIXME: Double-check the correctness of the above
Presumably this should go away before commit ^
Quoted 241 lines
+
+'''Implementation''': Toolstack adds the following to the qemu command- line:docs/designs/qemu-deprivilege.md new file +27
-0+ + -xen-domid-restrict + +'''How to test''': + +Use `fishdescriptor` to pull a file descriptor from a running QEMU, +then use `depriv-fd-checker` to check that it has the desired +properties, and that hypercalls which are meant to fail do fail. (In +Debian `fishdescriptor` can be found in the binary package +`chiark-scripts`; the `depriv-fd-checker` is included in the Xen +source tree.) + +'''Testing status''': Tested + +# Restrictions / improvements still to do + +This lists potential restrictions still to do. It is meant to be +listed in order of ease of implementation, with low-hanging fruit +first. + +## Chroot + +'''Description''': Qemu runs in its own chroot, such that even if it +could call an 'open' command of some sort, there would be nothing for +it to see. + +'''Implementation''': The toolstack creates a directory in the libxl"run-dir"; e.g.
+/var/run/xen/qemu-root-<domid>docs/designs/qemu-deprivilege.md new file +25
-0+ +Then adds the following to the qemu command-line: + + -chroot /var/run/xen/qemu-root-<domid> + +'''How to test''': Check `/proc/<qpid>/root` + +'''Tested''': Not tested + +## Namespaces for unused functionality (Linux only) + +'''Description''': QEMU doesn't use the functionality associated with +mount and IPC namespaces. (IPC namespaces contol non-file-based IPC +mechanisms within the kernel; unix and network sockets are not +affected by this.) Making separate namespaces for these for QEMU +won't affect normal operation, but it does mean that even if other +restrictions fail, the process won't be able to even name system mount +points or existing non-file-based IPC descriptors to attempt to attack +them. + +'''Implementation''': + +In theory this could be done in QEMU (similar to -sandbox, -runas, +-chroot, and so on), but a patch doing this in QEMU was NAKed upstream +(see [qemu-namespaces]). They preferred that this was done as a setupstep by
docs/designs/qemu-deprivilege.md new file +10
-0+whatever executes QEMU; i.e., have the process which exec's QEMU first +call: + + unshare(CLONE_NEWNS | CLONE_NEWIPC) + +'''How to test''': Check `/proc/<qpid>/ns/[ipc,mnt]` + +'''Tested''': Not tested + +[qemu-namespaces]: https://lists.gnu.org/archive/html/qemu-devel/2017-10/msg04723.html
docs/designs/qemu-deprivilege.md new file +98
-0+ +### Basic RLIMITs + +'''Description''': A number of limits on the resources that a given +process / userid is allowed to consume. These can limit the ability +of a compromised QEMU process to DoS domain 0 by exhausting various +resources available to it. + +'''Implementation''' + +Limits that can be implemented immediately without much effort: + - RLIMIT_FSIZE` (file size) to 256KiB. + - RLIMIT_NPROC (after uid changes to a unique uid) + +Probably not necessary but why not: + - RLIMIT_CORE: 0 + - RLIMIT_MSGQUEUE: 0 + - RLIMIT_LOCKS: 0 + - RLIMIT_MEMLOCK: 0 + +Note: mlock() is used by QEMU only when both "realtime" and "mlock" +are specified; this does not apply to QEMU running as a Xen DM. + +'''How to test''': Check `/proc/<qpid>/limits` + +'''Tested''': Not tested + +### Further RLIMITs + +RLIMIT_AS limits the total amount of memory; but this includes the +virtual memory which QEMU uses as a mapcache. xen-mapcache.c already +fiddles with this; it would be straightforward to make it *set* the +rlimit to what it thinks a sensible limit is. + +Other things that would take some cleverness / changes to QEMU to +utilize due to ordering constrants: + - RLIMIT_NOFILES (after all necessary files are opened) + +### libxl UID cleanup + +'''Description''': Domain IDs are reused, and thus restricted UIDs are +reused. If a compromised QEMU can fork (due to seccomp or +RLIMIT_NPROC limits being ineffective for some reason), it may avoid +being killed when its domain dies, then wait until the domain ID is +reused again, at which point it will have control over the domain in +question (which probably belongs to someone else). + +libxl should kill all UIDs associated with a domain both when the VM +is destroyed, and before starting a VM with the same UID. + +'''Implementation''': This is unnecessarily tricky. + +The kill() system call can have three kinds of targets: + - A single pid + - A process group + - "Every process except me to which I am allowed to send a signal" (-1) + +Targeting a single pid is racy and likely to be beaten by the +following loop: + + while(1) { + if(fork()) + _exit(0); + } + +That is, by the time you've read the process list and found the +process id you want to kill, that process has exited and there is a +new process whose pid you don't know about. + +Targeting a process group will be ineffective, as unprivileged +processes are allowed to make their own process groups. + +kill(-1) can be used but must be done with care. Consider the +following code, for example: + + setuid(target_uid); + kill(-1, 9); + +This looks like it will do the trick; but by setting all of the user +ids (effective, real, and saved), it opens the 'killing' process up to +being killed by the target process: + + while(1) { + if(fork()) + _exit(0); + else + kill(-1, 9); + } + +Fortunately there is an assymetry we can take advantage of. From the +POSIX spec: + +> For a process to have permission to send a signal to a process +> designated by pid, unless the sending process has appropriate +> privileges, the real or effective user ID of the sending process shall +> match the real or saved set-user-ID of the receiving process. + +The solution is to allocate a second "reaper" uid that is only used tokill
+target processes. We set the euid of the killing process to thetarget_uid,
+but the ruid of the killing process toreaper_uid, leaving the suid of thedocs/designs/qemu-deprivilege.md new file +33
-0+killing process as 0: + + setresuid(reaper_uid, target_uid, 0); + kill(-1, 9); + +NOTE: We cannot use `setreuid(reaper_uid, target_uid)` here, as that +will set *both* euid *and* suid to `target_uid`, making the killing +process vulnerable to the target process again. + +Since this will kill all other `reaper_uid` processes as well, we must +either allocate a separate `reaper_uid` per domain, or use locking to +ensure that only one killing process is active at a time. + +## libxl: Treat QMP connection as untrusted + +'''Description''': Currently libxl talks with QEMU via QMP; but its +interactions have not historically considered from a security point of +view. For example, qmp_synchronous_send() waits for a response from +QEMU, which a compromised QEMU could simply not send (thus preventing +the toolstack from making forward progress). + +'''Implementation''': Audit toolstack interactions with QEMU which +happen after the guest has started running, and assume QEMU has been +compromised. + +### seccomp filtering (Linux only) + +'''Description''': Turn on seccomp filtering to disable syscalls which +QEMU doesn't need. + +'''Implementation''': Enable from the command-line: + + -sandboxon,obsolete=deny,elevateprivileges=allow,spawn=deny,resourcecontrol=deny
docs/designs/qemu-deprivilege.md new file +35
-0+ +`elevateprivileges` is currently required to allow `-runas` to work. +Removing this requirement would mean making sure that the uid change +happened before the seccomp2 call, perhaps by changing the uid before +executing QEMU. (But this would then require other changes to create +the QMP socket, VNC socket, and so on). + +It should be noted that `-sandbox` is implemented as a blacklist, not +a whitelist; that is, it disables known-unsed functionality which may +be harmful, rather than disabling all functionality except that known +to be safe and needed. This is unfortunately necessary since qemu +doesn't know what system calls libraries might end up making. (See +[lwn-seccomp] for a more complete discussion.) + +This feature is not on by default and may not be available in all +environments. We therefore need to either: + 1. Require that this feature be enabled to build qemu + 2. Check for `-sandbox` support at runtime before + +[lwn-seccomp]: https://lwn.net/Articles/738694/ + +### Disks + +The chroot (and seccomp?) happens late enough such that QEMU can +initialize itself and open its disks. If you want to add a disk at run +time via or insert a CD, you can't pass a path because QEMU is +chrooted. Instead use the add-fd QMP command and use +/dev/fdset/<fdset-id> as the path. + +A further layer of restriction could be to set RLIMIT_NOFILES to '0', +and hand all disks over QMP. + +## Migration + +When calling xen-save-devices-state, since QEMU is running in a chroot
docs/designs/qemu-deprivilege.md line 290
Quoted 10 lines
docs/designs/qemu-deprivilege.md new file +10
-0+it is not useful to pass a filename (it doesn't even have write access +inside the chroot). Instead, give it an open fd using the add-fd +mechanism. + +Additionally, all the restrictions need to be applied to the qemu +started up on the post-migration side. One issue that needs to be +solved is how to signal the toolstack on restore that qemu is ready +for the domain to be started (since this is normally done via +xenstore, and at this point the xenstore connections will have been +closed).
I thought Anthony had fixed this now?
Paul
Quoted 190 lines
docs/designs/qemu-deprivilege.md new file +19
-0+ +### Network namespacing (Linux only) + +Enter QEMU into its own network namespace (in addition to mount & IPC +namespaces): + + unshare(CLONE_NEWNET); + +QEMU does actually use the network namespace as a Xen DM for two +purposes: 1) To set up network tap devices 2) To open vnc connections. + +#### Network + +If QEMU runs in its own network namespace, it can't open the tap +device itself because the interface won't be visible outside of its +own namespace. So instead, have the toolstack open the device and pass +it as an fd on the command-line: + + -device rtl8139,netdev=tapnet0,mac=... -netdevtap,id=tapnet0,fd=<tapfd>
docs/designs/qemu-deprivilege.md new file +13
-0+ +#### VNC + +If QEMU runs in its own network namespace, it is not straightforward +to listen on a TCP socket outside of its own network namespace. One +option would be to use VNC over a UNIX socket: + + -vnc unix:/var/run/xen/vnc-<domid> + +However, this would break functionality in the general case; I think +we need to have the toolstack open a socket and pass the fd to QEMU +(which requires changes to QEMU). +docs/features/qemu- renamed from docs/features/qemu-deprivilege.pandoc +0
-0deprivilege.pandoc
new file mode 100644
index 0000000000..f941525189docs/features/qemu-deprivilege.pandoc new file +51
-0@@ -0,0 +1,101 @@ +% QEMU Deprivileging / dm_restrict +% Revision 1 + +\clearpage + +# Basics + +---------------- ---------------------------------------------------- + Status: **Tech Preview** + +Architecture(s): x86 + + Component(s): toolstack + +---------------- ---------------------------------------------------- + +# Overview + +By default, the QEMU device model is run in domain 0. If an attacker +can gain control of a QEMU process, it could easily take control of a +system. + +dm_restrict is a set of operations to restrict QEMU running in domain +0. It consists of two halves: + + 1. Mechanisms to restrict QEMU to only being able to affect its own +domain + 2. Mechanisms to restruct QEMU's ability to interact with domain 0. + +# User details + +## Getting the right versions of software + +Linux: 4.11+ + +Qemu: 3.0+ (Or the version that comes with Xen 4.12+) + +## Setting up a group and userid range + +For maximum security, libxl needs to run the devicemodel for each +domain under a user id (UID) corresponding to its domain id. There +are 32752 possible domain IDs, and so libxl needs 32752 user ids set +aside for it. Setting up a group for all devicemodels to run at is +also recommended. + +The simplest and most effective way to do this is to allocate a +contiguous block of UIDs, and create a single user named +`xen-qemuuser-range-base` with the first UID. For example, under +Debian: + + adduser --system --uid 131072 --group --no-create-home xen-qemuuser-range-base
docs/features/qemu-deprivilege.pandoc new file +50
-0+ +Two comments on this method: + + 1. Most modern systems have 32-bit UIDs, and so can in theory go up +to 2^31 (or 2^32 if uids are unsigned). POSIX only guarantees 16-bit +UIDs however; UID 65535 is reserved for an invalid value, and 65534 is +normally allocated to "nobody". + 2. Additionally, some container systems have proposed using the +upper 16 bits of the uid for a container ID. Using a multiple of 2^16 +for the range base (as is done above) will result in all UIDs being +interpreted by such systems as a single container ID. + +Another, less-secure way is to run all QEMUs as the same UID. To do +this, create a user named `xen-qemuuser-shared`; for example: + + adduser --no-create-home --system xen-qemuuser-shared + +## Domain config changes + +The core domain config change is to add the following line to the +domain configuration: + + dm_restrict=1 + +This will perform a number of restrictions, outlined below in the +'Technical details' section. + +# Technical details + +See docs/design/qemu-deprivilege.md for technical details. + +# Limitations + +The following features still need to be implemented: + * Inserting a new cdrom while the guest is running (xl cdrom-insert) + * Migration / save / restore + +Additionally, getting PCI passthrough to work securely would require a +significant rework of how passthrough works at the moment. It may be +implemented at some point but is not a near-term priority. + +See SUPPORT.md for security support status. + +# History + +------------------------------------------------------------------------ +Date Revision Version Notes +---------- -------- -------- ------------------------------------------- +2018-09-14 1 Xen 4.12 Imported from docs/misc +---------- -------- -------- -------------------------------------------docs/misc/qemu- renamed from docs/misc/qemu-deprivilege.txt +0
-0deprivilege.txt
deleted file mode 100644
index 58b86a3908..0000000000docs/misc/qemu-deprivilege.txt deleted +0
-35@@ -1,36 +0,0 @@ -For security reasons, libxl tries to pass a non-root username to QEMU as -argument. During initialization QEMU calls setuid and setgid with the -user ID and the group ID of the user passed as argument. -Libxl looks for the following users in this order: - -1) a user named "xen-qemuuser-domid$domid", -Where $domid is the domid of the domain being created. -This requires the reservation of 65535 uids from xen-qemuuser-domid1 -to xen-qemuuser-domid65535. To use this mechanism, you might want to -create a large number of users at installation time. For example: - -for ((i=1; i<65536; i++)) -do - adduser --no-create-home --system xen-qemuuser-domid$i -done - -You might want to consider passing --group to adduser to create a new -group for each new user. - - -2) a user named "xen-qemuuser-shared" -As a fall back if both 1) fails, libxl will use a single user for -all QEMU instances. The user is named xen-qemuuser-shared. This is -less secure but still better than running QEMU as root. Using this is as -simple as creating just one more user on your host: - -adduser --no-create-home --system xen-qemuuser-shared - - -3) root -As a last resort, libxl will start QEMU as root. - - -Please note that running QEMU as non-root causes several features like -migration and PCI passthrough to not work properly and may prevent theguest
-from booting.
--
2.19.1
Quoted 4 lines
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Quoted 7 lines · Paul Durrant
Quoted 7 lines
docs/designs/qemu-deprivilege.md new file +7
-0+## Migration + +When calling xen-save-devices-state, since QEMU is running in a chroot +it is not useful to pass a filename (it doesn't even have write access +inside the chroot). Instead, give it an open fd using the add-fd +mechanism. +
Quoted 9 lines
docs/designs/qemu-deprivilege.md line 290
Quoted 6 lines
docs/designs/qemu-deprivilege.md new file +6
-0+Additionally, all the restrictions need to be applied to the qemu +started up on the post-migration side. One issue that needs to be +solved is how to signal the toolstack on restore that qemu is ready +for the domain to be started (since this is normally done via +xenstore, and at this point the xenstore connections will have been +closed).I thought Anthony had fixed this now?
It's work in progress. (libxl__ev_qmp_* patch series.)
Signature · 6 lines
Anthony PERARD _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
George Dunlap writes ("[PATCH v4 1/6] docs/qemu-deprivilege: Revise and update with status and future plans"):
Quoted 2 lines
docs/qemu-deprivilege.txthad some basic instructions for usingdm_restrict, but it was incomplete, misleading, and stale.
(Assuming you fix the commit message to drop the mention of nonexistent SUPPORT.md changes...)
Thanks,
Ian.
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel