Featuresets are supposed to be disappearing when the CPU policy infrastructure
is complete, but that has taken longer than expected, and isn't going to be
complete imminently either.
In the meantime, Xen does have proper default/max featuresets, and xen-cpuid
can even get them via the XEN_SYSCTL_cpu_policy_* interface, but only knows
now to render them nicely via the featureset interface.
Differences between default and max are a frequent source of errors,
frequently too in secret leading up to an embargo, so extend the featureset
sysctl to allow xen-cpuid to render them all nicely.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Christian Lindig <christian.lindig@cloud.com>
All caml_alloc_*() functions can throw exceptions, and longjump out of
context. If this happens, we leak the xch/xce handle.
Reorder the logic to allocate the the Ocaml object first.
Fixes: 8b3c06a3e5 ("tools/ocaml/xenctrl: OCaml 5 support, fix use-after-free")
Fixes: 22d5affdf0 ("tools/ocaml/evtchn: OCaml 5 support, fix potential resource leak")
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
The modification to the _H() macro for Ocaml 5 support introduced a subtle
bug. From the manual:
https://ocaml.org/manual/intfc.html#ss:parallel-execution-long-running-c-code
"After caml_release_runtime_system() was called and until
caml_acquire_runtime_system() is called, the C code must not access any OCaml
data, nor call any function of the run-time system, nor call back into OCaml
code."
Previously, the value was a naked C pointer, so dereferencing it wasn't
"accessing any Ocaml data", but the fix to avoid naked C pointers added a
layer of indirection through an Ocaml Custom object, meaning that the common
pattern of using _H() in a blocking section is unsafe.
In order to fix:
* Drop the _H() macro and replace it with a static inline xch_of_val().
* Opencode the assignment into Data_custom_val() in the constructors.
* Rename "value xch" parameters to "value xch_val" so we can consistently
have "xc_interface *xch" on the stack, and obtain the pointer with the GC
lock still held.
* Drop the _D() macro while at it, because it's just pointless indirection.
Fixes: 8b3c06a3e5 ("tools/ocaml/xenctrl: OCaml 5 support, fix use-after-free")
Signed-off-by: Edwin Török <edwin.torok@cloud.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
The intf->{addr,len} references in the xc_map_foreign_range() call are unsafe.
From the manual:
https://ocaml.org/manual/intfc.html#ss:parallel-execution-long-running-c-code
"After caml_release_runtime_system() was called and until
caml_acquire_runtime_system() is called, the C code must not access any OCaml
data, nor call any function of the run-time system, nor call back into OCaml
code."
More than what the manual says, the intf pointer is (potentially) invalidated
by caml_enter_blocking_section() if another thread happens to perform garbage
collection at just the right (wrong) moment.
Rewrite the logic. There's no need to stash data in the Ocaml object until
the success path at the very end.
Fixes: 8b7ce06a2d ("ocaml: Add XC bindings.")
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
The patch adding this binding was plain broken, and unreviewed. It modified
the C stub to add a 4th parameter without an equivalent adjustment in the
Ocaml side of the bindings.
In 64bit builds, this causes us to dereference whatever dead value is in %rcx
when trying to interpret the rflags parameter.
This has gone unnoticed because Xapi doesn't use this binding (it has its
own), but unbreak the binding by passing RDM_RELAXED unconditionally for
now (matching the libxl default behaviour).
Fixes: 9b34056cb4 ("tools: extend xc_assign_device() to support rdm reservation policy")
Signed-off-by: Edwin Török <edwin.torok@cloud.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
* Remove local integers when all we're returning is Val_int() of another
variable. The CAMLlocal*() can't be optimised automatically, as it's
registered with the GC.
* Rename "virq_type" to "virq" and "_port" to "port".
* In stub_eventchn_pending(), rename 'port' to 'rc', to be consistent with
all other stubs that return xenevtchn_port_or_error_t.
* In stub_eventchn_unmask(), check for rc == -1 to be consistent with all
other stubs.
No practical change.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
The modification to the _H() macro for Ocaml 5 support introduced a subtle
bug. From the manual:
https://ocaml.org/manual/intfc.html#ss:parallel-execution-long-running-c-code
"After caml_release_runtime_system() was called and until
caml_acquire_runtime_system() is called, the C code must not access any OCaml
data, nor call any function of the run-time system, nor call back into OCaml
code."
Previously, the value was a naked C pointer, so dereferencing it wasn't
"accessing any Ocaml data", but the fix to avoid naked C pointers added a
layer of indirection through an Ocaml Custom object, meaning that the common
pattern of using _H() in a blocking section is unsafe.
In order to fix:
* Drop the _H() macro and replace it with a static inline xce_of_val().
* Opencode the assignment into Data_custom_val() in the two constructors.
* Rename "value xce" parameters to "value xce_val" so we can consistently
have "xenevtchn_handle *xce" on the stack, and obtain the pointer with the
GC lock still held.
Fixes: 22d5affdf0 ("tools/ocaml/evtchn: OCaml 5 support, fix potential resource leak")
Signed-off-by: Edwin Török <edwin.torok@cloud.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
caml_alloc() takes units of Wsize (word size), not bytes. As a consequence,
we're allocating 4 or 8 times too much memory.
Ocaml has a helper, Wsize_bsize(), but it truncates cases which aren't an
exact multiple. Use a BUILD_BUG_ON() to cover the potential for truncation,
as there's no rounding-up form of the helper.
Fixes: 8b7ce06a2d ("ocaml: Add XC bindings.")
Fixes: d3e649277a ("ocaml: add mmap bindings implementation.")
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
There is no such thing as an Ocaml function (C stub or otherwise) taking no
parameters. In the absence of any other parameters, unit is still passed.
This doesn't explode with any ABI we care about, but would malfunction for an
ABI environment such as stdcall.
Fixes: c3afd398ba ("ocaml: Add XS bindings.")
Fixes: 8b7ce06a2d ("ocaml: Add XC bindings.")
Signed-off-by: Edwin Török <edwin.torok@cloud.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
The configuration file can contain typos or various errors that could prevent
live update from succeeding (e.g. a flag only valid on a different version).
Unknown entries in the config file would be ignored on startup normally,
add a strict --config-test that live-update can use to check that the config file
is valid *for the new binary*.
For compatibility with running old code during live update recognize
--live --help as an equivalent to --config-test.
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
Using `ocp-indent` for now to just make minimal modifications in
tabs vs spaces and get the right indentation.
This avoids perpetuating a formatting style that is inconsistent with
the rest of Xen, and that makes preparing and submitting patches more
difficult (OCaml indentation tools usually only support spaces, not tabs).
No functional change.
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
This function is unusued (only Xs_ring.write_substring is used), and the
bytes/string conversion here is backwards: the C stub implements the bytes
version and then we use a Bytes.unsafe_of_string to convert a string into
bytes.
However the operation here really is read-only: we read from the string and
write it to the ring, so the C stub should implement the read-only string
version, and if needed we could use Bytes.unsafe_to_string to be able to send
'bytes'. However that is not necessary as the 'bytes' version is dropped above.
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
The manually established link order is fragile to code changes. Use ocamldep
to calculate the appropriate link order.
No functional change.
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
This is not strictly necessary since it is essentially a no-op currently: a
cast to void * and value *, even in OCaml 5.0.
However it does make it clearer that what we have here is not a regular OCaml
value, but one allocated with Abstract_tag or Custom_tag, and follows the
example from the manual more closely:
https://v2.ocaml.org/manual/intfc.html#ss:c-outside-head
It also makes it clearer that these modules have been reviewed for
compat with OCaml 5.0.
We cannot use OCaml finalizers here, because we want exact control over when
to unmap these pages from remote domains.
No functional change.
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
Not to be confused which hvm_get_param, which also exists and has a
different, more error-prone interface.
This one always returns a 64-bit value, and that is retained in the
OCaml binding as well, returning 'int64' (and not int, or nativeint
which might have different sizes).
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
There is no API or ioctl to query event channel status, it is only
present in xenctrl.h
The C union is mapped to an OCaml variant exposing just the value from
the correct union tag. This causes the xc bindings to now depend on
evtchn to get a useful API for EVTCHNSTAT_virq.
The information provided here is similar to 'lsevtchn', but rather than
parsing its output it queries the underlying API directly.
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
fallback_exception_handler feeds a string with embedded newlines directly into
syslog(). While this is an improvement on getting nothing, syslogd escapes
all control characters it gets, and emits one (long) log line.
Fix the problem generally in the syslog stub. As we already have a local copy
of the string, split it in place and emit one syslog() call per line.
Also tweak Logging.msg_of to avoid putting an extra newline on a string which
already ends with one.
Fixes: ee7815f49f ("tools/oxenstored: Set uncaught exception handler")
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
The binding for xc_interface_close() free the underlying handle while leaving
the Ocaml object still in scope and usable. This would make it easy to suffer
a use-after-free, if it weren't for the fact that the typical usage is as a
singleton that lives for the lifetime of the program.
Ocaml 5 no longer permits storing a naked C pointer in an Ocaml value.
Therefore, use a Custom block. This allows us to use the finaliser callback
to call xc_interface_close(), if the Ocaml object goes out of scope.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
strdup() may return NULL. Check for this before passing to syslog().
Drop const from c_msg. It is bogus, as demonstrated by the need to cast to
void * in order to free the memory.
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
Unhandled exceptions go to stderr by default, but this doesn't typically work
for oxenstored because:
* daemonize reopens stderr as /dev/null
* systemd redirects stderr to /dev/null too
Debugging an unhandled exception requires reproducing the issue locally when
using --no-fork, and is not conducive to figuring out what went wrong on a
remote system.
Install a custom handler which also tries to render the backtrace to the
configured syslog facility, and DAEMON|ERR otherwise.
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
During live update, oxenstored tries a best effort approach to recover as many
domains and information as possible even if it encounters errors restoring
some domains.
However, logging about misunderstood input is more severe than simply info.
Log it at warning instead.
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
Closing the evtchn handle will unbind and free all local ports. The new
xenstored would need to rebind all evtchns, which is work that we don't want
or need to be doing during the critical handover period.
However, it turns out that the Windows PV drivers also rebind their local port
too across suspend/resume, leaving (o)xenstored with a stale idea of the
remote port to use. In this case, reusing the established connection is the
only robust option.
Therefore:
* Have oxenstored open /dev/xen/evtchn without CLOEXEC at start of day.
* Extend the handover information with the evtchn fd, domexc virq local port,
and the local port number for each domain connection.
* Have (the new) oxenstored recover the open handle using Xeneventchn.fdopen,
and use the provided local ports rather than trying to rebind them.
When this new information isn't present (i.e. live updating from an oxenstored
prior to this change), the best-effort status quo will have to do.
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
Inter-domain event channels are always a pair of local and remote ports.
Right now the handling is asymmetric, caused by the fact that the evtchn is
bound after the associated Domain object is constructed.
First, move binding of the event channel into the Domain.make() constructor.
This means the local port no longer needs to be an option. It also removes
the final callers of Domain.bind_interdomain.
Next, introduce a new port_pair type to encapsulate the fact that these two
should be updated together, and replace the previous port and remote_port
fields. This refactoring also changes the Domain.get_port interface (removing
an option) so take the opportunity to name it get_local_port instead.
Also, this fixes a use-after-free risk with Domain.close. Once the evtchn has
been unbound, the same local port number can be reused for a different
purpose, so explicitly invalidate the ports to prevent their accidental misuse
in the future.
This also cleans up some of the debugging, to always print a port pair.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
Generally speaking, the event channel local/remote port is fixed for the
lifetime of the associated domain object. The exception to this is a
secondary XS_INTRODUCE (defined to re-bind to a new event channel) which pokes
around at the domain object's internal state.
We need to refactor the evtchn handling to support live update, so start by
moving the relevant manipulation into Domain.
No practical change.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
This will make the logic clearer when we plumb local_port through these
functions.
While doing this, rearrange the construct in Domains.create0 to separate the
remote port handling from the interface handling. (The interface logic is
dubious in several ways, but not altered by this cleanup.)
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
Xenstored always needs to bind the DOM_EXC VIRQ.
Instead of doing it shortly after the call to Event.init(), do it in the
constructor directly. This removes the need for the field to be a mutable
option.
It will also simplify a future change to support live update. Rename the
field from virq_port (which could be any VIRQ) to it's proper name.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
This file has some style problems so severe that they interfere with the
readability of the subsequent bugfix patches.
Fix these issues ahead of time, to make the subsequent changes more readable.
No functional change.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
For live update, oxenstored wants to clear CLOEXEC on the evtchn handle, so it
survives the execve() into the new oxenstored.
Have the new interface match how cloexec works in other Ocaml standard
libraries.
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
There is no binding for xenevtchn_close(). In principle, this is a resource
leak, but the typical usage is as a singleton that lives for the lifetime of
the program.
Ocaml 5 no longer permits storing a naked C pointer in an Ocaml value.
Therefore, use a Custom block. This allows us to use the finaliser callback
to call xenevtchn_close(), if the Ocaml object goes out of scope.
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
A debug statement got inserted into a single-expression if statement.
Insert brackets to give the intended meaning, rather than the actual meaning
where the "let con = Connections..." is outside and executed unconditionally.
This results in some unnecessary ring checks for domains which otherwise have
IO credit.
Fixes: 42f0581a91 ("tools/oxenstored: Implement live update for socket connections")
Reported-by: Edwin Török <edvin.torok@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
domain_getinfolist() is quadratic with the number of domains, because of the
behaviour of the underlying hypercall. Nevertheless, getting domain info in
blocks of 1024 is far more efficient than blocks of 2.
In a scalability testing scenario with ~1000 VMs, a combination of this and
the previous change takes xenopsd's wallclock time in domain_getinfolist()
down from 88% to 0.02%
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Tested-by: Pau Ruiz Safont <pau.safont@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
domain_getinfolist() is quadratic with the number of domains, because of the
behaviour of the underlying hypercall. xenopsd was further observed to be
wasting excessive quantites of time manipulating the list of already-obtained
domains.
Implement a tail recursive `rev_concat` equivalent to `concat |> rev`, and use
it instead of calling `@` multiple times.
An incidental benefit is that the list of domains will now be in domid order,
instead of having pairs of 2 domains changing direction every time.
In a scalability testing scenario with ~1000 VMs, a combination of this and
the subsequent change takes xenopsd's wallclock time in domain_getinfolist()
down from 88% to 0.02%
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Tested-by: Pau Ruiz Safont <pau.safont@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
I was really hoping to avoid this, but its now too late in the 4.17 freeze and
we still don't have working fixes.
The in-Xen calculations for assistance capabilities are buggy. For the
avoidance of doubt, the original intention was to be able to control every
aspect of a APIC acceleration so we could comprehensively test Xen's support,
as it has proved to be buggy time and time again.
Even after a protracted discussion on what the new API ought to mean, attempts
to apply it to the existing logic have been unsuccessful, proving that the
API/ABI is too complicated for most people to reason about.
This reverts most of:
2ce11ce2496b2b9b3405
leaving in place the non-APIC specific changes (minimal as they are).
This takes us back to the behaviour of Xen 4.16 where APIC acceleration is
configured on a per system basis.
This work will be revisted in due course.
Fixes: 2ce11ce249 ("x86/HVM: allow per-domain usage of hardware virtualized APIC")
Fixes: 6b2b9b3405 ("x86: report Interrupt Controller Virtualization capabilities")
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
Release-acked-by: Henry Wang <Henry.Wang@arm.com>
Building with Dune in release mode fails with:
```
File "ocaml/xenstored/store.ml", line 464, characters 13-32:
Warning 18: this type-based record disambiguation is not principal.
File "ocaml/xenstored/store.ml", line 1:
Error: Some fatal warnings were triggered (1 occurrences)
```
This is a warning to help keep the code futureproof, quoting from its
documentation:
> Check information path during type-checking, to make sure that all types are
> derived in a principal way. When using labelled arguments and/or polymorphic
> methods, this flag is required to ensure future versions of the compiler will
> be able to infer types correctly, even if internal algorithms change. All
> programs accepted in -principal mode are also accepted in the default mode with
> equivalent types, but different binary signatures, and this may slow down type
> checking; yet it is a good idea to use it once before publishing source code.
Fixes: db471408ed "tools/ocaml/xenstored: Fix quota bypass on domain shutdown"
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
Release-acked-by: Henry Wang <Henry.Wang@arm.com>
Integers in Ocaml have 63 or 31 bits of signed precision.
On 64-bit builds of Ocaml, this is fine because a C uint32_t always fits
within a 63-bit signed integer.
In 32-bit builds of Ocaml, this goes wrong. The C uint32_t is truncated
first (loses the top bit), then has a unsigned/signed mismatch.
A "negative" value (i.e. a packet on the ring of between 1G and 2G in size)
will trigger an exception later in Bytes.make in xb.ml, and because the packet
is not removed from the ring, the exception re-triggers on every subsequent
query, creating a livelock.
Fix both the source of the exception in Xb, and as defence in depth, mark the
domain as bad for any Invalid_argument exceptions to avoid the risk of
livelock.
This is XSA-420 / CVE-2022-42324.
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
XSA-322 fixed a domid reuse vulnerability by assigning Dom0 as the owner of
any nodes left after a domain is shutdown (e.g. outside its /local/domain/N
tree).
However Dom0 has no quota on purpose, so this opened up another potential
attack vector. Avoid it by deleting these nodes instead of assigning them to
Dom0.
This is part of XSA-419 / CVE-2022-42323.
Fixes: c46eff9212 ("tools/ocaml/xenstored: clean up permissions for dead domains")
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
Introduce a limit on the number of outstanding reply packets in the xenbus
queue. This limits the number of in-flight requests: when the output queue is
full we'll stop processing inputs until the output queue has room again.
To avoid a busy loop on the Unix socket we only add it to the watched input
file descriptor set if we'd be able to call `input` on it. Even though Dom0
is trusted and exempt from quotas a flood of events might cause a backlog
where events are produced faster than daemons in Dom0 can consume them, which
could lead to an unbounded queue size and OOM.
Therefore the xenbus queue limit must apply to all connections, Dom0 is not
exempt from it, although if everything works correctly it will eventually
catch up.
This prevents a malicious guest from sending more commands while it has
outstanding watch events or command replies in its input ring. However if it
can cause the generation of watch events by other means (e.g. by Dom0, or
another cooperative guest) and stop reading its own ring then watch events
would've queued up without limit.
The xenstore protocol doesn't have a back-pressure mechanism, and doesn't
allow dropping watch events. In fact, dropping watch events is known to break
some pieces of normal functionality. This leaves little choice to safely
implement the xenstore protocol without exposing the xenstore daemon to
out-of-memory attacks.
Implement the fix as pipes with bounded buffers:
* Use a bounded buffer for watch events
* The watch structure will have a bounded receiving pipe of watch events
* The source will have an "overflow" pipe of pending watch events it couldn't
deliver
Items are queued up on one end and are sent as far along the pipe as possible:
source domain -> watch -> xenbus of target -> xenstore ring/socket of target
If the pipe is "full" at any point then back-pressure is applied and we prevent
more items from being queued up. For the source domain this means that we'll
stop accepting new commands as long as its pipe buffer is not empty.
Before we try to enqueue an item we first check whether it is possible to send
it further down the pipe, by attempting to recursively flush the pipes. This
ensures that we retain the order of events as much as possible.
We might break causality of watch events if the target domain's queue is full
and we need to start using the watch's queue. This is a breaking change in
the xenstore protocol, but only for domains which are not processing their
incoming ring as expected.
When a watch is deleted its entire pending queue is dropped (no code is needed
for that, because it is part of the 'watch' type).
There is a cache of watches that have pending events that we attempt to flush
at every cycle if possible.
Introduce 3 limits here:
* quota-maxwatchevents on watch event destination: when this is hit the
source will not be allowed to queue up more watch events.
* quota-maxoustanding which is the number of responses not read from the ring:
once exceeded, no more inputs are processed until all outstanding replies
are consumed by the client.
* overflow queue on the watch event source: all watches that cannot be stored
on destination are queued up here, a single command can trigger multiple
watches (e.g. due to recursion).
The overflow queue currently doesn't have an upper bound, it is difficult to
accurately calculate one as it depends on whether you are Dom0 and how many
watches each path has registered and how many watch events you can trigger
with a single command (e.g. a commit). However these events were already
using memory, this just moves them elsewhere, and as long as we correctly
block a domain it shouldn't result in unbounded memory usage.
Note that Dom0 is not excluded from these checks, it is important that Dom0 is
especially not excluded when it is the source, since there are many ways in
which a guest could trigger Dom0 to send it watch events.
This should protect against malicious frontends as long as the backend follows
the PV xenstore protocol and only exposes paths needed by the frontend, and
changes those paths at most once as a reaction to guest events, or protocol
state.
The queue limits are per watch, and per domain-pair, so even if one
communication channel would be "blocked", others would keep working, and the
domain itself won't get blocked as long as it doesn't overflow the queue of
watch events.
Similarly a malicious backend could cause the frontend to get blocked, but
this watch queue protects the frontend as well as long as it follows the PV
protocol. (Although note that protection against malicious backends is only a
best effort at the moment)
This is part of XSA-326 / CVE-2022-42318.
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
Ensures we cannot store more than [capacity] elements in a [Queue]. Replacing
all Queue with this module will then ensure at compile time that all Queues
are correctly bound checked.
Each element in the queue has a class with its own limits. This, in a
subsequent change, will ensure that command responses can proceed during a
flood of watch events.
No functional change.
This is part of XSA-326.
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
The queue here would only ever hold at most one element. This will simplify
follow-up patches.
This is part of XSA-326.
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
By default the OCaml garbage collector would return memory to the OS only
after unused memory is 5x live memory. Tweak this to 120% instead, which
would match the major GC speed.
This is part of XSA-326.
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
Previously we'd perform the operation, record the updated tree in the
transaction record, then try to insert a watchop path and the reply packet.
If we exceeded max requests we would've returned EQUOTA, but still:
* have performed the operation on the transaction's tree
* have recorded the watchop, making this queue effectively unbounded
It is better if we check whether we'd have room to store the operation before
performing the transaction, and raise EQUOTA there. Then the transaction
record won't grow.
This is part of XSA-326 / CVE-2022-42317.
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
We currently have 2 different set of defaults in upstream Xen git tree:
* defined in the source code, only used if there is no config file
* defined in the oxenstored.conf.in upstream Xen
An oxenstored.conf file is not mandatory, and if missing, maxrequests in
particular has an unsafe default.
Resync the defaults from oxenstored.conf.in into the source code.
This is part of XSA-326 / CVE-2022-42316.
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
During live update we will load the /tool/xenstored path from the previous binary,
and then try to mkdir /tool again which will fail with EEXIST.
Check for existence of the path before creating it.
The write call to /tool/xenstored should not need any changes
(and we do want to overwrite any previous path, in case it changed).
Prior to 7110192b1d live update would work only if the binary path was
specified, and with 7110192b1d and this live update also works when
no binary path is specified in `xenstore-control live-update`.
Fixes: 7110192b1d ("tools/oxenstored: Fix Oxenstored Live Update")
Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
Release-acked-by: Henry Wang <Henry.Wang@arm.com>
Using a full path to the C libraries when preparing one of the ocaml
binding for those libraries make the binding unusable by external
project. The full path is somehow embedded and reused by the external
project when linking against the binding.
Instead, we will use the proper way to link a library, by using '-l'.
For in-tree build, we also need to provide the search directory via
'-L'.
(The search path -L are still be embedded, but at least that doesn't
prevent the ocaml binding from been used.)
Related-to: xen-project/xen#96
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Acked-by: Christian Lindig <christian.lindig@citrix.com>
Tested-by: Andrew Cooper <andrew.cooper3@citrix.com>
Release-acked-by: Henry Wang <Henry.Wang@arm.com>
Andrew Cooper
Edwin Török
Edwin Török
Anthony PERARD