Gitlab Randconfig found:
arch/x86/mm/shadow/common.c: In function 'shadow_prealloc':
arch/x86/mm/shadow/common.c:1023:18: error: implicit declaration of function
'paging_logdirty_levels'; did you mean 'paging_log_dirty_init'? [-Werror=implicit-function-declaration]
1023 | count += paging_logdirty_levels();
| ^~~~~~~~~~~~~~~~~~~~~~
| paging_log_dirty_init
arch/x86/mm/shadow/common.c:1023:18: error: nested extern declaration of 'paging_logdirty_levels' [-Werror=nested-externs]
The '#if PG_log_dirty' expression is currently SHADOW_PAGING && !HVM &&
PV_SHIM_EXCLUSIVE. Move the declaration outside.
Fixes: 33fb3a6612 ("x86/shadow: account for log-dirty mode when pre-allocating")
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Since c/s 73b13705af ("firmware: provide a stand alone set of headers"),
we've had a proper stddef.h. Actually use it.
No functional change.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
HVMLoader is not relocatable in memory, and 32bit PIC code has a large
overhead. Override the compilers choice of pic/no-pic and force it to be
non-relocatable.
Bloat-o-meter reports a net:
add/remove: 0/0 grow/shrink: 3/107 up/down: 14/-3370 (-3356)
No functional change.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
There are two problems with domain_crash().
First, that it is frequently not preceded by a printk() at all, or only by a
dprintk(). Either way, critical diagnostic information is missing for an
event which is fatal to the guest.
Second, the embedded __LINE__ is an issue for livepatching, creating unwanted
churn in the binary diff. This is the final __LINE__ remaining in
livepatching-relevant contexts.
The end goal is to have domain_crash() require a print string which gets fed
to printk(), making it far less easy to omit relevant diagnostic information.
However, modifying all callers at once is far too big and complicated, so use
some macro magic to tolerate the old API (no print string) in the short term.
Adjust two callers in load_segments() to demonstrate the new API.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Julien Grall <jgrall@amazon.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
It turns out that evaluate_nospec() code generation is not safe under Clang.
Given:
void eval_nospec_test(int x)
{
if ( evaluate_nospec(x) )
asm volatile ("nop #true" ::: "memory");
else
asm volatile ("nop #false" ::: "memory");
}
Clang emits:
<eval_nospec_test>:
0f ae e8 lfence
85 ff test %edi,%edi
74 02 je <eval_nospec_test+0x9>
90 nop
c3 ret
90 nop
c3 ret
which is not safe because the lfence has been hoisted above the conditional
jump. Clang concludes that both barrier_nospec_true()'s have identical side
effects and can safely be merged.
Clang can be persuaded that the side effects are different if there are
different comments in the asm blocks. This is fragile, but no more fragile
that other aspects of this construct.
Introduce barrier_nospec_false() with a separate internal comment to prevent
Clang merging it with barrier_nospec_true() despite the otherwise-identical
content. The generated code now becomes:
<eval_nospec_test>:
85 ff test %edi,%edi
74 05 je <eval_nospec_test+0x9>
0f ae e8 lfence
90 nop
c3 ret
0f ae e8 lfence
90 nop
c3 ret
which has the correct number of lfence's, and in the correct place.
Link: https://github.com/llvm/llvm-project/issues/55084
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
We shouldn't be using two struct iovec's to write half of 'rec' each, and
there is no need to malloc() for two struct iovec's at all.
Simplify down to just two - one covering the whole of 'rec', and one covering
the pfns array.
No functional change.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Olaf Hering <olaf@aepfle.de>
OpenSBI already expands \n to \r\n. Don't repeat the expansion, as it doubles
the size of the resulting log with every other line being blank.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Rewrite argo_dprintk() so printk() format typechecking can always be
performed. This also fixes the fact that parameters are not evaulated at all
in the default case.
Emit the messages at XENLOG_DEBUG.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jason Andryuk <jandryuk@gmail.com>
XEN_DOMCTL_CDF_oos_off is forced set for PV domains, so the logic can't
ever be engaged for them. Conditionalize respective fields and remove
the respective bit from SHADOW_OPTIMIZATIONS when !HVM. As a result the
SH_type_oos_snapshot constant can disappear altogether in that case, and
a couple of #ifdef-s can also be dropped/combined.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
These aren't mode dependent (see 06f04f54ba ["x86/shadow:
sh_{write,cmpxchg}_guest_entry() are PV-only"], where they were moved
out of multi.c) and hence there's no need to have pointers to the
functions in struct shadow_paging_mode. Due to include dependencies,
however, the "paging" wrappers need to move out of paging.h; they're
needed from PV memory management code only anyway, so by moving them
their exposure is reduced at the same time.
By carefully placing the (moved and renamed) shadow function
declarations, #ifdef can also be dropped from the "paging" wrappers
(paging_mode_shadow() is constant false when !SHADOW_PAGING).
While moving the code, drop the (largely wrong) comment from
paging_write_guest_entry() and reduce that of
paging_cmpxchg_guest_entry().
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
vchan/init.c is defining offsetof privately. Remove that definition
and just use stddef.h instead.
Signed-off-by: Juergen Gross <jgross@suse.com>
Acked-by: Anthony PERARD <anthony.perard@citrix.com>
xfs/fsys_xfs.c is defining offsetof privately. Remove that definition
and just use stddef.h instead.
Signed-off-by: Juergen Gross <jgross@suse.com>
Acked-by: Anthony PERARD <anthony.perard@citrix.com>
util.h contains a definition of offsetof(), which isn't needed, as
firmware/include/stddef.h's doesn't really need overriding.
Remove it.
Signed-off-by: Juergen Gross <jgross@suse.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
Instead of having 3 identical copies of the definition of a
container_of() macro in different tools header files, add that macro
to xen-tools/common-macros.h and use that instead.
Delete the other copies of that macro.
Signed-off-by: Juergen Gross <jgross@suse.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Anthony PERARD <anthony.perard@citrix.com>
Defining min(), min_t(), max() and max_t() at other places than
xen-tools/common-macros.h isn't needed, as the definitions in said
header can be used instead.
Same applies to BUILD_BUG_ON() in hvmloader.
Signed-off-by: Juergen Gross <jgross@suse.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Anthony PERARD <anthony.perard@citrix.com>
sh_update_paging_modes() as its last action already invokes
sh_update_cr3(). Therefore there is no reason to invoke update_cr3()
another time immediately after calling paging_update_paging_modes(),
especially as sh_update_cr3() does not short-circuit the "nothing
changed" case.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Ordinary scalar operations are used in a multitude of other places, so
do so here as well. In fact take the opportunity and drop a local
variable then as well, first and foremost to get rid of a bogus cast.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
The "32b" and "pae" functions are identical at the source level (they
differ in what they get compiled to, due to differences in
SHADOW_FOREACH_L2E()), leaving aside a comment the PAE variant has and
the non-PAE one doesn't. Replace these infixes by the more usual l<N>
ones (and then also for the "64b" one for consistency; that'll also
allow for re-use once we support 5-level paging, if need be). The two
different instances are still distinguishable by their "level" suffix.
While fiddling with the names, convert the last parameter to boolean
as well.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
While no caller currently invokes the function without first making sure
there is at least one shadow [1], we'd better eliminate UB here:
find_first_set_bit() requires input to be non-zero to return a well-
defined result.
Further, using find_first_set_bit() isn't very efficient in the first
place for the intended purpose.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
[1] The function has exactly two uses, and both are from OOS code, which
is HVM-only. For HVM (but not for PV) sh_mfn_is_a_page_table(),
guarding the call to sh_unsync(), guarantees at least one shadow.
Hence even if sh_page_has_multiple_shadows() returned a bogus value
when invoked for a PV domain, the subsequent is_hvm_vcpu() and
oos_active checks (the former being redundant with the latter) will
compensate. (Arguably that oos_active check should come first, for
both clarity and efficiency reasons.)
Rework the config parsing of a p9 device to use the
split_string_into_pair() function instead of open coding it.
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Jason Andryuk <jandryuk@gmail.com>
Acked-by: Anthony PERARD <anthony.perard@citrix.com>
Today split_string_into_pair() will not really do what its name is
suggesting: instead of splitting a string into a pair of strings using
a delimiter, it will return the first two strings of the initial string
by using the delimiter.
This is never what the callers want, so modify split_string_into_pair()
to split the string only at the first delimiter found, resulting in
something like "x=a=b" to be split into "x" and "a=b" when being called
with "=" as the delimiter. Today the returned strings would be "x" and
"a".
At the same time switch the delimiter from "const char *" (allowing
multiple delimiter characters) to "char" (a single character only), as
this makes the function more simple without breaking any use cases.
Suggested-by: Anthony PERARD <anthony.perard@citrix.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Jason Andryuk <jandryuk@gmail.com>
Acked-by: Anthony PERARD <anthony.perard@citrix.com>
Instead of duplicating libxl__device_console_add() work in
init-xenstore-domain.c, just use libxenlight.
This requires to add a small wrapper function to libxenlight, as
libxl__device_console_add() is an internal function.
This at once removes a theoretical race between starting xenconsoled
and xenstore-stubdom, as the old code wasn't using a single
transaction for writing all the entries, leading to the possibility
that xenconsoled would see only some of the entries being written.
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Anthony PERARD <anthony.perard@citrix.com>
If the scope for IGD's IOMMU contains additional device that doesn't
actually exist, iommu=no-igfx would not disable that IOMMU. In this
particular case (Thinkpad x230) it included 00:02.1, but there is no
such device on this platform. Consider only existing devices for the
"gfx only" check as well as the establishing of IGD DRHD address
(underlying is_igd_drhd(), which is used to determine applicability of
two workarounds).
Fixes: 2d7f191b39 ("VT-d: generalize and correct "iommu=no-igfx" handling")
Signed-off-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Most use cases of split_string_into_pair() are requiring the returned
strings to be white space trimmed.
In order to avoid the same code pattern multiple times, add a predicate
parameter to split_string_into_pair() which can be specified to call
trim() with that predicate for the string pair returned. Specifying
NULL for the predicate will avoid the call of trim().
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Jason Andryuk <jandryuk@gmail.com>
Acked-by: Anthony PERARD <anthony.perard@citrix.com>
These are used from common code, so their signatures should be
consistent across architectures. This is achieved / guaranteed easiest
when their declarations are in a common header.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Julien Grall <jgrall@amazon.com>
In order to better reflect the contents of the header and to make it
more appropriate to use it for different runtime environments like
programs, libraries, and firmware, rename the libs.h include file to
common-macros.h. Additionally add a comment pointing out the need to be
self-contained.
Suggested-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Acked-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com> # tools/python/xen/lowlevel/xc/xc.c
Acked-by: Christian Lindig <christian.lindig@cloud.com>
Acked-by: Anthony PERARD <anthony.perard@citrix.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
As stated (correctly) by the comment next to SPEC_CTRL_ENTRY_FROM_PV, between
the two hunks visible in the patch, RET's are not safe prior to this point.
CR4_PV32_RESTORE hides a CALL/RET pair in certain configurations (PV32
compiled in, SMEP or SMAP active), and the RET can be attacked with one of
several known speculative issues.
Furthermore, CR4_PV32_RESTORE also hides a reference to the cr4_pv32_mask
global variable, which is not safe when XPTI is active before restoring Xen's
full pagetables.
This crash has gone unnoticed because it is only AMD CPUs which permit the
SYSCALL instruction in compatibility mode, and these are not vulnerable to
Meltdown so don't activate XPTI by default.
This is XSA-429 / CVE-2022-42331
Fixes: 5e79629011 ("x86/entry: Organise the use of MSR_SPEC_CTRL at each entry/exit point")
Fixes: 5784de3e20 ("x86: Meltdown band-aid against malicious 64-bit PV guests")
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
While the RCU variants of list insertion and removal allow lockless list
traversal (with RCU just read-locked), insertions and removals still
need serializing amongst themselves. To keep things simple, use the
domain lock for this purpose.
This is CVE-2022-42334 / part of XSA-428.
Fixes: 642123c512 ("x86/hvm: provide XEN_DMOP_pin_memory_cacheattr")
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Julien Grall <jgrall@amazon.com>
This is exposed via DMOP, i.e. to potentially not fully privileged
device models. With that we may not permit registration of an (almost)
unbounded amount of such regions.
This is CVE-2022-42333 / part of XSA-428.
Fixes: 642123c512 ("x86/hvm: provide XEN_DMOP_pin_memory_cacheattr")
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Pre-allocation is intended to ensure that in the course of constructing
or updating shadows there won't be any risk of just made shadows or
shadows being acted upon can disappear under our feet. The amount of
pages pre-allocated then, however, needs to account for all possible
subsequent allocations. While the use in sh_page_fault() accounts for
all shadows which may need making, so far it didn't account for
allocations coming from log-dirty tracking (which piggybacks onto the
P2M allocation functions).
Since shadow_prealloc() takes a count of shadows (or other data
structures) rather than a count of pages, putting the adjustment at the
call site of this function won't work very well: We simply can't express
the correct count that way in all cases. Instead take care of this in
the function itself, by "snooping" for L1 type requests. (While not
applicable right now, future new request sites of L1 tables would then
also be covered right away.)
It is relevant to note here that pre-allocations like the one done from
shadow_alloc_p2m_page() are benign when they fall in the "scope" of an
earlier pre-alloc which already included that count: The inner call will
simply find enough pages available then; it'll bail right away.
This is CVE-2022-42332 / XSA-427.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Tim Deegan <tim@xen.org>
This removes most of the opencoded bit logic on the exit qualification.
Unfortunately, size is 1-based not 0-based, so need adjusting in a separate
variable.
No functional change.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
An earlier change with the same title (commit 1ba66a870e) altered only
the path where x2apic_phys was already set to false (perhaps from the
command line). The same of course needs applying when the variable
wasn't modified yet from its initial value.
Reported-by: Elliott Mitchell <ehem+xen@m5p.com>
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Create 2 new test jobs, called qemu-smoke-dom0less-arm64-gcc-static-shared-mem
and qemu-smoke-dom0less-arm64-gcc-debug-static-shared-mem.
Adjust qemu-smoke-dom0less-arm64.sh script to accomodate the static
shared memory test as a new test variant. The test variant is determined
based on the first argument passed to the script. For testing static
shared memory, the argument is 'static-shared-mem'.
The test configures two dom0less DOMUs with a static shared memory
region and adds a check in the init script.
The check consists in comparing the contents of the /proc/device-tree/reserved-memory
xen-shmem entry with the static shared memory range and id with which
DOMUs were configured. If the memory layout is correct, a message gets
printed by DOMU.
At the end of the qemu run, the script searches for the specific message
in the logs and fails if not found.
Signed-off-by: Jiamei Xie <jiamei.xie@arm.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
Create 2 new test jobs, called qemu-smoke-dom0less-arm64-gcc-staticheap
and qemu-smoke-dom0less-arm64-gcc-debug-staticheap.
Add property "xen,static-heap" under /chosen node to enable static-heap.
If the domU can start successfully with static-heap enabled, then this
test pass.
ImageBuillder sets the kernel and ramdisk range based on the file size.
It will use the memory range between 0x45600000 to 0x47AED1E8. It uses
MEMORY_START and MEMORY_END from the cfg file as a range in which it can
instruct u-boot where to place the images.
Change MEMORY_END to 0x50000000 for all test cases.
Signed-off-by: Jiamei Xie <jiamei.xie@arm.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
At the moment, we direct serial input to hardware domain by default.
This does not make any sense when running in true dom0less mode, since
such domain does not exist. As a result, users wishing to write to
an emulated UART of a domU are always forced to execute CTRL-AAA first.
The same issue is when rotating among serial inputs, where we always
have to go through hardware domain case. This problem can be elaborated
further to all the domains that no longer exist.
Modify switch_serial_input() so that we skip switching serial input to
non existing domains. Take the opportunity to define and make use of
macro max_console_rx to make it clear what 'max_init_domid + 1' means
in the console code context. Also, modify call to printk() to use correct
format specifier for unsigned int.
For now, to minimize the required changes and to match the current
behavior with hwdom, the default input goes to the first real domain.
The choice is more or less arbitrary since dom0less domUs are supposedly
equal. This will be handled in the future by adding support in boot time
configuration for marking a specific domain preferred in terms of
directing serial input to.
Signed-off-by: Michal Orzel <michal.orzel@amd.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
The emulated PIIX3 uses a nybble for the status of each PCI function,
so the status for e.g. slot 0 functions 0 and 1 respectively can be
read as (\_GPE.PH00 & 0x0F), and (\_GPE.PH00 >> 0x04).
The AML that Xen gives to a guest gets the operand order for the odd-
numbered functions the wrong way round, returning (0x04 >> \_GPE.PH00)
instead.
As far as I can tell, this was the wrong way round in Xen from the
moment that PCI hotplug was first introduced in commit 83d82e6f35:
+ ShiftRight (0x4, \_GPE.PH00, Local1)
+ Return (Local1) /* IN status as the _STA */
Or maybe there's bizarre AML operand ordering going on there, like
Intel's wrong-way-round assembler, and it only broke later when it was
changed to being generated?
Either way, it's definitely wrong now, and instrumenting a Linux guest
shows that it correctly sees _STA being 0x00 in function 0 of an empty
slot, but then the loop in acpiphp_glue.c::get_slot_status() goes on to
look at function 1 and sees that _STA evaluates to 0x04. Thus reporting
an adapter is present in every slot in /sys/bus/pci/slots/*
Quite why Linux wants to look for function 1 being physically present
when function 0 isn't... I don't want to think about right now.
Fixes: 83d82e6f35 ("hvmloader: pass-through: multi-function PCI hot-plug")
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
This is mandated by the Fedora packaging guidelines because it is a security
vulnerability otherwise in suid scripts. While Xen doesn't have suid scripts,
it's a very good idea generally because it prevents the users local python
environment interfering from system packaged scripts.
pygrub is the odd-script-out, being installed by distutils rather than
manually with INSTALL_PYTHON_PROG. distutils has no nice way of editing the
shebang, so arrange to use INSTALL_PYTHON_PROG for pygrub too.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
* Add X86_{CPUID,MSR}_POLICY_FORMAT checks which were missed previously.
* Drop test_suite(). It hasn't been necessary since the Py2.3 era.
* Drop the __main__ logic. This can't be used without manually adjusting the
include path, and `make test` knows how to do the right thing.
* For `make test`, use `-v` to see which tests have been discovered and run.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
This script is not python3 compatible, but has its shebang altered to say
python3 by INSTALL_PYTHON_PROG.
The most recent reference I can find to this script (which isn't incidental
adjustments in the makefile) is from the Xen book, fileish 561e30b804 which
says
%% <snip> Alternatively, if the
%% Xen machine is connected to a serial-port server then we supply a
%% dumb TCP terminal client, {\tt xencons}.
So this a not-invented-here version of telnet. Delete it.
Resolves: xen-project/xen#159
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Anthony PERARD <anthony.perard@citrix.com>
Acked-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
This removes raw number manipulation, and makes the logic easier to follow.
No functional change.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Andrew Cooper