]> xenbits.xensource.com Git - xen.git/log
xen.git
13 months agochar: lpuart: Drop useless variables from UART structure
Michal Orzel [Thu, 4 Apr 2024 07:51:43 +0000 (09:51 +0200)]
char: lpuart: Drop useless variables from UART structure

These variables are useless. They are being assigned a value which is
never used since UART is expected to be pre-configured.

No functional change.

Signed-off-by: Michal Orzel <michal.orzel@amd.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
13 months agodrivers: char: Drop useless suspend/resume stubs in Arm drivers
Michal Orzel [Thu, 4 Apr 2024 07:51:42 +0000 (09:51 +0200)]
drivers: char: Drop useless suspend/resume stubs in Arm drivers

On Arm we don't use console_{suspend,resume} and the corresponding
stubs in serial drivers are being redundantly copied whenever a new
driver is added. Drop them as well as useless .endboot = NULL assignment.

Signed-off-by: Michal Orzel <michal.orzel@amd.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Stefano Stabellini <sstabellini@kernel.org>
13 months agovsprintf: address violations of MISRA C:2012 Rule 16.3
Federico Serafini [Thu, 4 Apr 2024 10:01:50 +0000 (12:01 +0200)]
vsprintf: address violations of MISRA C:2012 Rule 16.3

MISRA C:2012 Rule 16.3 states: "An unconditional `break' statement
shall terminate every switch-clause".

In order to meet the requirements to deviate the rule:
1) refactor the for loop to make the switch-clause ending with a
   return statement (note that adding a break at the end of the
   switch-clause would result in a violation of Rule 2.1
   "A project shall not contain unreachable code");
2) add pseudo-keyword fallthrough.

No functional change.

Signed-off-by: Federico Serafini <federico.serafini@bugseng.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen/sched: address violations of MISRA C:2012 Rule 16.3
Federico Serafini [Thu, 4 Apr 2024 10:01:27 +0000 (12:01 +0200)]
xen/sched: address violations of MISRA C:2012 Rule 16.3

Add break statement to address a violation of MISRA C:2012 Rule 16.3
("An unconditional `break' statement shall terminate every
switch-clause").
Replace deprecated comment /* FALLTHRU */ with pseudo-keyword
fallthrough to meet the requirements to deviate the rule.

No functional change.

Signed-off-by: Federico Serafini <federico.serafini@bugseng.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Acked-by: George Dunlap <george.dunlap@cloud.com>
13 months agotools: Move MB/GB() to common-macros.h
Jason Andryuk [Thu, 4 Apr 2024 10:01:13 +0000 (12:01 +0200)]
tools: Move MB/GB() to common-macros.h

Consolidate to a single set of common macros for tools.

MB() will gain another use in libelf, so this movement makes it
available.

Requested-by: Jan Beulich <jbeulich@suse.com>
Signed-off-by: Jason Andryuk <jason.andryuk@amd.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Anthony PERARD <anthony.perard@citrix.com>
13 months agotools/init-xenstore-domain: Replace variable MB() usage
Jason Andryuk [Thu, 4 Apr 2024 10:00:59 +0000 (12:00 +0200)]
tools/init-xenstore-domain: Replace variable MB() usage

The local MB() & GB() macros will be replaced with a common
implementation, but those only work with numeric values.  Introduce a
static inline mb_to_bytes() in place of the MB() macro to convert the
variable values, at the same time addressing the lack of
parenthesization of the prior macro's parameter.

Fixes: 134d53f57707 ("tools/init-xenstore-domain: fix memory map for PVH stubdom")
Signed-off-by: Jason Andryuk <jason.andryuk@amd.com>
Reviewed-by: Anthony PERARD <anthony.perard@citrix.com>
13 months agoxen/riscv: introduce extension support check by compiler
Oleksii Kurochko [Thu, 4 Apr 2024 10:00:17 +0000 (12:00 +0200)]
xen/riscv: introduce extension support check by compiler

Currently, RISC-V requires two extensions: _zbb and _zihintpause.

This patch introduces a compiler check to check if these extensions
are supported.
Additionally, it introduces the riscv/booting.txt file, which contains
information about the extensions that should be supported by the platform.

In the future, a feature will be introduced to check whether an extension
is supported at runtime.
However, this feature requires functionality for parsing device tree
source (DTS), which is not yet available.

Signed-off-by: Oleksii Kurochko <oleksii.kurochko@gmail.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agoautomation: introduce fixed randconfig for RISC-V
Oleksii Kurochko [Thu, 4 Apr 2024 09:59:15 +0000 (11:59 +0200)]
automation: introduce fixed randconfig for RISC-V

This patch introduces the anchor riscv-fixed-randconfig,
which includes all configurations that should be disabled for
randconfig builds.

Suggested-by: Stefano Stabellini <sstabellini@kernel.org>
Signed-off-by: Oleksii Kurochko <oleksii.kurochko@gmail.com>
Reviewed-by: Michal Orzel <michal.orzel@amd.com>
Acked-by: Stefano Stabellini <sstabellini@kernel.org>
13 months agodocs/misra: add 13.6 to rules.rst
Stefano Stabellini [Tue, 2 Apr 2024 23:21:41 +0000 (16:21 -0700)]
docs/misra: add 13.6 to rules.rst

As agreed during MISRA C meetings, add Rule 13.6 to rules.rst, with a
note about extending the rule to alignof and typeof. Also take the
opportunity to remove an older and out of order entry for the same rule.

Signed-off-by: Stefano Stabellini <stefano.stabellini@amd.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen/arm: ffa: support FFA_FEATURES
Jens Wiklander [Mon, 25 Mar 2024 09:39:04 +0000 (10:39 +0100)]
xen/arm: ffa: support FFA_FEATURES

Add support for the mandatory FF-A ABI function FFA_FEATURES.

Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>
Reviewed-by: Bertrand Marquis <bertrand.marquis@arm.com>
13 months agoxen/arm: ffa: separate rxtx buffer routines
Jens Wiklander [Mon, 25 Mar 2024 09:39:03 +0000 (10:39 +0100)]
xen/arm: ffa: separate rxtx buffer routines

Move rxtx buffer routines into a spearate file for easier navigation in
the source code.

Add ffa_rxtx_init(), ffa_rxtx_destroy(), and ffa_rxtx_domain_destroy() to
handle the ffa_rxtx internal things on initialization and teardown.

Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>
Reviewed-by: Bertrand Marquis <bertrand.marquis@arm.com>
13 months agoxen/arm: ffa: separate partition info get routines
Jens Wiklander [Mon, 25 Mar 2024 09:39:02 +0000 (10:39 +0100)]
xen/arm: ffa: separate partition info get routines

Move partition info get routines into a separate file for easier
navigation in the source code.

Add ffa_partinfo_init(), ffa_partinfo_domain_init(), and
ffa_partinfo_domain_destroy() to handle the ffa_partinfo internal things
on initialization and teardown.

Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>
Reviewed-by: Bertrand Marquis <bertrand.marquis@arm.com>
13 months agoxen/arm: ffa: separate memory sharing routines
Jens Wiklander [Mon, 25 Mar 2024 09:39:01 +0000 (10:39 +0100)]
xen/arm: ffa: separate memory sharing routines

Move memory sharing routines into a separate file for easier navigation
in the source code.

Add ffa_shm_domain_destroy() to isolate the ffa_shm things in
ffa_domain_teardown_continue().

Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>
[stefano: fix date in header]
Signed-off-by: Stefano Stabellini <stefano.stabellini@amd.com>
Reviewed-by: Bertrand Marquis <bertrand.marquis@arm.com>
13 months agoxen/arm: ffa: move common things to ffa_private.h
Jens Wiklander [Mon, 25 Mar 2024 09:39:00 +0000 (10:39 +0100)]
xen/arm: ffa: move common things to ffa_private.h

Prepare to separate ffa.c into modules by moving common things into the
new internal header file ffa_private.h.

Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>
Reviewed-by: Bertrand Marquis <bertrand.marquis@arm.com>
13 months agoxen/arm: ffa: rename functions to use ffa_ prefix
Jens Wiklander [Mon, 25 Mar 2024 09:38:59 +0000 (10:38 +0100)]
xen/arm: ffa: rename functions to use ffa_ prefix

Prepare to separate into modules by renaming functions that will need
new names when becoming non-static in the following commit.

Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>
Reviewed-by: Bertrand Marquis <bertrand.marquis@arm.com>
13 months agoxen/vm-event: address a violation of MISRA C:2012 Rule 16.3
Federico Serafini [Wed, 3 Apr 2024 07:38:41 +0000 (09:38 +0200)]
xen/vm-event: address a violation of MISRA C:2012 Rule 16.3

Add break statement to address a violation of MISRA C:2012 Rule 16.3
("An unconditional `break' statement shall terminate every
switch-clause ").

No functional change.

Signed-off-by: Federico Serafini <federico.serafini@bugseng.com>
Acked-by: Tamas K Lengyel <tamas@tklengyel.com>
13 months agoxen/evtchn: address a violation of MISRA C:2012 Rule 16.3
Federico Serafini [Wed, 3 Apr 2024 07:38:22 +0000 (09:38 +0200)]
xen/evtchn: address a violation of MISRA C:2012 Rule 16.3

Add break statement to address a violation of MISRA C:2012 Rule 16.3
("An unconditional `break' statement shall terminate every
switch-clause ").

No functional change.

Signed-off-by: Federico Serafini <federico.serafini@bugseng.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agoconsole: address a violation of MISRA C:2012 Rule 16.3
Federico Serafini [Wed, 3 Apr 2024 07:37:47 +0000 (09:37 +0200)]
console: address a violation of MISRA C:2012 Rule 16.3

Add break statement to address a violation of MISRA C:2012 Rule 16.3
("An unconditional `break' statement shall terminate every
switch-clause ").

No functional change.

Signed-off-by: Federico Serafini <federico.serafini@bugseng.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen/domctl: address a violation of MISRA C:2012 Rule 16.3
Federico Serafini [Wed, 3 Apr 2024 07:37:26 +0000 (09:37 +0200)]
xen/domctl: address a violation of MISRA C:2012 Rule 16.3

Add break statement to address a violation of MISRA C:2012 Rule 16.3
("An unconditional `break' statement shall terminate every
switch-clause ").

No functional change.

Signed-off-by: Federico Serafini <federico.serafini@bugseng.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agox86/amd: address violations of MISRA C Rule 20.7
Nicola Vetrini [Wed, 3 Apr 2024 07:37:00 +0000 (09:37 +0200)]
x86/amd: address violations of MISRA C Rule 20.7

MISRA C Rule 20.7 states: "Expressions resulting from the expansion
of macro parameters shall be enclosed in parentheses". Therefore, some
macro definitions should gain additional parentheses to ensure that all
current and future users will be safe with respect to expansions that
can possibly alter the semantics of the passed-in macro parameter.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen/mm: address violations of MISRA C Rule 20.7
Nicola Vetrini [Wed, 3 Apr 2024 07:36:30 +0000 (09:36 +0200)]
xen/mm: address violations of MISRA C Rule 20.7

MISRA C Rule 20.7 states: "Expressions resulting from the expansion
of macro parameters shall be enclosed in parentheses". Therefore, some
macro definitions should gain additional parentheses to ensure that all
current and future users will be safe with respect to expansions that
can possibly alter the semantics of the passed-in macro parameter.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agox86/hvm: address violations of MISRA C Rule 20.7
Nicola Vetrini [Wed, 3 Apr 2024 07:36:07 +0000 (09:36 +0200)]
x86/hvm: address violations of MISRA C Rule 20.7

MISRA C Rule 20.7 states: "Expressions resulting from the expansion
of macro parameters shall be enclosed in parentheses". Therefore, some
macro definitions should gain additional parentheses to ensure that all
current and future users will be safe with respect to expansions that
can possibly alter the semantics of the passed-in macro parameter.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agox86/vPMU: address violations of MISRA C Rule 20.7
Nicola Vetrini [Wed, 3 Apr 2024 07:35:37 +0000 (09:35 +0200)]
x86/vPMU: address violations of MISRA C Rule 20.7

MISRA C Rule 20.7 states: "Expressions resulting from the expansion
of macro parameters shall be enclosed in parentheses". Therefore, some
macro definitions should gain additional parentheses to ensure that all
current and future users will be safe with respect to expansions that
can possibly alter the semantics of the passed-in macro parameter.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agodrivers/char: mark extra reserved device memory in memory map
Marek Marczykowski-Górecki [Wed, 3 Apr 2024 07:34:22 +0000 (09:34 +0200)]
drivers/char: mark extra reserved device memory in memory map

The IOMMU driver checks if RMRR/IVMD are marked as reserved in memory
map. This should be true for addresses coming from the firmware, but
when extra pages used by Xen itself are included in the mapping, those
are taken from usable RAM used. Mark those pages as reserved too.

Not marking the pages as reserved didn't caused issues before due to
another a bug in IOMMU driver code, that was fixed in 83afa3135830
("amd-vi: fix IVMD memory type checks").

Failing to reserve memory will lead to panic in IOMMU setup code. And
not including the page in IOMMU mapping will lead to broken console (due
to IOMMU faults). The pages chosen by the XHCI console driver should
still be usable by the CPU though, and the console code already can deal
with too slow console by dropping characters (and console not printing
anything is a special case of "slow"). When reserving fails print an error
message showing which pages failed and who requested them. This should
be enough hint to find why XHCI console doesn't work.

Fixes: 3a1a7b809ffa "drivers/char: mark DMA buffers as reserved for the XHCI"
Signed-off-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agox86/spec-ctrl: Move __read_mostly data into __ro_after_init
Andrew Cooper [Thu, 28 Mar 2024 12:38:32 +0000 (12:38 +0000)]
x86/spec-ctrl: Move __read_mostly data into __ro_after_init

These variables predate the introduction of __ro_after_init, but all qualify.
Update them to be consistent with the rest of the file.

No functional change.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agox86/mm: Refine address alignment checks in modify_xen_mappings_lite()
Andrew Cooper [Tue, 19 Mar 2024 19:38:48 +0000 (19:38 +0000)]
x86/mm: Refine address alignment checks in modify_xen_mappings_lite()

Require 's' to be superpage aligned if given over a superpage mapping.  This
is expected to be the case in practice, and prevents v getting misaligned over
superpages during the processing loop.

Remove the requirement for 'e' to be page aligned.  All this is doing is
forcing work for the caller just to satisfy an assertion.

Reported-by: Jan Beulich <JBeulich@suse.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Roger Pau Monné <roger.pau@citrix.com>
13 months agotools/oxenstored: Re-format
Andrew Cooper [Mon, 26 Feb 2024 10:47:29 +0000 (10:47 +0000)]
tools/oxenstored: Re-format

Rerun make format.

Fixes: b6cf604207fd ("tools/oxenstored: Use Map instead of Hashtbl for quotas")
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Edwin Török <edwin.torok@cloud.com>
13 months agonestedsvm: Disable TscRateMSR
George Dunlap [Fri, 19 Jan 2024 09:27:54 +0000 (09:27 +0000)]
nestedsvm: Disable TscRateMSR

The primary purpose of TSC scaling, from our perspective, is to
maintain the fiction of an "invariant TSC" across migrates between
platforms with different clock speeds.

On AMD, the TscRateMSR CPUID bit is unconditionally enabled in the
"host cpuid", even if the hardware doesn't actually support it.
According to c/s fd14a1943c4 ("nestedsvm: Support TSC Rate MSR"),
testing showed that emulating TSC scaling in an L1 was more expensive
than emulating TSC scaling on an L0 (due to extra sets of vmexit /
vmenter).

However, the current implementation seems to be broken.

First of all, the final L2 scaling ratio should be a composition of
the L0 scaling ratio and the L1 scaling ratio; there's no indication
this is being done anywhere.

Secondly, it's not clear that the L1 tsc scaling ratio actually
affects the L0 tsc scaling ratio.  The stored value (ns_tscratio) is
used to affect the tsc *offset*, but doesn't seem to actually be
factored into d->hvm.tsc_scaling_ratio.  (Which shouldn't be
per-domain anyway, but per-vcpu.)  Having the *offset* scaled
according to the nested scaling without the actual RDTSC itself also
being scaled has got to produce inconsistent results.

For now, just disable the functionality entirely until we can
implement it properly:

- Don't set TSCRATEMSR in the HVM max CPUID policy

- Remove MSR_AMD64_TSC_RATIO emulation handling, so that the guest
  guests a #GP if it tries to access them (as it should when
  TSCRATEMSR is clear)

- Remove ns_tscratio from struct nestedhvm, and all code that touches
  it

Unfortunately this means ripping out the scaling calculation stuff as
well, since it's only used in the nested case; it's there in the git
tree if we need it for reference when we re-introduce it.

Signed-off-by: George Dunlap <george.dunlap@cloud.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agox86: Move SVM features exposed to guest into hvm_max_cpu_policy
George Dunlap [Mon, 26 Feb 2024 16:10:07 +0000 (16:10 +0000)]
x86: Move SVM features exposed to guest into hvm_max_cpu_policy

Currently (nested) SVM features we're willing to expose to the guest
are defined in calculate_host_policy, and stored in host_cpu_policy.
This is the wrong place for this; move it into
calculate_hvm_max_policy(), and store it in hvm_max_cpu_policy.

Signed-off-by: George Dunlap <george.dunlap@cloud.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen/efi: efibind: address violations of MISRA C Rule 20.7
Nicola Vetrini [Wed, 27 Mar 2024 08:20:29 +0000 (09:20 +0100)]
xen/efi: efibind: address violations of MISRA C Rule 20.7

MISRA C Rule 20.7 states: "Expressions resulting from the expansion
of macro parameters shall be enclosed in parentheses". Therefore, some
macro definitions should gain additional parentheses to ensure that all
current and future users will be safe with respect to expansions that
can possibly alter the semantics of the passed-in macro parameter.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen/elfnote: Specify ELF Notes are x86-specific
Jason Andryuk [Wed, 27 Mar 2024 08:14:17 +0000 (09:14 +0100)]
xen/elfnote: Specify ELF Notes are x86-specific

The Xen ELF Notes are only used with x86.  libelf's elf_xen_note_check()
exits early for ARM binaries with "ELF: Not bothering with notes on
ARM".

Add a comment to the top of elfnote.h specifying that Notes are only used
with x86 binaries.  This is to avoid adding disclaimers for individual
notes.

Signed-off-by: Jason Andryuk <jason.andryuk@amd.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agolibelf: Expand ELF note printing
Jason Andryuk [Wed, 27 Mar 2024 08:13:55 +0000 (09:13 +0100)]
libelf: Expand ELF note printing

The XEN_ELFNOTE_L1_MFN_VALID is an array of pairs of values, but it is
printed as:
(XEN) ELF: note: L1_MFN_VALID = 0

This is a limitation of only printing either string or numeric values.
Switch from the boolean to an enum which allows more flexibility in
printing the values.  Introduce ELFNOTE_NAME to only print the name
without a value like:
(XEN) ELF: note: L1_MFN_VALID

Details can optionally be printed for specific notes.

Signed-off-by: Jason Andryuk <jason.andryuk@amd.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agoRevert "xen/x86: bzImage parse kernel_alignment"
Jason Andryuk [Wed, 27 Mar 2024 08:13:34 +0000 (09:13 +0100)]
Revert "xen/x86: bzImage parse kernel_alignment"

A new ELF note will specify the alignment for a relocatable PVH kernel.
ELF notes are suitable for vmlinux and other ELF files, so this
Linux-specific bzImage parsing in unnecessary.

This reverts commit c44cac229067faeec8f49247d1cf281723ac2d40.

Signed-off-by: Jason Andryuk <jason.andryuk@amd.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agox86/public: address violations of MISRA C Rule 20.7
Nicola Vetrini [Wed, 27 Mar 2024 08:12:58 +0000 (09:12 +0100)]
x86/public: address violations of MISRA C Rule 20.7

MISRA C Rule 20.7 states: "Expressions resulting from the expansion
of macro parameters shall be enclosed in parentheses". Therefore, some
macro definitions should gain additional parentheses to ensure that all
current and future users will be safe with respect to expansions that
can possibly alter the semantics of the passed-in macro parameter.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agox86/hvm: address violations of Rule 20.7
Nicola Vetrini [Wed, 27 Mar 2024 08:12:30 +0000 (09:12 +0100)]
x86/hvm: address violations of Rule 20.7

MISRA C Rule 20.7 states: "Expressions resulting from the expansion
of macro parameters shall be enclosed in parentheses". Therefore, some
macro definitions should either gain additional parentheses or have
their expressions simplified to ensure that all current and future users
will be safe with respect to expansions that can possibly alter the
semantics of the passed-in macro parameter.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen/page_alloc: address violations of MISRA C Rule 20.7
Nicola Vetrini [Wed, 27 Mar 2024 08:12:06 +0000 (09:12 +0100)]
xen/page_alloc: address violations of MISRA C Rule 20.7

MISRA C Rule 20.7 states: "Expressions resulting from the expansion
of macro parameters shall be enclosed in parentheses". Therefore, some
macro definitions should either gain additional parentheses or have
their expressions simplified to ensure that all current and future users
will be safe with respect to expansions that can possibly alter the
semantics of the passed-in macro parameter.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen: address MISRA C Rule 20.7 violation in generated hypercall
Nicola Vetrini [Wed, 27 Mar 2024 08:11:03 +0000 (09:11 +0100)]
xen: address MISRA C Rule 20.7 violation in generated hypercall

MISRA C Rule 20.7 states: "Expressions resulting from the expansion
of macro parameters shall be enclosed in parentheses". Therefore, some
macro definitions should gain additional parentheses to ensure that all
current and future users will be safe with respect to expansions that
can possibly alter the semantics of the passed-in macro parameter.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agoxentrace: address violation of MISRA C Rule 20.7
Nicola Vetrini [Wed, 27 Mar 2024 08:10:23 +0000 (09:10 +0100)]
xentrace: address violation of MISRA C Rule 20.7

MISRA C Rule 20.7 states: "Expressions resulting from the expansion
of macro parameters shall be enclosed in parentheses". Therefore, some
macro definitions should gain additional parentheses to ensure that all
current and future users will be safe with respect to expansions that
can possibly alter the semantics of the passed-in macro parameter.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen/list: address violations of MISRA C Rule 20.7
Nicola Vetrini [Wed, 27 Mar 2024 08:09:50 +0000 (09:09 +0100)]
xen/list: address violations of MISRA C Rule 20.7

MISRA C Rule 20.7 states: "Expressions resulting from the expansion
of macro parameters shall be enclosed in parentheses". Therefore, some
macro definitions should gain additional parentheses to ensure that all
current and future users will be safe with respect to expansions that
can possibly alter the semantics of the passed-in macro parameter.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen/xsm: add parentheses to comply with MISRA C Rule 20.7
Nicola Vetrini [Mon, 25 Mar 2024 09:49:19 +0000 (10:49 +0100)]
xen/xsm: add parentheses to comply with MISRA C Rule 20.7

MISRA C Rule 20.7 states: "Expressions resulting from the expansion
of macro parameters shall be enclosed in parentheses". Therefore, some
macro definitions should gain additional parentheses to ensure that all
current and future users will be safe with respect to expansions that
can possibly alter the semantics of the passed-in macro parameter.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
Acked-by: Daniel P. Smith <dpsmith@apertussolutions.com>
13 months agoxen/console: add comment about external console lock helper
Roger Pau Monné [Mon, 25 Mar 2024 09:49:07 +0000 (10:49 +0100)]
xen/console: add comment about external console lock helper

The current console_lock_recursive_irqsave() implementation is not speculation
safe, however it's only used to prevent interleaved output.  Note this in the
function declaration in order for callers to be aware of the limitation.

No functional change.

Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agox86/vcpu: relax VCPUOP_initialise restriction for non-PV vCPUs
Roger Pau Monné [Mon, 25 Mar 2024 09:48:23 +0000 (10:48 +0100)]
x86/vcpu: relax VCPUOP_initialise restriction for non-PV vCPUs

There's no reason to force HVM guests to have a valid vcpu_info area when
initializing a vCPU, as the vCPU can also be brought online using the local
APIC, and on that path there's no requirement for vcpu_info to be setup ahead
of the bring up.  Note an HVM vCPU can operate normally without making use of
vcpu_info, and in fact does so when brought up via the local APIC.

Restrict the check against dummy_vcpu_info to only apply to PV guests.

Fixes: 192df6f9122d ('x86: allow HVM guests to use hypercalls to bring up vCPUs')
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
13 months agoxen/riscv: add minimal stuff to processor.h to build full Xen
Oleksii Kurochko [Mon, 25 Mar 2024 09:47:49 +0000 (10:47 +0100)]
xen/riscv: add minimal stuff to processor.h to build full Xen

The cpu_relax() function, introduced in this commit, is anticipated to
support _zihintpause by the CPU.

Signed-off-by: Oleksii Kurochko <oleksii.kurochko@gmail.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agoEFI: Introduce inline stub for efi_enabled on !X86 && !ARM
Shawn Anastasio [Mon, 25 Mar 2024 09:47:15 +0000 (10:47 +0100)]
EFI: Introduce inline stub for efi_enabled on !X86 && !ARM

On architectures with no EFI support, define an inline stub
implementation of efi_enabled in efi.h that always returns false.

Suggested-by: Jan Beulich <jbeulich@suse.com>
Signed-off-by: Shawn Anastasio <sanastasio@raptorengineering.com>
Acked-by: Julien Grall <jgrall@amazon.com>
13 months agoRevert "xen/page_alloc: introduce preserved page flags macro"
Julien Grall [Thu, 21 Mar 2024 16:02:25 +0000 (16:02 +0000)]
Revert "xen/page_alloc: introduce preserved page flags macro"

Osstest is reporting a crash:

Mar 21 12:00:29.533676 (XEN) pg[0] MFN 2211c5 c=0x2c00000000000000 o=0 v=0xe40000010007ffff t=0x24
Mar 21 12:00:42.829785 (XEN) Xen BUG at common/page_alloc.c:1033
Mar 21 12:00:42.829829 (XEN) ----[ Xen-4.19-unstable  x86_64  debug=y  Not tainted ]----
Mar 21 12:00:42.829857 (XEN) CPU:    12
Mar 21 12:00:42.841571 (XEN) RIP:    e008:[<ffff82d04022fe1f>] common/page_alloc.c#alloc_heap_pages+0x37f/0x6e2
Mar 21 12:00:42.841609 (XEN) RFLAGS: 0000000000010282   CONTEXT: hypervisor (d0v8)
Mar 21 12:00:42.853654 (XEN) rax: ffff83023e3ed06c   rbx: 000000000007ffff   rcx: 0000000000000028
Mar 21 12:00:42.853689 (XEN) rdx: ffff83047bec7fff   rsi: ffff83023e3ea3e8   rdi: ffff83023e3ea3e0
Mar 21 12:00:42.865657 (XEN) rbp: ffff83047bec7c10   rsp: ffff83047bec7b98   r8:  0000000000000000
Mar 21 12:00:42.877647 (XEN) r9:  0000000000000001   r10: 000000000000000c   r11: 0000000000000010
Mar 21 12:00:42.877682 (XEN) r12: 0000000000000001   r13: 0000000000000000   r14: ffff82e0044238a0
Mar 21 12:00:42.889652 (XEN) r15: 0000000000000000   cr0: 0000000080050033   cr4: 0000000000372660
Mar 21 12:00:42.901651 (XEN) cr3: 000000046fe34000   cr2: 00007fb72757610b
Mar 21 12:00:42.901685 (XEN) fsb: 00007fb726def380   gsb: ffff88801f200000   gss: 0000000000000000
Mar 21 12:00:42.913646 (XEN) ds: 0000   es: 0000   fs: 0000   gs: 0000   ss: e010   cs: e008
Mar 21 12:00:42.913680 (XEN) Xen code around <ffff82d04022fe1f> (common/page_alloc.c#alloc_heap_pages+0x37f/0x6e2):
Mar 21 12:00:42.925645 (XEN)  d1 1c 00 e8 ad dd 02 00 <0f> 0b 48 85 c9 79 36 0f 0b 41 89 cd 48 c7 47 f0
Mar 21 12:00:42.937649 (XEN) Xen stack trace from rsp=ffff83047bec7b98:
Mar 21 12:00:42.937683 (XEN)    0000000000000024 000000007bec7c20 0000000000000001 ffff83046ccda000
Mar 21 12:00:42.949653 (XEN)    ffff82e000000021 0000000000000016 0000000000000000 0000000000000000
Mar 21 12:00:42.949687 (XEN)    0000000000000000 0000000000000000 0000000000000028 0000000000000021
Mar 21 12:00:42.961652 (XEN)    ffff83046ccda000 0000000000000000 00007d2000000000 ffff83047bec7c48
Mar 21 12:00:42.961687 (XEN)    ffff82d0402302ff ffff83046ccda000 0000000000000100 0000000000000000
Mar 21 12:00:42.973655 (XEN)    ffff82d0405f0080 00007d2000000000 ffff83047bec7c80 ffff82d0402f626c
Mar 21 12:00:42.985656 (XEN)    ffff83046ccda000 ffff83046ccda640 0000000000000000 0000000000000000
Mar 21 12:00:42.985690 (XEN)    ffff83046ccda220 ffff83047bec7cb0 ffff82d0402f65a0 ffff83046ccda000
Mar 21 12:00:42.997662 (XEN)    0000000000000000 0000000000000000 0000000000000000 ffff83047bec7cc0
Mar 21 12:00:43.009660 (XEN)    ffff82d040311f8a ffff83047bec7ce0 ffff82d0402bd543 ffff83046ccda000
Mar 21 12:00:43.009695 (XEN)    ffff83047bec7dc8 ffff83047bec7d08 ffff82d04032c524 ffff83046ccda000
Mar 21 12:00:43.021653 (XEN)    ffff83047bec7dc8 0000000000000002 ffff83047bec7d58 ffff82d040206750
Mar 21 12:00:43.033642 (XEN)    0000000000000000 ffff82d040233fe5 ffff83047bec7d48 0000000000000000
Mar 21 12:00:43.033678 (XEN)    0000000000000002 00007fb72767f010 ffff82d0405e9120 0000000000000001
Mar 21 12:00:43.045654 (XEN)    ffff83047bec7e70 ffff82d040240728 0000000000000007 ffff83023e3b3000
Mar 21 12:00:43.045690 (XEN)    0000000000000246 ffff83023e2efa90 ffff83023e38e000 ffff83023e2efb40
Mar 21 12:00:43.057609 (XEN)    0000000000000007 ffff83023e3afb80 0000000000000206 ffff83047bec7dc0
Mar 21 12:00:43.069662 (XEN)    0000001600000001 000000000000ffff e75aaa8d0000000c ac0d6d864e487f62
Mar 21 12:00:43.069697 (XEN)    000000037fa48d76 0000000200000000 ffffffff000003ff 00000002ffffffff
Mar 21 12:00:43.081647 (XEN)    0000000000000000 00000000000001ff 0000000000000000 0000000000000000
Mar 21 12:00:43.093646 (XEN) Xen call trace:
Mar 21 12:00:43.093677 (XEN)    [<ffff82d04022fe1f>] R common/page_alloc.c#alloc_heap_pages+0x37f/0x6e2
Mar 21 12:00:43.093705 (XEN)    [<ffff82d0402302ff>] F alloc_domheap_pages+0x17d/0x1e4
Mar 21 12:00:43.105652 (XEN)    [<ffff82d0402f626c>] F hap_set_allocation+0x73/0x23c
Mar 21 12:00:43.105685 (XEN)    [<ffff82d0402f65a0>] F hap_enable+0x138/0x33c
Mar 21 12:00:43.117646 (XEN)    [<ffff82d040311f8a>] F paging_enable+0x2d/0x45
Mar 21 12:00:43.117679 (XEN)    [<ffff82d0402bd543>] F hvm_domain_initialise+0x185/0x428
Mar 21 12:00:43.129652 (XEN)    [<ffff82d04032c524>] F arch_domain_create+0x3e7/0x4c1
Mar 21 12:00:43.129687 (XEN)    [<ffff82d040206750>] F domain_create+0x4cc/0x7e2
Mar 21 12:00:43.141665 (XEN)    [<ffff82d040240728>] F do_domctl+0x1850/0x192d
Mar 21 12:00:43.141699 (XEN)    [<ffff82d04031a96a>] F pv_hypercall+0x617/0x6b5
Mar 21 12:00:43.153656 (XEN)    [<ffff82d0402012ca>] F lstar_enter+0x13a/0x140
Mar 21 12:00:43.153689 (XEN)
Mar 21 12:00:43.153711 (XEN)
Mar 21 12:00:43.153731 (XEN) ****************************************
Mar 21 12:00:43.165647 (XEN) Panic on CPU 12:
Mar 21 12:00:43.165678 (XEN) Xen BUG at common/page_alloc.c:1033
Mar 21 12:00:43.165703 (XEN) ****************************************
Mar 21 12:00:43.177633 (XEN)
Mar 21 12:00:43.177662 (XEN) Manual reset required ('noreboot' specified)

This reverts commit 75214d5e53f60a7b19e90ebdb090c20044a052ca.

Signed-off-by: Julien Grall <jgrall@amazon.com>
13 months agox86/cpuid: More AMD features
Andrew Cooper [Mon, 18 Mar 2024 18:07:05 +0000 (18:07 +0000)]
x86/cpuid: More AMD features

All of these are informational and require no further logic changes in Xen to
support.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agox86/p2m: Coding style cleanup
Andrew Cooper [Mon, 22 Oct 2018 13:53:26 +0000 (14:53 +0100)]
x86/p2m: Coding style cleanup

No functional change.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen/ppc: Ensure ELF sections' physical load addresses start at 0x0
Shawn Anastasio [Thu, 21 Mar 2024 08:49:20 +0000 (09:49 +0100)]
xen/ppc: Ensure ELF sections' physical load addresses start at 0x0

Some boot mechanisms, including the new linux file_load kexec systemcall
used by system firmware v2.10 on RaptorCS systems will try to honor the
physical address field of the ELF LOAD section header, which will fail
when the address is based off of XEN_VIRT_START (0xc000000000000000).

To ensure that the physical address of the LOAD section header starts at
0x0, move x86's DECL_SECTION macro to xen.lds.h and use it to declare
all sections.

Signed-off-by: Shawn Anastasio <sanastasio@raptorengineering.com>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
13 months agoAMD/IOMMU: drop remaining guest-IOMMU bits too
Jan Beulich [Thu, 21 Mar 2024 08:48:49 +0000 (09:48 +0100)]
AMD/IOMMU: drop remaining guest-IOMMU bits too

With a02174c6c885 ("amd/iommu: clean up unused guest iommu related
functions") having removed the sole place where d->g_iommu would be set
to non-NULL, guest_iommu_add_ppr_log() will unconditionally bail the
latest from its 2nd if(). With it dropped, all other stuff in the file
is unused, too. Delete iommu_guest.c altogether.

Further delete struct guest{_buffer,_dev_table,_iommu{,_msi}} as well as
struct mmio_reg for being unused with the unused g_iommu also dropped
from struct arch_iommu.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Roger Pau Monné <roger.pau@citrix.com>
13 months agox86/PoD: tie together P2M update and increment of entry count
Jan Beulich [Thu, 21 Mar 2024 08:48:10 +0000 (09:48 +0100)]
x86/PoD: tie together P2M update and increment of entry count

When not holding the PoD lock across the entire region covering P2M
update and stats update, the entry count - if to be incorrect at all -
should indicate too large a value in preference to a too small one, to
avoid functions bailing early when they find the count is zero. However,
instead of moving the increment ahead (and adjust back upon failure),
extend the PoD-locked region.

Fixes: 99af3cd40b6e ("x86/mm: Rework locking in the PoD layer")
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: George Dunlap <george.dunlap@cloud.com>
13 months agoxen/page_alloc: introduce page flag to stop buddy merging
Carlo Nonato [Thu, 21 Mar 2024 08:47:21 +0000 (09:47 +0100)]
xen/page_alloc: introduce page flag to stop buddy merging

Add a new PGC_no_buddy_merge flag that prevents the buddy algorithm in
free_heap_pages() from merging pages that have it set. As of now, only
PGC_static has this feature, but future work can extend it easier than
before.

Suggested-by: Jan Beulich <jbeulich@suse.com>
Signed-off-by: Carlo Nonato <carlo.nonato@minervasys.tech>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen/page_alloc: introduce preserved page flags macro
Carlo Nonato [Thu, 21 Mar 2024 08:46:42 +0000 (09:46 +0100)]
xen/page_alloc: introduce preserved page flags macro

PGC_static and PGC_extra needs to be preserved when assigning a page.
Define a new macro that groups those flags and use it instead of or'ing
every time.

To make preserved flags even more meaningful, they are kept also when
switching state in mark_page_free().

Signed-off-by: Carlo Nonato <carlo.nonato@minervasys.tech>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agotools/xl_parse: remove message for tsc mode string
Elliott Mitchell [Thu, 21 Mar 2024 08:46:13 +0000 (09:46 +0100)]
tools/xl_parse: remove message for tsc mode string

Normal behavior is to be silent.  Generating a message for the preferred
input can be mistaken for an error.  As such remove this message to match
other conditions.

Signed-off-by: Elliott Mitchell <ehem+xen@m5p.com>
Acked-by: Anthony PERARD <anthony.perard@citrix.com>
13 months agodocs/parse-support-md: Handle BulletList
Anthony PERARD [Tue, 19 Mar 2024 15:15:09 +0000 (15:15 +0000)]
docs/parse-support-md: Handle BulletList

Commit d638e304f13a introduced a bullet list, but parse-support-md
choke on it as it doesn't know what to do about it.

Introduce ri_BulletList() so that r_content() will find this new
function and call it instead of calling process_unknown().

Reported-by: Julien Grall <julien@xen.org>
Fixes: d638e304f13a ("SUPPORT.MD: Fix matrix generation after 43c416d0d819 and 77c39a53cf5b")
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Tested-by: Julien Grall <jgrall@amazon.com>
Acked-by: Julien Grall <jgrall@amazon.com>
13 months agoxen/vpci: Improve code generation in mask_write()
Andrew Cooper [Fri, 15 Mar 2024 11:31:33 +0000 (11:31 +0000)]
xen/vpci: Improve code generation in mask_write()

The use of __clear_bit() forces dmask to be spilled to the stack, and
interferes with the compiler heuristcs for some upcoming improvements to the
ffs() code generation.

First, shrink dmask to just the active vectors by making out the upper bits.
This replaces the "i < msi->vectors" part of the loop condition.

Next, use a simple while() loop with "clear bottom bit" expressed in plane C,
which affords the optimiser a far better understanding of what the loop is
doing.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Roger Pau Monné <roger.pau@citrix.com>
13 months agox86/boot: Support the watchdog on newer AMD systems
Andrew Cooper [Tue, 19 Mar 2024 14:11:12 +0000 (14:11 +0000)]
x86/boot: Support the watchdog on newer AMD systems

The MSRs used by setup_k7_watchdog() are architectural in 64bit.  The Unit
Select (0x76, cycles not in halt state) isn't, but it hasn't changed in 25
years, making this a trend likely to continue.

Drop the family check.  If the Unit Select does happen to change meaning in
the future, check_nmi_watchdog() will still notice the watchdog not operating
as expected.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agox86/boot: Fix setup_apic_nmi_watchdog() to fail more cleanly
Andrew Cooper [Tue, 19 Mar 2024 13:29:30 +0000 (13:29 +0000)]
x86/boot: Fix setup_apic_nmi_watchdog() to fail more cleanly

Right now, if the user requests the watchdog on the command line,
setup_apic_nmi_watchdog() will blindly assume that setting up the watchdog
worked.  Reuse nmi_perfctr_msr to identify when the watchdog has been
configured.

Rearrange setup_p6_watchdog() to not set nmi_perfctr_msr until the sanity
checks are complete.  Turn setup_p4_watchdog() into a void function, matching
the others.

If the watchdog isn't set up, inform the user and override to NMI_NONE, which
will prevent check_nmi_watchdog() from claiming that all CPUs are stuck.

e.g.:

  (XEN) alt table ffff82d040697c38 -> ffff82d0406a97f0
  (XEN) Failed to configure NMI watchdog
  (XEN) Brought up 512 CPUs
  (XEN) Scheduling granularity: cpu, 1 CPU per sched-resource

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agox86/boot: Improve the boot watchdog determination of stuck cpus
Andrew Cooper [Fri, 15 Mar 2024 19:32:39 +0000 (19:32 +0000)]
x86/boot: Improve the boot watchdog determination of stuck cpus

Right now, check_nmi_watchdog() has two processing loops over all online CPUs
using prev_nmi_count as storage.

Use a cpumask_t instead (1/32th as much initdata) and have wait_for_nmis()
make the determination of whether it is stuck, rather than having both
functions needing to agree on how many ticks mean stuck.

More importantly though, it means we can use the standard cpumask
infrastructure, including turning this:

  (XEN) Brought up 512 CPUs
  (XEN) Testing NMI watchdog on all CPUs: {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125,126,127,128,129,130,131,132,133,134,135,136,137,138,139,140,141,142,143,144,145,146,147,148,149,150,151,152,153,154,155,156,157,158,159,160,161,162,163,164,165,166,167,168,169,170,171,172,173,174,175,176,177,178,179,180,181,182,183,184,185,186,187,188,189,190,191,192,193,194,195,196,197,198,199,200,201,202,203,204,205,206,207,208,209,210,211,212,213,214,215,216,217,218,219,220,221,222,223,224,225,226,227,228,229,230,231,232,233,234,235,236,237,238,239,240,241,242,243,244,245,246,247,248,249,250,251,252,253,254,255,256,257,258,259,260,261,262,263,264,265,266,267,268,269,270,271,272,273,274,275,276,277,278,279,280,281,282,283,284,285,286,287,288,289,290,291,292,293,294,295,296,297,298,299,300,301,302,303,304,305,306,307,308,309,310,311,312,313,314,315,316,317,318,319,320,321,322,323,324,325,326,327,328,329,330,331,332,333,334,335,336,337,338,339,340,341,342,343,344,345,346,347,348,349,350,351,352,353,354,355,356,357,358,359,360,361,362,363,364,365,366,367,368,369,370,371,372,373,374,375,376,377,378,379,380,381,382,383,384,385,386,387,388,389,390,391,392,393,394,395,396,397,398,399,400,401,402,403,404,405,406,407,408,409,410,411,412,413,414,415,416,417,418,419,420,421,422,423,424,425,426,427,428,429,430,431,432,433,434,435,436,437,438,439,440,441,442,443,444,445,446,447,448,449,450,451,452,453,454,455,456,457,458,459,460,461,462,463,464,465,466,467,468,469,470,471,472,473,474,475,476,477,478,479,480,481,482,483,484,485,486,487,488,489,490,491,492,493,494,495,496,497,498,499,500,501,502,503,504,505,506,507,508,509,510,511} stuck

into the rather more manageable:

  (XEN) Brought up 512 CPUs
  (XEN) Testing NMI watchdog on all CPUs: {0-511} stuck

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Roger Pau Monné <roger.pau@citrix.com>
13 months agoxen/nospec: Include <xen/stdbool.h>
Michal Orzel [Tue, 19 Mar 2024 10:30:30 +0000 (11:30 +0100)]
xen/nospec: Include <xen/stdbool.h>

After introduction of lock_evaluate_nospec() using bool type, building
Xen on Arm with UBSAN enabled fails:

In file included from ./include/xen/spinlock.h:4,
                 from common/ubsan/ubsan.c:13:
./include/xen/nospec.h:79:22: error: unknown type name ‘bool’
   79 | static always_inline bool lock_evaluate_nospec(bool condition)

There is no issue on x86, as xen/stdbool.h is included somewhere along
the asm/nospec.h path, which is not the case for other architectures.

Fixes: 7ef0084418e1 ("x86/spinlock: introduce support for blocking speculation into critical regions")
Signed-off-by: Michal Orzel <michal.orzel@amd.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen/wait: address violations of MISRA C Rule 20.7
Nicola Vetrini [Tue, 19 Mar 2024 10:30:01 +0000 (11:30 +0100)]
xen/wait: address violations of MISRA C Rule 20.7

MISRA C Rule 20.7 states: "Expressions resulting from the expansion
of macro parameters shall be enclosed in parentheses". Therefore, some
macro definitions should gain additional parentheses to ensure that all
current and future users will be safe with respect to expansions that
can possibly alter the semantics of the passed-in macro parameter.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen/notifier: address violations of MISRA C Rule 20.7
Nicola Vetrini [Tue, 19 Mar 2024 10:29:12 +0000 (11:29 +0100)]
xen/notifier: address violations of MISRA C Rule 20.7

MISRA C Rule 20.7 states: "Expressions resulting from the expansion
of macro parameters shall be enclosed in parentheses". Therefore, some
macro definitions should gain additional parentheses to ensure that all
current and future users will be safe with respect to expansions that
can possibly alter the semantics of the passed-in macro parameter.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen/arm: smmu: address violations of MISRA C Rule 20.7
Nicola Vetrini [Tue, 19 Mar 2024 10:28:44 +0000 (11:28 +0100)]
xen/arm: smmu: address violations of MISRA C Rule 20.7

MISRA C Rule 20.7 states: "Expressions resulting from the expansion
of macro parameters shall be enclosed in parentheses". Therefore, some
macro definitions should gain additional parentheses to ensure that all
current and future users will be safe with respect to expansions that
can possibly alter the semantics of the passed-in macro parameter.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
13 months agoEFI: address violations of MISRA C Rule 20.7
Nicola Vetrini [Tue, 19 Mar 2024 10:28:28 +0000 (11:28 +0100)]
EFI: address violations of MISRA C Rule 20.7

MISRA C Rule 20.7 states: "Expressions resulting from the expansion
of macro parameters shall be enclosed in parentheses". Therefore, some
macro definitions should gain additional parentheses to ensure that all
current and future users will be safe with respect to expansions that
can possibly alter the semantics of the passed-in macro parameter.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen/device_tree: address violations of MISRA C Rule 20.7
Nicola Vetrini [Tue, 19 Mar 2024 10:27:58 +0000 (11:27 +0100)]
xen/device_tree: address violations of MISRA C Rule 20.7

MISRA C Rule 20.7 states: "Expressions resulting from the expansion
of macro parameters shall be enclosed in parentheses". Therefore, some
macro definitions should gain additional parentheses to ensure that all
current and future users will be safe with respect to expansions that
can possibly alter the semantics of the passed-in macro parameter.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
13 months agox86/cpufeature: add parentheses to comply with Rule 20.7
Nicola Vetrini [Tue, 19 Mar 2024 10:27:29 +0000 (11:27 +0100)]
x86/cpufeature: add parentheses to comply with Rule 20.7

MISRA C Rule 20.7 states: "Expressions resulting from the expansion
of macro parameters shall be enclosed in parentheses". Therefore, some
macro definitions should gain additional parentheses to ensure that all
current and future users will be safe with respect to expansions that
can possibly alter the semantics of the passed-in macro parameter.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agoamd/iommu: clean up unused guest iommu related functions
Nicola Vetrini [Tue, 19 Mar 2024 10:27:07 +0000 (11:27 +0100)]
amd/iommu: clean up unused guest iommu related functions

Delete unused functions from 'iommu_guest.c'.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen/spinlock: make struct lock_profile rspinlock_t aware
Juergen Gross [Tue, 19 Mar 2024 10:26:34 +0000 (11:26 +0100)]
xen/spinlock: make struct lock_profile rspinlock_t aware

Struct lock_profile contains a pointer to the spinlock it is associated
with. Prepare support of differing spinlock_t and rspinlock_t types by
adding a type indicator of the pointer. Use the highest bit of the
block_cnt member for this indicator in order to not grow the struct
while hurting only the slow path with slightly less performant code.
Note that this requires a cast when printing the value in order to be
format compliant.

Signed-off-by: Juergen Gross <jgross@suse.com>
Acked-by: Alejandro Vallejo <alejandro.vallejo@cloud.com>
Acked-by: Julien Grall <jgrall@amazon.com>
13 months agoxen/spinlock: add rspin_[un]lock_irq[save|restore]()
Juergen Gross [Tue, 19 Mar 2024 10:25:59 +0000 (11:25 +0100)]
xen/spinlock: add rspin_[un]lock_irq[save|restore]()

Instead of special casing rspin_lock_irqsave() and
rspin_unlock_irqrestore() for the console lock, add those functions
to spinlock handling and use them where needed.

Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen/spinlock: rename recursive lock functions
Juergen Gross [Tue, 19 Mar 2024 10:25:07 +0000 (11:25 +0100)]
xen/spinlock: rename recursive lock functions

Rename the recursive spin_lock() functions by replacing the trailing
"_recursive" with a leading "r".

Switch the parameter to be a pointer to rspinlock_t.

Suggested-by: Jan Beulich <jbeulich@suse.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Acked-by: Julien Grall <jgrall@amazon.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen/spinlock: introduce new type for recursive spinlocks
Juergen Gross [Tue, 19 Mar 2024 10:23:53 +0000 (11:23 +0100)]
xen/spinlock: introduce new type for recursive spinlocks

Introduce a new type "rspinlock_t" to be used for recursive spinlocks.

For now it is only an alias of spinlock_t, so both types can still be
used for recursive spinlocks. This will be changed later, though.

Switch all recursive spinlocks to the new type.

Define the initializer helpers and use them where appropriate.

Signed-off-by: Juergen Gross <jgross@suse.com>
Acked-by: Julien Grall <jgrall@amazon.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agoIOMMU: store name for extra reserved device memory
Marek Marczykowski-Górecki [Tue, 19 Mar 2024 10:23:06 +0000 (11:23 +0100)]
IOMMU: store name for extra reserved device memory

It will be useful for error reporting in a subsequent patch.

Signed-off-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agox86/mm: use block_lock_speculation() in _mm_write_lock()
Jan Beulich [Mon, 18 Mar 2024 12:53:37 +0000 (13:53 +0100)]
x86/mm: use block_lock_speculation() in _mm_write_lock()

I can only guess that using block_speculation() there was a leftover
from, earlier on, SPECULATIVE_HARDEN_LOCK depending on
SPECULATIVE_HARDEN_BRANCH.

Fixes: 197ecd838a2a ("locking: attempt to ensure lock wrappers are always inline")
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
13 months agoxen/sched: address violations of MISRA C Rule 20.7
Nicola Vetrini [Mon, 18 Mar 2024 12:53:07 +0000 (13:53 +0100)]
xen/sched: address violations of MISRA C Rule 20.7

MISRA C Rule 20.7 states: "Expressions resulting from the expansion
of macro parameters shall be enclosed in parentheses". Therefore, some
macro definitions should gain additional parentheses to ensure that all
current and future users will be safe with respect to expansions that
can possibly alter the semantics of the passed-in macro parameter.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Acked-by: George Dunlap <george.dunlap@cloud.com>
13 months agoautomation/eclair: allow parameter name "unused"
Federico Serafini [Mon, 18 Mar 2024 12:52:53 +0000 (13:52 +0100)]
automation/eclair: allow parameter name "unused"

Update ECLAIR configuration of MISRA C:2012 Rule 8.3 to deviate
violations involving parameter name "unused" (with an optional
numeric suffix): it makes explicit the intention of not using such
parameter within the function.

Signed-off-by: Federico Serafini <federico.serafini@bugseng.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
13 months agoMAINTAINERS: remove myself from various positions
Wei Liu [Mon, 18 Mar 2024 12:52:35 +0000 (13:52 +0100)]
MAINTAINERS: remove myself from various positions

I have not been active in the Xen community for some time. Update the
MAINTAINERS file to reflect reality.

Signed-off-by: Wei Liu <wl@xen.org>
Acked-by: Stefano Stabellini <sstabellini@kernel.org>
13 months agoMAINTAINERS: drop AMD IOMMU and Intel VT-d sections
Jan Beulich [Mon, 18 Mar 2024 12:52:09 +0000 (13:52 +0100)]
MAINTAINERS: drop AMD IOMMU and Intel VT-d sections

We'd like to thank Kevin, so far the VT-d maintainer, for his past
contributions, while at the same time we'd like to reflect reality as it
has been for quite some time. Have VT-d maintainership (and for symmetry
also AMD IOMMU's) fall back to the maintainers of general x86. This also
brings things in sync with how they are for Arm.

While touching the X86 section also append two so far missing slashes.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Kevin Tian <kevin.tian@intel.com>
13 months agoperfc: add pseudo-keyword fallthrough
Federico Serafini [Mon, 18 Mar 2024 12:51:17 +0000 (13:51 +0100)]
perfc: add pseudo-keyword fallthrough

Add pseudo-keyword fallthrough to make explicit the intention of the
code and meet requirements to deviate MISRA C:2012 Rule 16.3
("An unconditional `break' statement shall terminate every
switch-clause").

No functional change.

Signed-off-by: Federico Serafini <federico.serafini@bugseng.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agox86/svmdebug: Print np, sev and sev_es vmcb bits
Vaishali Thakkar [Mon, 18 Mar 2024 12:50:33 +0000 (13:50 +0100)]
x86/svmdebug: Print np, sev and sev_es vmcb bits

Currently only raw _np_ctrl is being printed. It can
be informational to know about which particular bits
are enabled. So, this commit adds the bit-by-bit decode
for np, sev and sev_es bits.

Note that while, only np is enabled in certain scenarios
at the moment, work for enabling sev and sev_es is in
progress. And it'll be useful to have this information as
part of svmdebug.

Signed-off-by: Vaishali Thakkar <vaishali.thakkar@vates.tech>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agox86/svm: Drop the suffix _guest from vmcb bit
Vaishali Thakkar [Mon, 18 Mar 2024 12:49:46 +0000 (13:49 +0100)]
x86/svm: Drop the suffix _guest from vmcb bit

The suffix _guest is redundant for asid bit. Drop it
to avoid adding extra code volume.

While we're here, replace 0/1 with false/true and use
VMCB accessors instead of open coding.

Suggested-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Vaishali Thakkar <vaishali.thakkar@vates.tech>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agoSUPPORT.MD: Fix matrix generation after 43c416d0d819 and 77c39a53cf5b
Julien Grall [Thu, 14 Mar 2024 17:39:29 +0000 (17:39 +0000)]
SUPPORT.MD: Fix matrix generation after 43c416d0d819 and 77c39a53cf5b

The script docs/support-matrix-generate throw the following error on the
latest staging.

MDCCSR_EL0
DBGDTR_EL0
DBGDTRTX_EL0
DBGDTRRX_EL0
DBGDSCRINT
DBGDTRTXINT
DBGDTRRXINT
^ cannot parse status codeblock line:
MDCCSR_EL0
 ? at docs/parse-support-md line 172, <F> chunk 1.

With the current indentation, it looks like the script thinks this is
a status code block.

Solve the issue by switching to a bullet list.

Fixes: 43c416d0d819 ("xen/arm: arm64: Add emulation of Debug Data Transfer Registers")
Fixes: 77c39a53cf5b ("xen/arm: arm32: Add emulation of Debug Data Transfer Registers")
Signed-off-by: Julien Grall <jgrall@amazon.com>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
13 months agoxen/arm: arm32: Add emulation of Debug Data Transfer Registers
Ayan Kumar Halder [Thu, 7 Mar 2024 12:39:43 +0000 (12:39 +0000)]
xen/arm: arm32: Add emulation of Debug Data Transfer Registers

When user enables HVC_DCC config option in Linux, it invokes access to debug
transfer register (i.e. DBGDTRTXINT). As this register is not emulated, Xen
injects an undefined exception to the guest and Linux crashes.

To prevent this crash, introduce a partial emulation of DBGDTR[TR]XINT (these
registers share the same encoding) as RAZ/WI and DBGDSCRINT as TXfull.

Refer ARM DDI 0487J.a ID042523, G8.3.19, DBGDTRTXint:
"If TXfull is set to 1, set DTRTX to UNKNOWN".

As a pre-requisite, DBGOSLSR should be emulated in the same way as its AArch64
variant (i.e. OSLSR_EL1). This is to ensure that DBGOSLSR.OSLK is 0, which
allows us to skip the emulation of DBGDSCREXT (TXfull is treated as UNK/SBZP)
and focus on DBGDSCRINT. DBGOSLSR.OSLM[1] is set to 1 to mantain consistency
with Arm64.

Take the opportunity to fix the minimum EL for DBGDSCRINT, which should be 0.

Signed-off-by: Ayan Kumar Halder <ayan.kumar.halder@amd.com>
Signed-off-by: Michal Orzel <michal.orzel@amd.com>
Acked-by: Julien Grall <jgrall@amazon.com>
13 months agoxen/arm: arm64: Add emulation of Debug Data Transfer Registers
Michal Orzel [Thu, 7 Mar 2024 12:39:42 +0000 (12:39 +0000)]
xen/arm: arm64: Add emulation of Debug Data Transfer Registers

Currently, if user enables HVC_DCC config option in Linux, it invokes access
to debug data transfer registers (i.e. DBGDTRTX_EL0 on arm64, DBGDTRTXINT on
arm32). As these registers are not emulated, Xen injects an undefined
exception to the guest and Linux crashes.

To prevent this crash, introduce a partial emulation of DBGDTR[TR]X_EL0
(these registers share the same encoding) as RAZ/WI and MDCCSR_EL0 as TXfull.

Refer ARM DDI 0487J.a ID042523, D19.3.8, DBGDTRTX_EL0
"If TXfull is set to 1, set DTRRX and DTRTX to UNKNOWN".

Thus, any OS is expected to read MDCCSR_EL0 and check for TXfull before
using DBGDTRTX_EL0. Linux does it via hvc_dcc_init() ---> hvc_dcc_check(),
and returns -ENODEV in case TXfull bit is still set after writing a test
character. This way we prevent the guest from making use of HVC DCC as a
console.

Signed-off-by: Michal Orzel <michal.orzel@amd.com>
Signed-off-by: Ayan Kumar Halder <ayan.kumar.halder@amd.com>
Acked-by: Julien Grall <jgrall@amazon.com>
13 months agoxen/arm: Introduce CONFIG_PARTIAL_EMULATION and "partial-emulation" cmd option
Ayan Kumar Halder [Thu, 7 Mar 2024 12:39:41 +0000 (12:39 +0000)]
xen/arm: Introduce CONFIG_PARTIAL_EMULATION and "partial-emulation" cmd option

There can be situations when the registers cannot be emulated to their full
functionality. This can be due to the complexity involved. In such cases, one
can emulate those registers as RAZ/WI for example. We call them as partial
emulation.

Some registers are non-optional and as such there is nothing preventing an OS
from accessing them.
Instead of injecting undefined exception (which might result in crashing a
guest), one may want to prefer a partial emulation to let the guest running
(in some cases accepting the fact that it might result in unwanted behavior).

A suitable example of this (as seen in subsequent patches) is emulation of
DBGDTRTX_EL0 (on Arm64) and DBGDTRTXINT(on Arm32). These non-optional
registers can be emulated as RAZ/WI and they can be enclosed within
CONFIG_PARTIAL_EMULATION.

Further, "partial-emulation" command line option allows us to
enable/disable partial emulation at run time. While CONFIG_PARTIAL_EMULATION
enables support for partial emulation at compile time (i.e. adds code for
partial emulation), this option may be enabled or disabled by Yocto or other
build systems. However if the build system turns this option on, users
can use scripts like Imagebuilder to generate uboot-script which will append
"partial-emulation=true" to xen command line to turn on the partial emulation.
Thus, it helps to avoid rebuilding xen.

By default, "CONFIG_PARTIAL_EMULATION=y" and "partial-emulation=false".
This is done so that Xen supports partial emulation. However, customers are
fully aware when they enable partial emulation. It's important to note that
enabling such support might result in unwanted/non-spec compliant behavior.

Added a note in SUPPORT.md to clarify the security support for partial
emulation.

Signed-off-by: Ayan Kumar Halder <ayan.kumar.halder@amd.com>
Signed-off-by: Michal Orzel <michal.orzel@amd.com>
Acked-by: Julien Grall <jgrall@amazon.com>
13 months agoxen/spinlock: remove misra rule 21.1 violations
Juergen Gross [Thu, 14 Mar 2024 12:55:55 +0000 (13:55 +0100)]
xen/spinlock: remove misra rule 21.1 violations

In xen spinlock code there are several violations of MISRA rule 21.1
(identifiers starting with "__" or "_[A-Z]").

Fix them by using trailing underscores instead.

Signed-off-by: Juergen Gross <jgross@suse.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen: avoid grep fodder define and undef
Federico Serafini [Thu, 14 Mar 2024 12:55:27 +0000 (13:55 +0100)]
xen: avoid grep fodder define and undef

Place an "#if 0" before grep fodder #define-s and remove
the #undef-s.

This addresses violations of MISRA C:2012 Rules 20.5 ("#undef should
not be used") and 5.5 ("Identifiers shall be distinct from macro
names").

No functional change.

Signed-off-by: Federico Serafini <federico.serafini@bugseng.com>
Acked-by: Andrew Cooper <andrew.cooper3@citrix.com>
13 months agoMAINTAINERS: Remove myself from several subsystems
Paul Durrant [Thu, 14 Mar 2024 12:55:11 +0000 (13:55 +0100)]
MAINTAINERS: Remove myself from several subsystems

I am not as actively involved with the Xen project as I once was so I need
to relinquish some of my maintainer responsibilities: IOMMU and I/O
emulation.

NOTE: Since I am sole maintainer for I/O EMULATION (IOREQ) and
      X86 I/O EMULATION, these sections are removed.

Signed-off-by: Paul Durrant <pdurrant@amazon.com>
Acked-by: Stefano Stabellini <sstabellini@kernel.org>
13 months agox86/svm: Drop the _enabled suffix from vmcb bits
Vaishali Thakkar [Thu, 14 Mar 2024 12:54:43 +0000 (13:54 +0100)]
x86/svm: Drop the _enabled suffix from vmcb bits

The suffix is redundant for np/sev/sev-es bits. Drop it
to avoid adding extra code volume.

Also, while we're here, drop the double negations in one
of the instances of _np bit, replace 0/1 with false/true
in the use cases of _np and use VMCB accessors instead
of open coding.

Suggested-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Vaishali Thakkar <vaishali.thakkar@vates.tech>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
13 months agoxen/compiler: deviate the inline macro for MISRA C Rule 20.4
Nicola Vetrini [Thu, 14 Mar 2024 12:54:14 +0000 (13:54 +0100)]
xen/compiler: deviate the inline macro for MISRA C Rule 20.4

Rule 20.4 states: "A macro shall not be defined with the same name
as a keyword".

Defining this macro with the same name as the inline keyword
allows for additionally checking that out-of-lined static inline
functions end up in the correct section while minimizing churn and
has a positive impact on the overall safety. See [1] for additional
context on the motivation of this deviation.

No functional change.

[1] https://lore.kernel.org/xen-devel/adaa6d55-266d-4df8-8967-9340080d17e4@citrix.com/

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
13 months agoxen/lz4: address violations of MISRA C:2012 Directive 4.10
Maria Celeste Cesario [Thu, 14 Mar 2024 12:52:46 +0000 (13:52 +0100)]
xen/lz4: address violations of MISRA C:2012 Directive 4.10

Add inclusion guard to address violations of
MISRA C:2012 Directive 4.10 ("Precautions shall be taken in order to
prevent the contents of a header file being included more than once").
Mechanical change.

Signed-off-by: Maria Celeste Cesario <maria.celeste.cesario@bugseng.com>
Signed-off-by: Simone Ballarin <simone.ballarin@bugseng.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agox86/mtrr: address violations of MISRA C:2012 Directive 4.10
Maria Celeste Cesario [Thu, 14 Mar 2024 12:52:32 +0000 (13:52 +0100)]
x86/mtrr: address violations of MISRA C:2012 Directive 4.10

Add inclusion guard to address violations of
MISRA C:2012 Directive 4.10 ("Precautions shall be taken in order to
prevent the contents of a header file being included more than once").
Mechanical change.

Signed-off-by: Maria Celeste Cesario <maria.celeste.cesario@bugseng.com>
Signed-off-by: Simone Ballarin <simone.ballarin@bugseng.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agorestrict concept of pIRQ to x86
Jan Beulich [Thu, 14 Mar 2024 12:51:19 +0000 (13:51 +0100)]
restrict concept of pIRQ to x86

... by way of a new arch-selectable Kconfig control.

Note that some smaller pieces of code are left without #ifdef, to keep
things better readable. Hence items like ECS_PIRQ, nr_static_irqs, or
domain_pirq_to_irq() remain uniformly.

As to XEN_DOMCTL_irq_permission - this, despite having a uniformly
available wrapper in libxc, is unused on Arm: libxl bypasses those
calls, and the Python and OCamL bindings have no users at all.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Julien Grall <jgrall@amazon.com>
13 months agoautomation/eclair_analysis: deviate certain macros for Rule 20.12
Nicola Vetrini [Thu, 15 Feb 2024 13:06:17 +0000 (14:06 +0100)]
automation/eclair_analysis: deviate certain macros for Rule 20.12

Certain macros are allowed to violate the Rule, since their meaning and
intended use is well-known to all Xen developers.

Variadic macros that rely on the GCC extension for removing a trailing
comma when token pasting the variable argument are similarly
well-understood and therefore allowed.

No functional change.

Signed-off-by: Nicola Vetrini <nicola.vetrini@bugseng.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
13 months agox86: protect conditional lock taking from speculative execution
Roger Pau Monné [Mon, 4 Mar 2024 15:24:21 +0000 (16:24 +0100)]
x86: protect conditional lock taking from speculative execution

Conditionally taken locks that use the pattern:

if ( lock )
    spin_lock(...);

Need an else branch in order to issue an speculation barrier in the else case,
just like it's done in case the lock needs to be acquired.

eval_nospec() could be used on the condition itself, but that would result in a
double barrier on the branch where the lock is taken.

Introduce a new pair of helpers, {gfn,spin}_lock_if() that can be used to
conditionally take a lock in a speculation safe way.

This is part of XSA-453 / CVE-2024-2193

Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agox86/mm: add speculation barriers to open coded locks
Roger Pau Monné [Mon, 4 Mar 2024 17:08:48 +0000 (18:08 +0100)]
x86/mm: add speculation barriers to open coded locks

Add a speculation barrier to the clearly identified open-coded lock taking
functions.

Note that the memory sharing page_lock() replacement (_page_lock()) is left
as-is, as the code is experimental and not security supported.

This is part of XSA-453 / CVE-2024-2193

Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agolocking: attempt to ensure lock wrappers are always inline
Roger Pau Monné [Mon, 4 Mar 2024 13:29:36 +0000 (14:29 +0100)]
locking: attempt to ensure lock wrappers are always inline

In order to prevent the locking speculation barriers from being inside of
`call`ed functions that could be speculatively bypassed.

While there also add an extra locking barrier to _mm_write_lock() in the branch
taken when the lock is already held.

Note some functions are switched to use the unsafe variants (without speculation
barrier) of the locking primitives, but a speculation barrier is always added
to the exposed public lock wrapping helper.  That's the case with
sched_spin_lock_double() or pcidevs_lock() for example.

This is part of XSA-453 / CVE-2024-2193

Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agopercpu-rwlock: introduce support for blocking speculation into critical regions
Roger Pau Monné [Tue, 13 Feb 2024 16:57:38 +0000 (17:57 +0100)]
percpu-rwlock: introduce support for blocking speculation into critical regions

Add direct calls to block_lock_speculation() where required in order to prevent
speculation into the lock protected critical regions.  Also convert
_percpu_read_lock() from inline to always_inline.

Note that _percpu_write_lock() has been modified the use the non speculation
safe of the locking primites, as a speculation is added unconditionally by the
calling wrapper.

This is part of XSA-453 / CVE-2024-2193

Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agorwlock: introduce support for blocking speculation into critical regions
Roger Pau Monné [Tue, 13 Feb 2024 15:08:52 +0000 (16:08 +0100)]
rwlock: introduce support for blocking speculation into critical regions

Introduce inline wrappers as required and add direct calls to
block_lock_speculation() in order to prevent speculation into the rwlock
protected critical regions.

Note the rwlock primitives are adjusted to use the non speculation safe variants
of the spinlock handlers, as a speculation barrier is added in the rwlock
calling wrappers.

trylock variants are protected by using lock_evaluate_nospec().

This is part of XSA-453 / CVE-2024-2193

Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agox86/spinlock: introduce support for blocking speculation into critical regions
Roger Pau Monné [Tue, 13 Feb 2024 12:08:05 +0000 (13:08 +0100)]
x86/spinlock: introduce support for blocking speculation into critical regions

Introduce a new Kconfig option to block speculation into lock protected
critical regions.  The Kconfig option is enabled by default, but the mitigation
won't be engaged unless it's explicitly enabled in the command line using
`spec-ctrl=lock-harden`.

Convert the spinlock acquire macros into always-inline functions, and introduce
a speculation barrier after the lock has been taken.  Note the speculation
barrier is not placed inside the implementation of the spin lock functions, as
to prevent speculation from falling through the call to the lock functions
resulting in the barrier also being skipped.

trylock variants are protected using a construct akin to the existing
evaluate_nospec().

This patch only implements the speculation barrier for x86.

Note spin locks are the only locking primitive taken care in this change,
further locking primitives will be adjusted by separate changes.

This is part of XSA-453 / CVE-2024-2193

Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
13 months agoxen: Swap order of actions in the FREE*() macros
Andrew Cooper [Fri, 2 Feb 2024 00:39:42 +0000 (00:39 +0000)]
xen: Swap order of actions in the FREE*() macros

Wherever possible, it is a good idea to NULL out the visible reference to an
object prior to freeing it.  The FREE*() macros already collect together both
parts, making it easy to adjust.

This has a marginal code generation improvement, as some of the calls to the
free() function can be tailcall optimised.

No functional change.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
13 months agox86/spec-ctrl: Mitigation Register File Data Sampling
Andrew Cooper [Thu, 22 Jun 2023 22:32:19 +0000 (23:32 +0100)]
x86/spec-ctrl: Mitigation Register File Data Sampling

RFDS affects Atom cores, also branded E-cores, between the Goldmont and
Gracemont microarchitectures.  This includes Alder Lake and Raptor Lake hybrid
clien systems which have a mix of Gracemont and other types of cores.

Two new bits have been defined; RFDS_CLEAR to indicate VERW has more side
effets, and RFDS_NO to incidate that the system is unaffected.  Plenty of
unaffected CPUs won't be getting RFDS_NO retrofitted in microcode, so we
synthesise it.  Alder Lake and Raptor Lake Xeon-E's are unaffected due to
their platform configuration, and we must use the Hybrid CPUID bit to
distinguish them from their non-Xeon counterparts.

Like MD_CLEAR and FB_CLEAR, RFDS_CLEAR needs OR-ing across a resource pool, so
set it in the max policies and reflect the host setting in default.

This is part of XSA-452 / CVE-2023-28746.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>