rtm@csail.mit.edu reports:
> nfsd4_decode_bitmap4() will write beyond bmval[bmlen-1] if the RPC
> directs it to do so. This can cause nfsd4_decode_state_protect4_a()
> to write client-supplied data beyond the end of
> nfsd4_exchange_id.spo_must_allow[] when called by
> nfsd4_decode_exchange_id().
Rewrite the loops so nfsd4_decode_bitmap() cannot iterate beyond
@bmlen.
Reported by: rtm@csail.mit.edu
Fixes: d1c263a031 ("NFSD: Replace READ* macros in nfsd4_decode_fattr()")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
There's a perfectly fine prototype in the asm/setup.h header. Use it.
No functional changes.
Signed-off-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Brijesh Singh <brijesh.singh@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: https://lkml.kernel.org/r/20211110220731.2396491-8-brijesh.singh@amd.com
Carve it out so that it is abstracted out of the main boot path. All
other encrypted guest-relevant processing should be placed in there.
No functional changes.
Signed-off-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Brijesh Singh <brijesh.singh@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: https://lkml.kernel.org/r/20211110220731.2396491-7-brijesh.singh@amd.com
Currently set_pcie_thunderbolt() open-codes pci_find_vsec_capability().
Refactor the former to use the latter. No functional change intended.
Link: https://lore.kernel.org/r/20211115112902.24033-1-andriy.shevchenko@linux.intel.com
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Krzysztof Wilczyński <kw@linux.com>
Remove all the defines of masks and bit positions for the GHCB MSR
protocol and use comments instead which correspond directly to the spec
so that following those can be a lot easier and straightforward with the
spec opened in parallel to the code.
Aligh vertically while at it.
No functional changes.
Signed-off-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Brijesh Singh <brijesh.singh@amd.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: https://lkml.kernel.org/r/20211110220731.2396491-6-brijesh.singh@amd.com
In order to make sure new function cast mismatches are not introduced
in the kernel (to avoid tripping CFI checking), the kernel should be
globally built with -Wcast-function-type.
Link: https://github.com/KSPP/linux/issues/20
Co-developed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Each thread executing in an enclave is associated with a Thread Control
Structure (TCS). The SGX test enclave contains two hardcoded TCS, thus
supporting two threads in the enclave.
Add a test to ensure it is possible to enter enclave at both entrypoints.
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lkml.kernel.org/r/7be151a57b4c7959a2364753b995e0006efa3da1.1636997631.git.reinette.chatre@intel.com
Each thread executing in an enclave is associated with a Thread Control
Structure (TCS). The test enclave contains two hardcoded TCS. Each TCS
contains meta-data used by the hardware to save and restore thread specific
information when entering/exiting the enclave.
The two TCS structures within the test enclave share their SSA (State Save
Area) resulting in the threads clobbering each other's data. Fix this by
providing each TCS their own SSA area.
Additionally, there is an 8K stack space and its address is
computed from the enclave entry point which is correctly done for
TCS #1 that starts on the first address inside the enclave but
results in out of bounds memory when entering as TCS #2. Split 8K
stack space into two separate pages with offset symbol between to ensure
the current enclave entry calculation can continue to be used for both
threads.
While using the enclave with multiple threads requires these fixes the
impact is not apparent because every test up to this point enters the
enclave from the first TCS.
More detail about the stack fix:
-------------------------------
Before this change the test enclave (test_encl) looks as follows:
.tcs (2 pages):
(page 1) TCS #1
(page 2) TCS #2
.text (1 page)
One page of code
.data (5 pages)
(page 1) encl_buffer
(page 2) encl_buffer
(page 3) SSA
(page 4 and 5) STACK
encl_stack:
As shown above there is a symbol, encl_stack, that points to the end of the
.data segment (pointing to the end of page 5 in .data) which is also the
end of the enclave.
The enclave entry code computes the stack address by adding encl_stack to
the pointer to the TCS that entered the enclave. When entering at TCS #1
the stack is computed correctly but when entering at TCS #2 the stack
pointer would point to one page beyond the end of the enclave and a #PF
would result when TCS #2 attempts to enter the enclave.
The fix involves moving the encl_stack symbol between the two stack pages.
Doing so enables the stack address computation in the entry code to compute
the correct stack address for each TCS.
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lkml.kernel.org/r/a49dc0d85401db788a0a3f0d795e848abf3b1f44.1636997631.git.reinette.chatre@intel.com
The Enclave Page Cache Map (EPCM) is a secure structure used by the
processor to track the contents of the enclave page cache. The EPCM
contains permissions with which enclave pages can be accessed. SGX
support allows EPCM and PTE page permissions to differ - as long as
the PTE permissions do not exceed the EPCM permissions.
Add a test that:
(1) Creates an SGX enclave page with writable EPCM permission.
(2) Changes the PTE permission on the page to read-only. This should
be permitted because the permission does not exceed the EPCM
permission.
(3) Attempts a write to the page. This should generate a page fault
(#PF) because of the read-only PTE even though the EPCM
permissions allow the page to be written to.
This introduces the first test of SGX exception handling. In this test
the issue that caused the exception (PTE page permissions) can be fixed
from outside the enclave and after doing so it is possible to re-enter
enclave at original entrypoint with ERESUME.
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lkml.kernel.org/r/3bcc73a4b9fe8780bdb40571805e7ced59e01df7.1636997631.git.reinette.chatre@intel.com
SGX selftests prepares a data structure outside of the enclave with
the type of and data for the operation that needs to be run within
the enclave. At this time only two complementary operations are supported
by the enclave: copying a value from outside the enclave into a default
buffer within the enclave and reading a value from the enclave's default
buffer into a variable accessible outside the enclave.
In preparation for more operations supported by the enclave the names of the
current enclave operations are changed to more accurately reflect the
operations and more easily distinguish it from future operations:
* The enums ENCL_OP_PUT and ENCL_OP_GET are renamed to ENCL_OP_PUT_TO_BUFFER
and ENCL_OP_GET_FROM_BUFFER respectively.
* The structs encl_op_put and encl_op_get are renamed to encl_op_put_to_buf
and encl_op_get_from_buf respectively.
* The enclave functions do_encl_op_put and do_encl_op_get are renamed to
do_encl_op_put_to_buf and do_encl_op_get_from_buf respectively.
No functional changes.
Suggested-by: Jarkko Sakkinen <jarkko@kernel.org>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Acked-by: Jarkko Sakkinen <jarkko@kernel.org>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lkml.kernel.org/r/023fda047c787cf330b88ed9337705edae6a0078.1636997631.git.reinette.chatre@intel.com
To add more operations to the test enclave, the protocol needs to allow
to have operations with varying parameters. Create a separate parameter
struct for each existing operation, with the shared parameters in struct
encl_op_header.
[reinette: rebased to apply on top of oversubscription test series]
Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lkml.kernel.org/r/f9a4a8c436b538003b8ebddaa66083992053cef1.1636997631.git.reinette.chatre@intel.com
Add a variation of the unclobbered_vdso test.
In the new test, create a heap for the test enclave, which has the same
size as all available Enclave Page Cache (EPC) pages in the system. This
will guarantee that all test_encl.elf pages *and* SGX Enclave Control
Structure (SECS) have been swapped out by the page reclaimer during the
load time.
This test will trigger both the page reclaimer and the page fault handler.
The page reclaimer triggered, while the heap is being created during the
load time. The page fault handler is triggered for all the required pages,
while the test case is executing.
Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lkml.kernel.org/r/41f7c508eea79a3198b5014d7691903be08f9ff1.1636997631.git.reinette.chatre@intel.com
Create the test enclave inside each TEST_F(), instead of FIXTURE_SETUP(),
so that the heap size can be defined per test.
Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lkml.kernel.org/r/70ca264535d2ca0dc8dcaf2281e7d6965f8d4a24.1636997631.git.reinette.chatre@intel.com
Introduce setup_test_encl() so that the enclave creation can be moved to
TEST_F()'s. This is required for a reclaimer test where the heap size needs
to be set large enough to triger the page reclaimer.
Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lkml.kernel.org/r/bee0ca867a95828a569c1ba2a8e443a44047dc71.1636997631.git.reinette.chatre@intel.com
Logging is always a compromise between clarity and detail. The main use
case for dumping VMA's is when FIXTURE_SETUP() fails, and is less important
for enclaves that do initialize correctly. Therefore, print the segments
and /proc/self/maps only in the error case.
Finally, if a single test ever creates multiple enclaves, the amount of
log lines would become enormous.
Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lkml.kernel.org/r/23cef0ae1de3a8a74cbfbbe74eca48ca3f300fde.1636997631.git.reinette.chatre@intel.com
Create a heap for the test enclave, which is allocated from /dev/null,
and left unmeasured. This is beneficial by its own because it verifies
that an enclave built from multiple choices, works properly. If LSM
hooks are added for SGX some day, a multi source enclave has higher
probability to trigger bugs on access control checks.
The immediate need comes from the need to implement page reclaim tests.
In order to trigger the page reclaimer, one can just set the size of
the heap to high enough.
Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lkml.kernel.org/r/e070c5f23578c29608051cab879b1d276963a27a.1636997631.git.reinette.chatre@intel.com
For a heap makes sense to leave its contents "unmeasured" in the SGX
enclave build process, meaning that they won't contribute to the
cryptographic signature (a RSA-3072 signed SHA56 hash) of the enclave.
Enclaves are signed blobs where the signature is calculated both from
page data and also from "structural properties" of the pages. For
instance a page offset of *every* page added to the enclave is hashed.
For data, this is optional, not least because hashing a page has a
significant contribution to the enclave load time. Thus, where there is
no reason to hash, do not. The SGX ioctl interface supports this with
SGX_PAGE_MEASURE flag. Only when the flag is *set*, data is measured.
Add seg->measure boolean flag to struct encl_segment. Only when the
flag is set, include the segment data to the signature (represented
by SIGSTRUCT architectural structure).
Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lkml.kernel.org/r/625b6fe28fed76275e9238ec4e15ec3c0d87de81.1636997631.git.reinette.chatre@intel.com
Define source per segment so that enclave pages can be added from different
sources, e.g. anonymous VMA for zero pages. In other words, add 'src' field
to struct encl_segment, and assign it to 'encl->src' for pages inherited
from the enclave binary.
Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lkml.kernel.org/r/7850709c3089fe20e4bcecb8295ba87c54cc2b4a.1636997631.git.reinette.chatre@intel.com
The enclave binary (test_encl.elf) is built with only three sections (tcs,
text, and data) as controlled by its custom linker script.
If gcc is built with "--enable-linker-build-id" (this appears to be a
common configuration even if it is by default off) then gcc
will pass "--build-id" to the linker that will prompt it (the linker) to
write unique bits identifying the linked file to a ".note.gnu.build-id"
section.
The section ".note.gnu.build-id" does not exist in the test enclave
resulting in the following warning emitted by the linker:
/usr/bin/ld: warning: .note.gnu.build-id section discarded, --build-id
ignored
The test enclave does not use the build id within the binary so fix the
warning by passing a build id of "none" to the linker that will disable the
setting from any earlier "--build-id" options and thus disable the attempt
to write the build id to a ".note.gnu.build-id" section that does not
exist.
Link: https://lore.kernel.org/linux-sgx/20191017030340.18301-2-sean.j.christopherson@intel.com/
Suggested-by: Cedric Xing <cedric.xing@intel.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lkml.kernel.org/r/ca0f8a81fc1e78af9bdbc6a88e0f9c37d82e53f2.1636997631.git.reinette.chatre@intel.com
To ensure clean unload of the machine driver, components and topology, do
the unregister before we free IPC and debugfs.
It is a possibility that part of the unregister we would have IPC
communication with the firmware.
Suggested-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20211102094756.9317-1-peter.ujfalusi@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Considering the current transition of the GPIO subsystem, remove all
dependencies of the legacy GPIO interface (linux/gpio.h and linux
/of_gpio.h) and replace it with the descriptor-based GPIO approach.
Signed-off-by: Maíra Canal <maira.canal@usp.br>
Link: https://lore.kernel.org/r/YXWo/9o7ye9a11aR@fedora
Signed-off-by: Mark Brown <broonie@kernel.org>
This patch removes unused included header files and moves others into
cs_dsp.h to ensure that types referenced in the header file are properly
described to prevent compiler warnings.
Signed-off-by: Simon Trimmer <simont@opensource.cirrus.com>
Acked-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20211115120215.56824-1-simont@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
This adds an ALSA control so that the slow-start audio ramp feature
can be disabled. This is useful for high-definition audio applications.
The register field is unusual in that it is a 3-bit field with only
two valid values, 000=off and 111=on.
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20211101101006.13092-1-rf@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The machine driver is a generic machine driver for SOF with nau8825
codec w or w/o speaker additionally. Depending on the SOC
HDMI, DMIC, Bluetooth offload support are added dynamically.
Only add information related to SOF since the machine driver was
only tested with SOF.
There are currently 4 i2s machine variants of ADL.
This supports the headphone NUA8825(SSP0) alone or with smart or dumb
speakers.
Board 2,3,4 use SSP2 for Bluetooth offload support except board 1.
Board 1 : NAU8825 + RT1019P(SSP2)
Board 2 : NAU8825 + MAX98373(SSP1)
Board 3 : NAU8825 + MAX98360A(SSP1)
Board 4 : NAU8825
Signed-off-by: David Lin <CTLIN0@nuvoton.com>
Co-developed-by: Mac Chiang <mac.chiang@intel.com>
Signed-off-by: Mac Chiang <mac.chiang@intel.com>
Acked-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20211109133808.8729-1-mac.chiang@intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
'hphpa_on' is know to be false, so this is just dead code that should be
removed.
Suggested-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Reviewed-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/57a89cc31eb2312addd3c77896d7df8206aef138.1635967035.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Mark Brown <broonie@kernel.org>
Replace the old .txt binding with a new schema binding.
At the same time, some of the descriptions are updated to make them
clearer, fix errors, or just make them fit better into the style
of schema binding.
The cirrus,hs-bias-ramp-rate property was missing from the old .txt
binding and has been added to the yaml.
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20211028140902.11786-4-rf@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
This series provides a way to use constant dailink numbers for different
devices. So that we don't need to renumber them in topologies.
Some patches with different purpose are sent together in this series
since they are dependent.
Bard Liao (8):
ASoC: intel: sof_sdw: return the original error number
ASoC: intel: sof_sdw: rename be_index/link_id to link_index
ASoC: intel: sof_sdw: Use a fixed DAI link id for AMP
ASoC: intel: sof_sdw: move DMIC link id overwrite to
create_sdw_dailink
ASoC: intel: sof_sdw: remove SOF_RT715_DAI_ID_FIX quirk
ASoC: intel: sof_sdw: remove sof_sdw_mic_codec_mockup_init
ASoC: intel: sof_sdw: remove get_next_be_id
ASoC: intel: sof_sdw: add link adr order check
Pierre-Louis Bossart (2):
ASoC: Intel: sof_sdw: fix jack detection on HP Spectre x360
convertible
ASoC: Intel: sof_sdw: add SKU for Dell Latitude 9520
sound/soc/intel/boards/sof_sdw.c | 152 ++++++++++----------
sound/soc/intel/boards/sof_sdw_common.h | 7 +-
sound/soc/intel/boards/sof_sdw_rt715.c | 7 -
sound/soc/intel/boards/sof_sdw_rt715_sdca.c | 7 -
4 files changed, 85 insertions(+), 88 deletions(-)
--
2.17.1
This is the continuation of a previous series [1] where
- patch 1/4 is removed in favor of using pin switch
This will be posted independently of tfa989x support,
since it mainly require changes to sound/soc/qcom/common.c
and device DTS.
- patch 2/4 is already merged
so here are reworked patch 3/4 (bindings fixed and example added)
and patch 4/4 unchanged.
[1] https://patchwork.kernel.org/project/alsa-devel/cover/20211024085840.1536438-1-vincent.knecht@mailoo.org/
Vincent Knecht (2):
ASoC: dt-bindings: nxp, tfa989x: Add rcv-gpios property for tfa9897
ASoC: codecs: tfa989x: Add support for tfa9897 optional rcv-gpios
.../bindings/sound/nxp,tfa989x.yaml | 41 +++++++++++++++++++
sound/soc/codecs/tfa989x.c | 20 ++++++++-
2 files changed, 60 insertions(+), 1 deletion(-)
--
2.31.1
SGX EPC pages do not have a "struct page" associated with them so the
pfn_valid() sanity check fails and results in a warning message to
the console.
Add an additional check to skip the warning if the address of the error
is in an SGX EPC page.
Signed-off-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Tested-by: Reinette Chatre <reinette.chatre@intel.com>
Link: https://lkml.kernel.org/r/20211026220050.697075-8-tony.luck@intel.com
SGX reserved memory does not appear in the standard address maps.
Add hook to call into the SGX code to check if an address is located
in SGX memory.
There are other challenges in injecting errors into SGX. Update the
documentation with a sequence of operations to inject.
Signed-off-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Tested-by: Reinette Chatre <reinette.chatre@intel.com>
Link: https://lkml.kernel.org/r/20211026220050.697075-7-tony.luck@intel.com
Add a call inside memory_failure() to call the arch specific code
to check if the address is an SGX EPC page and handle it.
Note the SGX EPC pages do not have a "struct page" entry, so the hook
goes in at the same point as the device mapping hook.
Pull the call to acquire the mutex earlier so the SGX errors are also
protected.
Make set_mce_nospec() skip SGX pages when trying to adjust
the 1:1 map.
Signed-off-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Reviewed-by: Naoya Horiguchi <naoya.horiguchi@nec.com>
Tested-by: Reinette Chatre <reinette.chatre@intel.com>
Link: https://lkml.kernel.org/r/20211026220050.697075-6-tony.luck@intel.com
Provide a recovery function sgx_memory_failure(). If the poison was
consumed synchronously then send a SIGBUS. Note that the virtual
address of the access is not included with the SIGBUS as is the case
for poison outside of SGX enclaves. This doesn't matter as addresses
of code/data inside an enclave is of little to no use to code executing
outside the (now dead) enclave.
Poison found in a free page results in the page being moved from the
free list to the per-node poison page list.
Signed-off-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Tested-by: Reinette Chatre <reinette.chatre@intel.com>
Link: https://lkml.kernel.org/r/20211026220050.697075-5-tony.luck@intel.com
A memory controller patrol scrubber can report poison in a page
that isn't currently being used.
Add "poison" field in the sgx_epc_page that can be set for an
sgx_epc_page. Check for it:
1) When sanitizing dirty pages
2) When freeing epc pages
Poison is a new field separated from flags to avoid having to make all
updates to flags atomic, or integrate poison state changes into some
other locking scheme to protect flags (Currently just sgx_reclaimer_lock
which protects the SGX_EPC_PAGE_RECLAIMER_TRACKED bit in page->flags).
In both cases place the poisoned page on a per-node list of poisoned
epc pages to make sure it will not be reallocated.
Signed-off-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Tested-by: Reinette Chatre <reinette.chatre@intel.com>
Link: https://lkml.kernel.org/r/20211026220050.697075-4-tony.luck@intel.com
X86 machine check architecture reports a physical address when there
is a memory error. Handling that error requires a method to determine
whether the physical address reported is in any of the areas reserved
for EPC pages by BIOS.
SGX EPC pages do not have Linux "struct page" associated with them.
Keep track of the mapping from ranges of EPC pages to the sections
that contain them using an xarray. N.B. adds CONFIG_XARRAY_MULTI to
the SGX dependecies. So "select" that in arch/x86/Kconfig for X86/SGX.
Create a function arch_is_platform_page() that simply reports whether an
address is an EPC page for use elsewhere in the kernel. The ACPI error
injection code needs this function and is typically built as a module,
so export it.
Note that arch_is_platform_page() will be slower than other similar
"what type is this page" functions that can simply check bits in the
"struct page". If there is some future performance critical user of
this function it may need to be implemented in a more efficient way.
Note also that the current implementation of xarray allocates a few
hundred kilobytes for this usage on a system with 4GB of SGX EPC memory
configured. This isn't ideal, but worth it for the code simplicity.
Signed-off-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Tested-by: Reinette Chatre <reinette.chatre@intel.com>
Link: https://lkml.kernel.org/r/20211026220050.697075-3-tony.luck@intel.com
SGX EPC pages go through the following life cycle:
DIRTY ---> FREE ---> IN-USE --\
^ |
\-----------------/
Recovery action for poison for a DIRTY or FREE page is simple. Just
make sure never to allocate the page. IN-USE pages need some extra
handling.
Add a new flag bit SGX_EPC_PAGE_IS_FREE that is set when a page
is added to a free list and cleared when the page is allocated.
Notes:
1) These transitions are made while holding the node->lock so that
future code that checks the flags while holding the node->lock
can be sure that if the SGX_EPC_PAGE_IS_FREE bit is set, then the
page is on the free list.
2) Initially while the pages are on the dirty list the
SGX_EPC_PAGE_IS_FREE bit is cleared.
Signed-off-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Tested-by: Reinette Chatre <reinette.chatre@intel.com>
Link: https://lkml.kernel.org/r/20211026220050.697075-2-tony.luck@intel.com
A test of the form:
while true; do modprobe -r cxl_pmem; modprobe cxl_pmem; done
May lead to a crash signature of the form:
BUG: unable to handle page fault for address: ffffffffc0660030
#PF: supervisor instruction fetch in kernel mode
#PF: error_code(0x0010) - not-present page
[..]
Workqueue: cxl_pmem 0xffffffffc0660030
RIP: 0010:0xffffffffc0660030
Code: Unable to access opcode bytes at RIP 0xffffffffc0660006.
[..]
Call Trace:
? process_one_work+0x4ec/0x9c0
? pwq_dec_nr_in_flight+0x100/0x100
? rwlock_bug.part.0+0x60/0x60
? worker_thread+0x2eb/0x700
In that report the 0xffffffffc0660030 address corresponds to the former
function address of cxl_nvb_update_state() from a previous load of the
module, not the current address. Fix that by arranging for ->state_work
in the 'struct cxl_nvdimm_bridge' object to be reinitialized on cxl_pmem
module reload.
Details:
Recall that CXL subsystem wants to link a CXL memory expander device to
an NVDIMM sub-hierarchy when both a persistent memory range has been
registered by the CXL platform driver (cxl_acpi) *and* when that CXL
memory expander has published persistent memory capacity (Get Partition
Info). To this end the cxl_nvdimm_bridge driver arranges to rescan the
CXL bus when either of those conditions change. The helper
bus_rescan_devices() can not be called underneath the device_lock() for
any device on that bus, so the cxl_nvdimm_bridge driver uses a workqueue
for the rescan.
Typically a driver allocates driver data to hold a 'struct work_struct'
for a driven device, but for a workqueue that may run after ->remove()
returns, driver data will have been freed. The 'struct
cxl_nvdimm_bridge' object holds the state and work_struct directly.
Unfortunately it was only arranging for that infrastructure to be
initialized once per device creation rather than the necessary once per
workqueue (cxl_pmem_wq) creation.
Introduce is_cxl_nvdimm_bridge() and cxl_nvdimm_bridge_reset() in
support of invalidating stale references to a recently destroyed
cxl_pmem_wq.
Cc: <stable@vger.kernel.org>
Fixes: 8fdcb1704f ("cxl/pmem: Add initial infrastructure for pmem support")
Reported-by: Vishal Verma <vishal.l.verma@intel.com>
Tested-by: Vishal Verma <vishal.l.verma@intel.com>
Link: https://lore.kernel.org/r/163665474585.3505991.8397182770066720755.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
During NUMA init, CXL memory defined in the SRAT Memory Affinity
subtable may be assigned to a NUMA node. Since there is no
requirement that the SRAT be comprehensive for CXL memory another
mechanism is needed to assign NUMA nodes to CXL memory not identified
in the SRAT.
Use the CXL Fixed Memory Window Structure (CFMWS) of the ACPI CXL
Early Discovery Table (CEDT) to find all CXL memory ranges.
Create a NUMA node for each CFMWS that is not already assigned to
a NUMA node. Add a memblk attaching its host physical address
range to the node.
Note that these ranges may not actually map any memory at boot time.
They may describe persistent capacity or may be present to enable
hot-plug.
Consumers can use phys_to_target_node() to discover the NUMA node.
Signed-off-by: Alison Schofield <alison.schofield@intel.com>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/163553711933.2509508.2203471175679990.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Now that cxl_acpi has been converted to use the core ACPI CEDT sub-table
parser, update cxl_test to inject CFMWS and CHBS data directly into
cxl_acpi's handlers.
Cc: Alison Schofield <alison.schofield@intel.com>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/163553711363.2509508.17428994087868269952.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
The cxl_acpi driver originally open-coded its table parsing since the
ACPI subtable helpers were marked __init and only used in early NUMA
initialization. Now that those helpers have been exported for driver
usage replace the open-coded solution with the common one.
Cc: Alison Schofield <alison.schofield@intel.com>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/163553710810.2509508.14686373989517930921.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
In preparation for drivers reusing the core table parsing
infrastructure, arrange for handlers to take a context argument. This
allows driver table parsing to wrap ACPI table entries in
driver-specific data.
The first consumer of this infrastructure is the CEDT parsing that
happens in the cxl_acpi driver, add a conditional
(CONFIG_ACPI_TABLE_LIB=y) export of a acpi_table_parse_cedt() helper for
this case.
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Len Brown <lenb@kernel.org>
Tested-by: Alison Schofield <alison.schofield@intel.com>
Reviewed-by: Alison Schofield <alison.schofield@intel.com>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/163553710257.2509508.14310494417463866020.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
The CEDT adds yet one more unique subtable header type where the length
is a 16-bit value. Extend the subtable helpers to detect this scenario.
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Len Brown <lenb@kernel.org>
Tested-by: Alison Schofield <alison.schofield@intel.com>
Reviewed-by: Alison Schofield <alison.schofield@intel.com>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/163553709742.2509508.5177761945441327574.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
The NFIT driver and now the CXL ACPI driver have both open-coded ACPI
table parsing. Before another instance is added arrange for the core
ACPI sub-table parsing to be optionally available to drivers via the
CONFIG_ACPI_TABLE_LIB symbol. If no drivers select the symbol then the
infrastructure reverts back to being tagged __init via the
__init_or_acpilib annotation.
For now, only tag the core sub-table routines and data that the CEDT parsing in
the cxl_acpi driver would want to reuse, a CEDT parsing helper is added
in a later change.
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Len Brown <lenb@kernel.org>
Cc: Alison Schofield <alison.schofield@intel.com>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/163553709227.2509508.8215196520233473814.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Add mocked health information for cxl_test memdevs. This allows
cxl-cli's 'list' command to display the canned health_info fields.
Cc: Dan Williams <dan.j.williams@intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Vishal Verma <vishal.l.verma@intel.com>
Link: https://lore.kernel.org/r/20211018051251.2289112-1-vishal.l.verma@intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
It turns out that the usb example of specifying the subsystem namespace
at build time is not preferred. The rationale for that preference has
become more apparent as CXL patches with plain EXPORT_SYMBOL_GPL beg the
question, "why would any code other than CXL care about this symbol?".
Make the namespace explicit.
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Link: https://lore.kernel.org/r/163676356810.3618264.601632777702192938.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
The 'struct cxl_mem' object actually represents the state of a CXL
device within the driver. Comments indicating that 'struct cxl_mem' is a
device itself are incorrect. It is data layered on top of a CXL Memory
Expander class device. Rename it 'struct cxl_dev_state'. The 'struct'
cxl_memdev' structure represents a Linux CXL memory device object, and
it uses services and information provided by 'struct cxl_dev_state'.
Update the structure name, function names, and the kdocs to reflect the
real uses of this structure.
Some helper functions that were previously prefixed "cxl_mem_" are
renamed to just "cxl_".
Acked-by: Ben Widawsky <ben.widawsky@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Link: https://lore.kernel.org/r/20211102202901.3675568-3-ira.weiny@intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
__cxl_mem_mbox_send_cmd() no longer exists. Remove the reference.
Acked-by: Ben Widawsky <ben.widawsky@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Link: https://lore.kernel.org/r/20211102202901.3675568-2-ira.weiny@intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
There is a potential race between queue_work() returning and the
queued-work running that could result in put_device() running before
get_device(). Introduce the cxl_nvdimm_bridge_state_work() helper that
takes the reference unconditionally, but drops it if no new work was
queued, to keep the references balanced.
Fixes: 8fdcb1704f ("cxl/pmem: Add initial infrastructure for pmem support")
Cc: <stable@vger.kernel.org>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Ben Widawsky <ben.widawsky@intel.com>
Link: https://lore.kernel.org/r/163553734757.2509761.3305231863616785470.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
