With the cleanup of the misleading clock value to avoid extra
calculations to convert between link_bit_rate and clock, use
one standard "clock" field for the c20 pll which works with
crtc_state->port_clock field.
Cc: Clint Taylor <clinton.a.taylor@intel.com>
Cc: Mika Kahola <mika.kahola@intel.com>
Signed-off-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Reviewed-by: Mika Kahola <mika.kahola@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231207221025.2032207-4-radhakrishna.sripada@intel.com
The field link_bit_rate serves as the actual clock value for the C20
pll_state structure. Remove the misleading clock field. The subsequent
patch would rename the link_bit_rate as the clock field.
Cc: Clint Taylor <clinton.a.taylor@intel.com>
Cc: Mika Kahola <mika.kahola@intel.com>
Signed-off-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Reviewed-by: Mika Kahola <mika.kahola@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231207221025.2032207-3-radhakrishna.sripada@intel.com
In C20 pll_state link_bit_rate and clock fields are bit redundant. Since
many of the helpers assume the clock values, which are different from
link_bit_rate for dp2.0, convert the helpers to use the numbers that
are compatible with link_bit_rate.
Currently link_bit_rate is compatible with crtc_state->port_clock. The
function intel_c20pll_calc_port_clock returns the number which is
compatible with crtc_state->port_clock. In order to avoid extra
conversions b/ween clock and link_bit_rate, remove "clock" field from the
C20 pll_state and then rename "link_bit_rate" as "clock".
While at it rely on crtc_state->port_clock during C20 Pll programming.
Cc: Clint Taylor <clinton.a.taylor@intel.com>
Cc: Mika Kahola <mika.kahola@intel.com>
Signed-off-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Reviewed-by: Mika Kahola <mika.kahola@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231207221025.2032207-2-radhakrishna.sripada@intel.com
On board the sdmmc0 interface is wired to a SD Card socket.
According with mmc-controller bindings, the mmc-ddr-3_3v property
is used for eMMC devices to enable high-speed DDR mode (3.3V I/O).
Remove the mmc-ddr-3_3v property from sdmmc0 node.
Signed-off-by: Mihai Sain <mihai.sain@microchip.com>
Link: https://lore.kernel.org/r/20231204072537.2991-1-mihai.sain@microchip.com
Signed-off-by: Claudiu Beznea <claudiu.beznea@tuxon.dev>
- Fix an annoying goof where the NX hugepage test prints out garbage
instead of the magic token needed to run the text.
- Fix build errors when a header is delete/moved due to a missing flag
in the Makefile.
- Detect if KVM bugged/killed a selftest's VM and print out a helpful
message instead of complaining that a random ioctl() failed.
- Annotate the guest printf/assert helpers with __printf(), and fix the
various bugs that were lurking due to lack of said annotation.
-----BEGIN PGP SIGNATURE-----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=Ur6p
-----END PGP SIGNATURE-----
Merge tag 'kvm-x86-selftests-6.7-rcN' of https://github.com/kvm-x86/linux into HEAD
KVM selftests fixes for 6.8 merge window:
- Fix an annoying goof where the NX hugepage test prints out garbage
instead of the magic token needed to run the text.
- Fix build errors when a header is delete/moved due to a missing flag
in the Makefile.
- Detect if KVM bugged/killed a selftest's VM and print out a helpful
message instead of complaining that a random ioctl() failed.
- Annotate the guest printf/assert helpers with __printf(), and fix the
various bugs that were lurking due to lack of said annotation.
A small subset of these was included in 6.7-rc as well.
With migration disabled, one function becomes unused:
virt/kvm/guest_memfd.c:262:12: error: 'kvm_gmem_migrate_folio' defined but not used [-Werror=unused-function]
262 | static int kvm_gmem_migrate_folio(struct address_space *mapping,
| ^~~~~~~~~~~~~~~~~~~~~~
Remove the #ifdef around the reference so that fallback_migrate_folio()
is never used. The gmem implementation of the hook is trivial; since
the gmem mapping is unmovable, the pages should not be migrated anyway.
Fixes: a7800aa80e ("KVM: Add KVM_CREATE_GUEST_MEMFD ioctl() for guest-specific backing memory")
Reported-by: Arnd Bergmann <arnd@arndb.de>
Suggested-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
KVM/Arm supports readonly memslots; fix the calculation of
supported_flags in set_memory_region_test.c, otherwise the
test fails.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
In __team_options_register, team_options are allocated and appended to
the team's option_list.
If one option instance allocation fails, the "inst_rollback" cleanup
path frees the previously allocated options but doesn't remove them from
the team's option_list.
This leaves dangling pointers that can be dereferenced later by other
parts of the team driver that iterate over options.
This patch fixes the cleanup path to remove the dangling pointers from
the list.
As far as I can tell, this uaf doesn't have much security implications
since it would be fairly hard to exploit (an attacker would need to make
the allocation of that specific small object fail) but it's still nice
to fix.
Cc: stable@vger.kernel.org
Fixes: 80f7c6683f ("team: add support for per-port options")
Signed-off-by: Florent Revest <revest@chromium.org>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Hangbin Liu <liuhangbin@gmail.com>
Link: https://lore.kernel.org/r/20231206123719.1963153-1-revest@chromium.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Commit 227b60f510 added a seqlock to ensure that the low and high
port numbers were always updated together.
This is overkill because the two 16bit port numbers can be held in
a u32 and read/written in a single instruction.
More recently 91d0b78c51 added support for finer per-socket limits.
The user-supplied value is 'high << 16 | low' but they are held
separately and the socket options protected by the socket lock.
Use a u32 containing 'high << 16 | low' for both the 'net' and 'sk'
fields and use READ_ONCE()/WRITE_ONCE() to ensure both values are
always updated together.
Change (the now trival) inet_get_local_port_range() to a static inline
to optimise the calling code.
(In particular avoiding returning integers by reference.)
Signed-off-by: David Laight <david.laight@aculab.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Acked-by: Mat Martineau <martineau@kernel.org>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://lore.kernel.org/r/4e505d4198e946a8be03fb1b4c3072b0@AcuMS.aculab.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In general, activating long mode involves setting the EFER_LME bit in
the EFER register and then enabling the X86_CR0_PG bit in the CR0
register. At this point, the EFER_LMA bit will be set automatically by
hardware.
In the case of SVM/SEV guests where writes to CR0 are intercepted, it's
necessary for the host to set EFER_LMA on behalf of the guest since
hardware does not see the actual CR0 write.
In the case of SEV-ES guests where writes to CR0 are trapped instead of
intercepted, the hardware *does* see/record the write to CR0 before
exiting and passing the value on to the host, so as part of enabling
SEV-ES support commit f1c6366e30 ("KVM: SVM: Add required changes to
support intercepts under SEV-ES") dropped special handling of the
EFER_LMA bit with the understanding that it would be set automatically.
However, since the guest never explicitly sets the EFER_LMA bit, the
host never becomes aware that it has been set. This becomes problematic
when userspace tries to get/set the EFER values via
KVM_GET_SREGS/KVM_SET_SREGS, since the EFER contents tracked by the host
will be missing the EFER_LMA bit, and when userspace attempts to pass
the EFER value back via KVM_SET_SREGS it will fail a sanity check that
asserts that EFER_LMA should always be set when X86_CR0_PG and EFER_LME
are set.
Fix this by always inferring the value of EFER_LMA based on X86_CR0_PG
and EFER_LME, regardless of whether or not SEV-ES is enabled.
Fixes: f1c6366e30 ("KVM: SVM: Add required changes to support intercepts under SEV-ES")
Reported-by: Peter Gonda <pgonda@google.com>
Cc: stable@vger.kernel.org
Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20210507165947.2502412-2-seanjc@google.com>
[A two year old patch that was revived after we noticed the failure in
KVM_SET_SREGS and a similar patch was posted by Michael Roth. This is
Sean's patch, but with Michael's more complete commit message. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
It seems that when the driver is built-in, the HID bus is
initialized after the driver is loaded, which whould cause
module_hid_driver() to fail.
Fix this by registering the driver after the HID bus using
late_initcall() in accordance with other hwmon HID drivers.
Signed-off-by: Armin Wolf <W_Armin@gmx.de>
Link: https://lore.kernel.org/r/20231207210723.222552-1-W_Armin@gmx.de
[groeck: Dropped "compile tested" comment; the patch has been tested
but the tester did not provide a Tested-by: tag]
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Since the plane_state variable is declared outside the scaler_users
loop in intel_atomic_setup_scalers(), and it's never reset back to
NULL inside the loop we may end up calling intel_atomic_setup_scaler()
with a non-NULL plane state for the pipe scaling case. That is bad
because intel_atomic_setup_scaler() determines whether we are doing
plane scaling or pipe scaling based on plane_state!=NULL. The end
result is that we may miscalculate the scaler mode for pipe scaling.
The hardware becomes somewhat upset if we end up in this situation
when scanning out a planar format on a SDR plane. We end up
programming the pipe scaler into planar mode as well, and the
result is a screenfull of garbage.
Fix the situation by making sure we pass the correct plane_state==NULL
when calculating the scaler mode for pipe scaling.
Cc: stable@vger.kernel.org
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231207193441.20206-2-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
On ADL+ the hardware automagically calculates the CCS AUX surface
stride from the main surface stride, so when remapping we can't
really play a lot of tricks with the main surface stride, or else
the AUX surface stride would get miscalculated and no longer
match the actual data layout in memory.
Supposedly we could remap in 256 main surface tile units
(AUX page(4096)/cachline(64)*4(4x1 main surface tiles per
AUX cacheline)=256 main surface tiles), but the extra complexity
is probably not worth the hassle.
So let's just make sure our mapping stride is calculated from
the full framebuffer stride (instead of the framebuffer width).
This way the stride we program into PLANE_STRIDE will be the
original framebuffer stride, and thus there will be no change
to the AUX stride/layout.
Cc: stable@vger.kernel.org
Cc: Imre Deak <imre.deak@intel.com>
Cc: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231205180308.7505-1-ville.syrjala@linux.intel.com
Reviewed-by: Imre Deak <imre.deak@intel.com>
Using -MD without -MP causes build failures when a header file is deleted
or moved. With -MP, the compiler will emit phony targets for the header
files it lists as dependencies, and the Makefiles won't refuse to attempt
to rebuild a C unit which no longer includes the deleted header.
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Link: https://lore.kernel.org/r/9fc8b5395321abbfcaf5d78477a9a7cd350b08e4.camel@infradead.org
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Pass MAGIC_TOKEN to __TEST_REQUIRE() when printing the help message about
needing to pass a magic value to manually run the NX hugepages test,
otherwise the help message will contain garbage.
In file included from x86_64/nx_huge_pages_test.c:15:
x86_64/nx_huge_pages_test.c: In function ‘main’:
include/test_util.h:40:32: error: format ‘%d’ expects a matching ‘int’ argument [-Werror=format=]
40 | ksft_exit_skip("- " fmt "\n", ##__VA_ARGS__); \
| ^~~~
x86_64/nx_huge_pages_test.c:259:9: note: in expansion of macro ‘__TEST_REQUIRE’
259 | __TEST_REQUIRE(token == MAGIC_TOKEN,
| ^~~~~~~~~~~~~~
Signed-off-by: angquan yu <angquan21@gmail.com>
Link: https://lore.kernel.org/r/20231128221105.63093-1-angquan21@gmail.com
[sean: rewrite shortlog+changelog]
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
- When checking if a _running_ vCPU is "in-kernel", i.e. running at CPL0,
get the CPL directly instead of relying on preempted_in_kernel, which
is valid if and only if the vCPU was preempted, i.e. NOT running.
- Set .owner for various KVM file_operations so that files refcount the
KVM module until KVM is done executing _all_ code, including the last
few instructions of kvm_put_kvm(). And then revert the misguided
attempt to rely on "struct kvm" refcounts to pin KVM-the-module.
- Fix a benign "return void" that was recently introduced.
-----BEGIN PGP SIGNATURE-----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=gWp0
-----END PGP SIGNATURE-----
Merge tag 'kvm-x86-fixes-6.7-rcN' of https://github.com/kvm-x86/linux into kvm-master
KVM fixes for 6.7-rcN:
- When checking if a _running_ vCPU is "in-kernel", i.e. running at CPL0,
get the CPL directly instead of relying on preempted_in_kernel, which
is valid if and only if the vCPU was preempted, i.e. NOT running.
- Set .owner for various KVM file_operations so that files refcount the
KVM module until KVM is done executing _all_ code, including the last
few instructions of kvm_put_kvm(). And then revert the misguided
attempt to rely on "struct kvm" refcounts to pin KVM-the-module.
- Fix a benign "return void" that was recently introduced.
-----BEGIN PGP SIGNATURE-----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=2oOa
-----END PGP SIGNATURE-----
Merge tag 'kvm-s390-master-6.7-1' of https://git.kernel.org/pub/scm/linux/kernel/git/kvms390/linux into kvm-master
Two small but important bugfixes.
- Avoid mapping vLPIs that have already been mapped
-----BEGIN PGP SIGNATURE-----
iHUEABYIAB0WIQSNXHjWXuzMZutrKNKivnWIJHzdFgUCZWub4QAKCRCivnWIJHzd
FnZPAP9KlyKZTy3RvczIy4imn+DebrASAApmREkAIvfmz9wFzQD+Lg8AfxPAMz8E
e9eNpTZ48Zqzmas93/eu1t6ZpcJOZQA=
=fyc9
-----END PGP SIGNATURE-----
Merge tag 'kvmarm-fixes-6.7-1' of git://git.kernel.org/pub/scm/linux/kernel/git/kvmarm/kvmarm into kvm-master
KVM/arm64 fixes for 6.7, take #1
- Avoid mapping vLPIs that have already been mapped
When building whole selftests on arm64, rsync gives an erorr about sgx:
rsync: [sender] link_stat "/root/linux-next/tools/testing/selftests/sgx/test_encl.elf" failed: No such file or directory (2)
rsync error: some files/attrs were not transferred (see previous errors) (code 23) at main.c(1327) [sender=3.2.5]
The root casue is sgx only used on X86_64, and shall be skipped on other
platforms.
Fix this by moving TEST_CUSTOM_PROGS and TEST_FILES inside the if check,
then the build result will be "Skipping non-existent dir: sgx".
Fixes: 2adcba79e6 ("selftests/x86: Add a selftest for SGX")
Signed-off-by: Zhao Mengmeng <zhaomengmeng@kylinos.cn>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Link: https://lore.kernel.org/all/20231206025605.3965302-1-zhaomzhao%40126.com
Building the test enclave with -static-pie may produce a dynamic symbol
table, but this is not supported for enclaves and any relocations need to
happen manually (e.g., as for "encl_op_array"). Thus, opportunistically
discard ".dyn*" and ".gnu.hash" which the enclave loader cannot handle.
Signed-off-by: Jo Van Bulck <jo.vanbulck@cs.kuleuven.be>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Link: https://lore.kernel.org/all/20231005153854.25566-13-jo.vanbulck%40cs.kuleuven.be
The external tests manipulating page permissions expect encl_buffer to be
placed at the start of the test enclave's .data section. As this is not
guaranteed per the C standard, explicitly place encl_buffer in a separate
section that is explicitly placed at the start of the .data segment in the
linker script to avoid the compiler placing it somewhere else in .data.
Signed-off-by: Jo Van Bulck <jo.vanbulck@cs.kuleuven.be>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Acked-by: Kai Huang <kai.huang@intel.com>
Link: https://lore.kernel.org/all/20231005153854.25566-12-jo.vanbulck%40cs.kuleuven.be
DEFINED only considers symbols, not section names. Hence, replace the
check for .got.plt with the _GLOBAL_OFFSET_TABLE_ symbol and remove other
(non-essential) asserts.
Signed-off-by: Jo Van Bulck <jo.vanbulck@cs.kuleuven.be>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Link: https://lore.kernel.org/all/20231005153854.25566-10-jo.vanbulck%40cs.kuleuven.be
Static-pie binaries normally include a startup routine to perform any ELF
relocations from .rela.dyn. Since the enclave loading process is different
and glibc is not included, do the necessary relocation for encl_op_array
entries manually at runtime relative to the enclave base to ensure correct
function pointers.
When keeping encl_op_array as a local variable on the stack, gcc without
optimizations generates code that explicitly gets the right function
addresses and stores them to create the array on the stack:
encl_body:
/* snipped */
lea do_encl_op_put_to_buf(%rip), %rax
mov %rax, -0x50(%rbp)
lea do_encl_op_get_from_buf(%rip), %rax
mov %rax,-0x48(%rbp)
lea do_encl_op_put_to_addr(%rip), %rax
/* snipped */
However, gcc -Os or clang generate more efficient code that initializes
encl_op_array by copying a "prepared copy" containing the absolute
addresses of the functions (i.e., relative to the image base starting from
0) generated by the compiler/linker:
encl_body:
/* snipped */
lea prepared_copy(%rip), %rsi
lea -0x48(%rsp), %rdi
mov $0x10,%ecx
rep movsl %ds:(%rsi),%es:(%rdi)
/* snipped */
When building the enclave with -static-pie, the compiler/linker includes
relocation entries for the function symbols in the "prepared copy":
Relocation section '.rela.dyn' at offset 0x4000 contains 12 entries:
Offset Info Type Symbol
/* snipped; "prepared_copy" starts at 0x6000 */
000000006000 000000000008 R_X86_64_RELATIVE <do_encl_emodpe>
000000006008 000000000008 R_X86_64_RELATIVE <do_encl_eaccept>
000000006010 000000000008 R_X86_64_RELATIVE <do_encl_op_put_to_buf>
000000006018 000000000008 R_X86_64_RELATIVE <do_encl_op_get_from_buf>
000000006020 000000000008 R_X86_64_RELATIVE <do_encl_op_put_to_addr>
000000006028 000000000008 R_X86_64_RELATIVE <do_encl_op_get_from_addr>
000000006030 000000000008 R_X86_64_RELATIVE <do_encl_op_nop>
000000006038 000000000008 R_X86_64_RELATIVE <do_encl_init_tcs_page>
Static-pie binaries normally include a glibc "_dl_relocate_static_pie"
routine that will perform these relocations as part of the startup.
However, since the enclave loading process is different and glibc is not
included, we cannot rely on these relocations to be performed. Without
relocations, the code would erroneously jump to the _absolute_ function
address loaded from the local copy.
Thus, declare "encl_op_array" as global and manually relocate the loaded
function-pointer entries relative to the enclave base at runtime. This
generates the following code:
encl_body:
/* snipped */
lea encl_op_array(%rip), %rcx
lea __encl_base(%rip), %rax
add (%rcx,%rdx,8),%rax
jmp *%rax
Signed-off-by: Jo Van Bulck <jo.vanbulck@cs.kuleuven.be>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Acked-by: Kai Huang <kai.huang@intel.com>
Link: https://lore.kernel.org/all/150d8ca8-2c66-60d1-f9fc-8e6279824e94@cs.kuleuven.be/
Link: https://lore.kernel.org/all/5c22de5a-4b3b-1f38-9771-409b4ec7f96d@cs.kuleuven.be/#r
Link: https://lore.kernel.org/all/20231005153854.25566-9-jo.vanbulck%40cs.kuleuven.be
The current combination of -static and -fPIC creates a static executable
with position-dependent addresses for global variables. Use -static-pie
and -fPIE to create a proper static position independent executable that
can be loaded at any address without a dynamic linker.
When building the original "lea (encl_stack)(%rbx), %rax" assembly code
with -static-pie -fPIE, the linker complains about a relocation it cannot
resolve:
/usr/local/bin/ld: /tmp/cchIWyfG.o: relocation R_X86_64_32S against
`.data' can not be used when making a PIE object; recompile with -fPIE
collect2: error: ld returned 1 exit status
Thus, since only RIP-relative addressing is legit for local symbols, use
"encl_stack(%rip)" and declare an explicit "__encl_base" symbol at the
start of the linker script to be able to calculate the stack address
relative to the current TCS in the enclave assembly entry code.
Signed-off-by: Jo Van Bulck <jo.vanbulck@cs.kuleuven.be>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Acked-by: Kai Huang <kai.huang@intel.com>
Link: https://lore.kernel.org/all/f9c24d89-ed72-7d9e-c650-050d722c6b04@cs.kuleuven.be/
Link: https://lore.kernel.org/all/20231005153854.25566-8-jo.vanbulck%40cs.kuleuven.be
Remove redundant push/pop pair that stores and restores the enclave base
address in the test enclave, as it is never used after the pop and can
anyway be easily retrieved via the __encl_base symbol.
Signed-off-by: Jo Van Bulck <jo.vanbulck@cs.kuleuven.be>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Acked-by: Kai Huang <kai.huang@intel.com>
Link: https://lore.kernel.org/all/20231005153854.25566-7-jo.vanbulck%40cs.kuleuven.be
Use -ffreestanding to assert the enclave compilation targets a
freestanding environment (i.e., without "main" or standard libraries).
This fixes clang reporting "undefined reference to `memset'" after
erroneously optimizing away the provided memset/memcpy implementations.
Still need to instruct the linker from using standard system startup
functions, but drop -nostartfiles as it is implied by -nostdlib.
Signed-off-by: Jo Van Bulck <jo.vanbulck@cs.kuleuven.be>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Acked-by: Kai Huang <kai.huang@intel.com>
Link: https://lore.kernel.org/all/20231005153854.25566-6-jo.vanbulck%40cs.kuleuven.be
Fixes "'linker' input unused [-Wunused-command-line-argument]" errors when
compiling with clang.
Signed-off-by: Jo Van Bulck <jo.vanbulck@cs.kuleuven.be>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Link: https://lore.kernel.org/all/20231005153854.25566-5-jo.vanbulck%40cs.kuleuven.be
Add the "memory" clobber to the EMODPE and EACCEPT asm blocks to tell the
compiler the assembly code accesses to the secinfo struct. This ensures
the compiler treats the asm block as a memory barrier and the write to
secinfo will be visible to ENCLU.
Fixes: 20404a8085 ("selftests/sgx: Add test for EPCM permission changes")
Signed-off-by: Jo Van Bulck <jo.vanbulck@cs.kuleuven.be>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Kai Huang <kai.huang@intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Link: https://lore.kernel.org/all/20231005153854.25566-4-jo.vanbulck%40cs.kuleuven.be
Ensure sym_tab and sym_names are zero-initialized and add an early-out
condition in the unlikely (erroneous) case that the enclave ELF file would
not contain a symbol table.
This addresses -Werror=maybe-uninitialized compiler warnings for gcc -O2.
Fixes: 33c5aac3bf ("selftests/sgx: Test complete changing of page type flow")
Signed-off-by: Jo Van Bulck <jo.vanbulck@cs.kuleuven.be>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Link: https://lore.kernel.org/all/20231005153854.25566-3-jo.vanbulck%40cs.kuleuven.be
Ensure ctx is zero-initialized, such that the encl_measure function will
not call EVP_MD_CTX_destroy with an uninitialized ctx pointer in case of an
early error during key generation.
Fixes: 2adcba79e6 ("selftests/x86: Add a selftest for SGX")
Signed-off-by: Jo Van Bulck <jo.vanbulck@cs.kuleuven.be>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Acked-by: Kai Huang <kai.huang@intel.com>
Link: https://lore.kernel.org/all/20231005153854.25566-2-jo.vanbulck%40cs.kuleuven.be
The address/size cells in PWM node are needed only if individual LEDs
are listed. If multi-led is used, then this leads to dtc W=1 warnings:
pm8550.dtsi:65.19-73.5: Warning (avoid_unnecessary_addr_size): /soc@0/spmi@c400000/pmic@1/pwm:
unnecessary #address-cells/#size-cells without "ranges" or child "reg" property
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20231208124332.48636-1-krzysztof.kozlowski@linaro.org
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
In libbpf, when determining whether we need to load vmlinux btf, we're
currently (among other things) checking whether there is any struct_ops
program present in the object. This works for most realistic struct_ops
maps, as a struct_ops map is of course typically composed of one or more
struct_ops programs. However, that technically need not be the case. A
struct_ops interface could be defined which allows a map to be specified
which one or more non-prog fields, and which provides default behavior
if no struct_ops progs is actually provided otherwise. For sched_ext,
for example, you technically only need to specify the name of the
scheduler in the struct_ops map, with the core scheduler logic providing
default behavior if no prog is actually specified.
If we were to define and try to load such a struct_ops map, we would
crash in libbpf when initializing it as obj->btf_vmlinux will be NULL:
Reading symbols from minimal...
(gdb) r
Starting program: minimal_example
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/usr/lib/libthread_db.so.1".
Program received signal SIGSEGV, Segmentation fault.
0x000055555558308c in btf__type_cnt (btf=0x0) at btf.c:612
612 return btf->start_id + btf->nr_types;
(gdb) bt
type_name=0x5555555d99e3 "sched_ext_ops", kind=4) at btf.c:914
kind=4) at btf.c:942
type=0x7fffffffe558, type_id=0x7fffffffe548, ...
data_member=0x7fffffffe568) at libbpf.c:948
kern_btf=0x0) at libbpf.c:1017
at libbpf.c:8059
So as to account for such bare-bones struct_ops maps, let's update
obj_needs_vmlinux_btf() to also iterate over an obj's maps and check
whether any of them are struct_ops maps.
Signed-off-by: David Vernet <void@manifault.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Reviewed-by: Alan Maguire <alan.maguire@oracle.com>
Link: https://lore.kernel.org/bpf/20231208061704.400463-1-void@manifault.com
TDX is incompatible with hibernation and some ACPI sleep states.
Users must disable hibernation to use TDX. Users must also disable
TDX if they want to use ACPI S3 sleep.
This feels a bit wonky and asymmetric, but it avoids adding any new
command-line parameters for now. It can be improved if users hate it
too much.
Long version:
TDX cannot survive from S3 and deeper states. The hardware resets and
disables TDX completely when platform goes to S3 and deeper. Both TDX
guests and the TDX module get destroyed permanently.
The kernel uses S3 to support suspend-to-ram, and S4 or deeper states to
support hibernation. The kernel also maintains TDX states to track
whether it has been initialized and its metadata resource, etc. After
resuming from S3 or hibernation, these TDX states won't be correct
anymore.
Theoretically, the kernel can do more complicated things like resetting
TDX internal states and TDX module metadata before going to S3 or
deeper, and re-initialize TDX module after resuming, etc, but there is
no way to save/restore TDX guests for now.
Until TDX supports full save and restore of TDX guests, there is no big
value to handle TDX module in suspend and hibernation alone. To make
things simple, just choose to make TDX mutually exclusive with S3 and
hibernation.
Note the TDX module is initialized at runtime. To avoid having to deal
with the fuss of determining TDX state at runtime, just choose TDX vs S3
and hibernation at kernel early boot. It's a bad user experience if the
choice of TDX and S3/hibernation is done at runtime anyway, i.e., the
user can experience being able to do S3/hibernation but later becoming
unable to due to TDX being enabled.
Disable TDX in kernel early boot when hibernation support is available.
Currently there's no mechanism exposed by the hibernation code to allow
other kernel code to disable hibernation once for all. Users that want
TDX must disable hibernation, like using hibername=no on the command
line.
Disable ACPI S3 when TDX is enabled by the BIOS. For now the user needs
to disable TDX in the BIOS to use ACPI S3. A new kernel command line
can be added in the future if there's a need to let user disable TDX
host via kernel command line.
Alternatively, the kernel could disable TDX when ACPI S3 is supported
and request the user to disable S3 to use TDX. But there's no existing
kernel command line to do that, and BIOS doesn't always have an option
to disable S3.
[ dhansen: subject / changelog tweaks ]
Signed-off-by: Kai Huang <kai.huang@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reviewed-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lore.kernel.org/all/20231208170740.53979-16-dave.hansen%40intel.com
After the global KeyID has been configured on all packages, initialize
all TDMRs to make all TDX-usable memory regions that are passed to the
TDX module become usable.
This is the last step of initializing the TDX module.
Initializing TDMRs can be time consuming on large memory systems as it
involves initializing all metadata entries for all pages that can be
used by TDX guests. Initializing different TDMRs can be parallelized.
For now to keep it simple, just initialize all TDMRs one by one. It can
be enhanced in the future.
Signed-off-by: Kai Huang <kai.huang@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Isaku Yamahata <isaku.yamahata@intel.com>
Reviewed-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reviewed-by: Yuan Yao <yuan.yao@intel.com>
Reviewed-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lore.kernel.org/all/20231208170740.53979-15-dave.hansen%40intel.com
After the list of TDMRs and the global KeyID are configured to the TDX
module, the kernel needs to configure the key of the global KeyID on all
packages using TDH.SYS.KEY.CONFIG.
This SEAMCALL cannot run parallel on different cpus. Loop all online
cpus and use smp_call_on_cpu() to call this SEAMCALL on the first cpu of
each package.
To keep things simple, this implementation takes no affirmative steps to
online cpus to make sure there's at least one cpu for each package. The
callers (aka. KVM) can ensure success by ensuring sufficient CPUs are
online for this to succeed.
Intel hardware doesn't guarantee cache coherency across different
KeyIDs. The PAMTs are transitioning from being used by the kernel
mapping (KeyId 0) to the TDX module's "global KeyID" mapping.
This means that the kernel must flush any dirty KeyID-0 PAMT cachelines
before the TDX module uses the global KeyID to access the PAMTs.
Otherwise, if those dirty cachelines were written back, they would
corrupt the TDX module's metadata. Aside: This corruption would be
detected by the memory integrity hardware on the next read of the memory
with the global KeyID. The result would likely be fatal to the system
but would not impact TDX security.
Following the TDX module specification, flush cache before configuring
the global KeyID on all packages. Given the PAMT size can be large
(~1/256th of system RAM), just use WBINVD on all CPUs to flush.
If TDH.SYS.KEY.CONFIG fails, the TDX module may already have "converted"
some memory for TDX module use. Convert the memory back so that it can
be safely used by the kernel again. Note that this is slower than it
should be because of the "partial write machine check" erratum which
affects TDX-capable hardware.
Also refactor and introduce a new helper: tdmr_do_pamt_func(). This
takes a TDMR and runs a function on its PAMT. It looks a _bit_ odd to
pass a function pointer around like this, but its use is pretty narrow
and it does eliminate what would otherwise be some copying and pasting.
[ dhansen: * munge changelog as usual
* remove weird (*pamd_func)() syntax ]
Signed-off-by: Kai Huang <kai.huang@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Isaku Yamahata <isaku.yamahata@intel.com>
Reviewed-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reviewed-by: Yuan Yao <yuan.yao@intel.com>
Reviewed-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lore.kernel.org/all/20231208170740.53979-14-dave.hansen%40intel.com
The TDX module uses a private KeyID as the "global KeyID" for mapping
things like the PAMT and other TDX metadata. This KeyID has already
been reserved when detecting TDX during the kernel early boot.
Now that the "TD Memory Regions" (TDMRs) are fully built, pass them to
the TDX module together with the global KeyID.
Signed-off-by: Kai Huang <kai.huang@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Isaku Yamahata <isaku.yamahata@intel.com>
Reviewed-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reviewed-by: Yuan Yao <yuan.yao@intel.com>
Reviewed-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lore.kernel.org/all/20231208170740.53979-13-dave.hansen%40intel.com
As the last step of constructing TDMRs, populate reserved areas for all
TDMRs. Cover all memory holes and PAMTs with a TMDR reserved area.
[ dhansen: trim down chagnelog ]
Signed-off-by: Kai Huang <kai.huang@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Isaku Yamahata <isaku.yamahata@intel.com>
Reviewed-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reviewed-by: Yuan Yao <yuan.yao@intel.com>
Reviewed-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lore.kernel.org/all/20231208170740.53979-12-dave.hansen%40intel.com
The TDX module uses additional metadata to record things like which
guest "owns" a given page of memory. This metadata, referred as
Physical Address Metadata Table (PAMT), essentially serves as the
'struct page' for the TDX module. PAMTs are not reserved by hardware
up front. They must be allocated by the kernel and then given to the
TDX module during module initialization.
TDX supports 3 page sizes: 4K, 2M, and 1G. Each "TD Memory Region"
(TDMR) has 3 PAMTs to track the 3 supported page sizes. Each PAMT must
be a physically contiguous area from a Convertible Memory Region (CMR).
However, the PAMTs which track pages in one TDMR do not need to reside
within that TDMR but can be anywhere in CMRs. If one PAMT overlaps with
any TDMR, the overlapping part must be reported as a reserved area in
that particular TDMR.
Use alloc_contig_pages() since PAMT must be a physically contiguous area
and it may be potentially large (~1/256th of the size of the given TDMR).
The downside is alloc_contig_pages() may fail at runtime. One (bad)
mitigation is to launch a TDX guest early during system boot to get
those PAMTs allocated at early time, but the only way to fix is to add a
boot option to allocate or reserve PAMTs during kernel boot.
It is imperfect but will be improved on later.
TDX only supports a limited number of reserved areas per TDMR to cover
both PAMTs and memory holes within the given TDMR. If many PAMTs are
allocated within a single TDMR, the reserved areas may not be sufficient
to cover all of them.
Adopt the following policies when allocating PAMTs for a given TDMR:
- Allocate three PAMTs of the TDMR in one contiguous chunk to minimize
the total number of reserved areas consumed for PAMTs.
- Try to first allocate PAMT from the local node of the TDMR for better
NUMA locality.
Also dump out how many pages are allocated for PAMTs when the TDX module
is initialized successfully. This helps answer the eternal "where did
all my memory go?" questions.
[ dhansen: merge in error handling cleanup ]
Signed-off-by: Kai Huang <kai.huang@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Isaku Yamahata <isaku.yamahata@intel.com>
Reviewed-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reviewed-by: Yuan Yao <yuan.yao@intel.com>
Link: https://lore.kernel.org/all/20231208170740.53979-11-dave.hansen%40intel.com