aboutsummaryrefslogtreecommitdiff
path: root/sys/amd64
Commit message (Collapse)AuthorAgeFilesLines
...
* amd64: on any fault during call to EFI RT, restore execution and print fault ↵Konstantin Belousov2024-12-251-12/+18
| | | | | | | | | | | | | | details The fault info should be useful to see what specifically BIOS tried to do and why it faulted. E.g. it might allow to see which EFI memory segment needs to be mapped in addition to normal runtime segments, to work around the fault. Reviewed by: kevans, markj Sponsored by: The FreeBSD Foundation MFC after: 1 week Differential revision: https://reviews.freebsd.org/D48186
* amd64: extract code to print fault details from trap_fatal() into a new helperKonstantin Belousov2024-12-251-7/+22
| | | | | | | | Suggested by: markj Reviewed by: kevans, markj Sponsored by: The FreeBSD Foundation MFC after: 1 week Differential revision: https://reviews.freebsd.org/D48186
* x86 atomics: Remove unused WANT_FUNCTIONSOlivier Certner2024-12-171-3/+0
| | | | | | | | | | This macro has not been in use since commit "inline atomics and allow tied modules to inline locks" (r335873, f4b3640475cec929). Reviewed by: markj, kib, emaste, imp MFC after: 5 days Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D48061
* atomics: Constify loadsOlivier Certner2024-12-161-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | In order to match reality, allow using these functions with pointers on const objects, and bring us closer to C11. Remove the '+' modifier in the atomic_load_acq_64_i586()'s inline asm statement's constraint for '*p' (the value to load). CMPXCHG8B always writes back some value, even when the value exchange does not happen in which case what was read is written back. atomic_load_acq_64_i586() further takes care of the operation atomically writing back the same value that was read in any case. All in all, this makes the inline asm's write back undetectable by any other code, whether executing on other CPUs or code on the same CPU before and after the call to atomic_load_acq_64_i586(), except for the fact that CMPXCHG8B will trigger a #GP(0) if the memory address is part of a read-only mapping. This unfortunate property is however out of scope of the C abstract machine, and in particular independent of whether the 'uint64_t' pointed to is declared 'const' or not. Approved by: markj (mentor) MFC after: 5 days Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D46887
* x86: Add routines for querying XSAVE feature informationBojan Novković2024-12-151-1/+110
| | | | | | | | | | | | This patch adds several routines that track and expose information about various XSAVE-related features. More specifically, it adds the ability to check whether a given XFEATURE is supported and which XSAVE extensions are supported. Furthermore, it adds several routines for calculating the size and offsets within a save area given a XSAVE feature bitmap. Reviewed by: kib Differential Revision: https://reviews.freebsd.org/D47394
* amd64: Add wrappers for XRSTORS and XSAVESBojan Novković2024-12-151-0/+23
| | | | | Reviewed by: kib Differential Revision: https://reviews.freebsd.org/D46984
* x86: Refactor kernel-mode NMI handlingBojan Novković2024-12-151-28/+4
| | | | | | | | | | This refactor aims to add the ability to share performance counter interrupts by refactoring the kernel-mode NMI handler. The handler now allows multiple drivers to service the same interrupt (e.g. hwpmc(4) and hwt(4)'s Intel Processor Trace backend). Reviewed by: kib, avg Differential Revision: https://reviews.freebsd.org/D46421
* vmcb.h: fix typo in the comment, try twoKonstantin Belousov2024-12-101-1/+1
| | | | | | Noted by: rm Sponsored by: The FreeBSD Foundation MFC after: 3 days
* vmcb.h: fix typo in the commentKonstantin Belousov2024-12-101-1/+1
| | | | | Sponsored by: The FreeBSD Foundation MFC after: 3 days
* amd64: re-enable la57, try twoKonstantin Belousov2024-12-091-0/+1
| | | | | | | | | | If you need to use libc with jemalloc without commit eade2001aa9d91440886de8359a4dec9edcde2a9 (head, or b90d68c9b28b0a458bdf4a963397817974ed0f90 stable/14), you have to set vm.pmap.la57 to 0 for system-global effect, or use proccontrol -m la48 on affected image. Sponsored by: The FreeBSD Foundation
* Replace calls to bus_generic_attach with bus_attach_childrenJohn Baldwin2024-12-061-1/+2
| | | | | Reviewed by: imp Differential Revision: https://reviews.freebsd.org/D47675
* Replace calls to bus_generic_probe with bus_identify_childrenJohn Baldwin2024-12-061-1/+1
| | | | | Reviewed by: imp Differential Revision: https://reviews.freebsd.org/D47674
* amd64/vmm: Add svm_modsuspend to the vmm_ops_amd tableMark Johnston2024-12-031-2/+1
| | | | | Fixes: 0b32ef71f9f1 ("vmm: Correctly suspend and resume the vmm driver.") Sponsored by: Klara, Inc.
* vmm: Correctly suspend and resume the vmm driver.Joshua Rogers2024-11-297-2/+28
| | | | | | | | | | | | | | | Previously, VMXON would be executed on a resume, contrary to proper initalization. The contents of MSR_IA32_FEATURE_CONTROL may be lost on suspension, therefore must be restored. Likewise, the VMX Enable bit may be cleared upon suspend, requiring it to be re-set. Concretely disable VMX on suspend, and re-enable it on resume. Note: any IOMMU context will remain lost for any enabled vmm devices. Signed-off-by: Joshua Rogers <Joshua@Joshua.Hu> Reviewed by: jhb,imp Pull Request: https://github.com/freebsd/freebsd-src/pull/1419
* amd64: add machine/pte.hKonstantin Belousov2024-11-262-57/+105
| | | | | | | | | | | Following arm64 and risc-v, move definitions that describe hardware-enforced layout of PTEs and #PF error bits, into a dedicated header. Reviewed by: markj Sponsored by: The FreeBSD Foundation MFC after: 1 week Differential revision: https://reviews.freebsd.org/D47749
* amd64: switch pmap_map_io_transient() to use pmap_kenter_attr()Konstantin Belousov2024-11-251-7/+3
| | | | | | | | | | | instead of constructing transient pte itself. This pre-set PG_A and PG_M bits, avoiding atomic pte update on access and modification. Also it set the nx bit, the mapping is not supposed to be used for executing. Reviewed by: alc Sponsored by: The FreeBSD Foundation MFC after: 1 week Differential revision: https://reviews.freebsd.org/D47717
* amd64 pmap: assert and explain why pmap_qremove() is safe WRT supermappingsKonstantin Belousov2024-11-251-0/+12
| | | | | | | | | Based on alc@ comments from https://reviews.freebsd.org/D47678. Reviewed by: alc Sponsored by: The FreeBSD Foundation MFC after: 1 week Differential revision: https://reviews.freebsd.org/D47717
* amd64 efi rt: handle #BPKonstantin Belousov2024-11-211-0/+12
| | | | | | | | PR: 282860 Reviewed by: markj Sponsored by: The FreeBSD Foundation MFC after: 1 week Differential revision: https://reviews.freebsd.org/D47694
* atomic(9): Implement atomic_testand(clear|set)_ptrJohn Baldwin2024-11-191-0/+2
| | | | | | | | | | For current architectures, these are just aliases for the existing operation on the relevant scalar integer. Reviewed by: imp, kib Obtained from: CheriBSD Sponsored by: AFRL, DARPA Differential Revision: https://reviews.freebsd.org/D47631
* vmm: Postpone vmm module initialization to after SI_SUB_DEVFSMark Johnston2024-11-071-1/+2
| | | | | | | | | | | | | vmmops_modinit() needs to create a device file, and this must happen after SI_SUB_DEVFS. On non-EARLY_AP_STARTUP platforms (i.e., !x86) this happens already by accident, but we shouldn't rely on it. On riscv, remove the current SI_SUB_SMP ordering since that was copied from arm64 and isn't needed. In particular, riscv's vmmops_modinit() does not call smp_rendezvous(). Reported by: Oleksandr Kryvulia <shuriku@shurik.kiev.ua> Fixes: a97f683fe3c4 ("vmm: Add a device file interface for creating and destroying VMs")
* vmm: Add a device file interface for creating and destroying VMsMark Johnston2024-11-051-1/+3
| | | | | | | | | | | | | | | | | | | | | | | | | This supersedes the sysctl interface, which has the limitations of being root-only and not supporting automatic resource destruction, i.e., we cannot easily destroy VMs automatically when bhyve terminates. For now, two ioctls are implemented VMMCTL_VM_CREATE and VMMCTL_VM_DESTROY. Eventually I would like to support tying a VM's lifetime to that of the descriptor, so that it is automatically destroyed when the descriptor is closed. However, this will require some work in bhyve: when the guest wants to reboot, bhyve exits with a status that indicates that it is to be restarted. This is incompatible with the idea of tying a VM's lifetime to that of a descriptor, since we want to avoid creating and destroying a VM across each reboot (as this involves freeing all of the guest memory, among other things). One possible design would be to decompose bhyve into two processes, a parent which handles reboots, and a child which runs in capability mode and handles guest execution. In any case, this gets us closer to addressing the shortcomings mentioned above. Reviewed by: jhb Differential Revision: https://reviews.freebsd.org/D47028
* vmm: Rename the amdiommu driver to amdviiommuMark Johnston2024-11-041-19/+19
| | | | | | | | To avoid a conflict with the new amdiommu driver imported recently. Fixes: 0f5116d7efe3 ("AMD IOMMU driver") Reviewed by: kib Differential Revision: https://reviews.freebsd.org/D47415
* x86: Remove invalid DEVMETHOD methods for leaf devicesJohn Baldwin2024-11-011-4/+0
| | | | | | | | | None of these drivers are for bus devices, so bus_generic_* is not appropriate. Most of these were nops except that detach would actually "succeed" (but not do any cleanup). Reviewed by: imp Differential Revision: https://reviews.freebsd.org/D47374
* vmm: fix vcpu atomic loadRuslan Bukin2024-10-291-1/+2
| | | | | | | | | | | Load vcpu with acquire semantics as we are making a critical code section between creating vcpu and using it. Tested on risc-v only. Pointed out by: markj Reviewed by: jhb, markj Differential Revision: https://reviews.freebsd.org/D47306
* sysent: regen for typo fixBrooks Davis2024-10-222-2/+2
|
* la57: explain how the trampoline worksKonstantin Belousov2024-10-201-20/+41
| | | | | | | Reviewed by: markj, imp (previous version) Sponsored by: The FreeBSD Foundation MFC after: 3 days Differential revision: https://reviews.freebsd.org/D47208
* vmm: avoid potential KASSERT kernel panic in vm_handle_dbPierre Pronchery2024-10-021-1/+2
| | | | | | | | | | | | | | | | | | | | | | | | | If the guest VM emits the exit code VM_EXITCODE_DB the kernel will execute the function named vm_handle_db. If the value of rsp is not page aligned and if rsp+sizeof(uint64_t) spans across two pages, the function vm_copy_setup will need two structs vm_copyinfo to prepare the copy operation. For instance is rsp value is 0xFFC, two vm_copyinfo objects are needed: * address=0xFFC, len=4 * address=0x1000, len=4 The vulnerability was addressed by commit 51fda658baa ("vmm: Properly handle writes spanning across two pages in vm_handle_db"). Still, replace the KASSERT with an error return as a more defensive approach. Reported by: Synacktiv Reviewed by markj, emaste Security: HYP-09 Sponsored by: The Alpha-Omega Project Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D46133
* vmm: Properly handle writes spanning across two pages in vm_handle_dbBojan Novković2024-10-021-5/+5
| | | | | | | | | | | | The vm_handle_db function is responsible for writing correct status register values into memory when a guest VM is being single-stepped using the RFLAGS.TF mechanism. However, it currently does not properly handle an edge case where the resulting write spans across two pages. This commit fixes this by making vm_handle_db use two vm_copy_info structs. Security: HYP-09 Reviewed by: markj
* sysent: regen commentsBrooks Davis2024-10-012-2/+4
|
* sysent: regen removing comment alignmentBrooks Davis2024-10-012-151/+151
|
* amd64/mp_machdep.c: styleKonstantin Belousov2024-10-011-6/+5
| | | | | | | | Wrap long lines. Remove redundand declaration. Sponsored by: The FreeBSD Foundation MFC after: 3 days
* bhyve: initialize register valuePierre Pronchery2024-09-271-1/+1
| | | | | | | | | | | | | | | | | | | In case of an error in a code pattern like ``` uint64_t val; error = memread(vcpu, gpa, &val, 1, arg); error = vie_update_register(vcpu, reg, val, size); ``` uninitialized stack data would be used. Reported by: Synacktiv Reviewed by: markj Security: HYP-21 Sponsored by: The Alpha-Omega Project Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D46107
* Remove stray whitespaces from sys/amd64/Joshua Rogers2024-09-2141-121/+121
| | | | | | Signed-off-by: Joshua Rogers <Joshua@Joshua.Hu> Reviewed by: imp Pull Request: https://github.com/freebsd/freebsd-src/pull/1418
* Add a new sysctl in order to diffrentiate UEFI architecturesAhmad Khalifa2024-09-201-0/+21
| | | | | | | | With the new 32-bit UEFI loader, it's convenient to have a sysctl to figure out how we booted. Can be accessed at machdep.efi_arch Reviewed by: imp Pull Request: https://github.com/freebsd/freebsd-src/pull/1098
* sysarch: improve checks for max user addressKonstantin Belousov2024-09-161-2/+3
| | | | | | | | making LA48 processes have the same limit as with the pre-LA57 kernels. Sponsored by: Advanced Micro Devices (AMD) Sponsored by: The FreeBSD Foundation MFC after: 1 week
* amd64 sysarch(2): styleKonstantin Belousov2024-09-161-4/+4
| | | | | | Sponsored by: Advanced Micro Devices (AMD) Sponsored by: The FreeBSD Foundation MFC after: 1 week
* amd64: pml5 entries do not support PAT bitsKonstantin Belousov2024-09-161-6/+3
| | | | | | Sponsored by: Advanced Micro Devices (AMD) Sponsored by: The FreeBSD Foundation MFC after: 1 week
* amd64 pmap: do not set PG_G for usermode pmap pml5 kernel entryKonstantin Belousov2024-09-161-2/+2
| | | | | | Sponsored by: Advanced Micro Devices (AMD) Sponsored by: The FreeBSD Foundation MFC after: 1 week
* pmap_bootstrap_la57(): reload IDTKonstantin Belousov2024-09-161-0/+1
| | | | | | | | after the trip through protected mode. This is required by AMD64 ARM. Sponsored by: Advanced Micro Devices (AMD) Sponsored by: The FreeBSD Foundation MFC after: 1 week
* la57: do not set global bit for PML5 entryKonstantin Belousov2024-09-161-1/+1
| | | | | | | | | The bit is reserved for PLM5, causing #PF on KVA access on real hardware, unlike QEMU. Sponsored by: Advanced Micro Devices (AMD) Sponsored by: The FreeBSD Foundation MFC after: 1 week
* amd64 la57_trampoline: save registers in memoryKonstantin Belousov2024-09-161-9/+12
| | | | | | | | | AMD64 ARM states that 64bit part of the architectural state is undefined after 32<->64 mode switching. Sponsored by: Advanced Micro Devices (AMD) Sponsored by: The FreeBSD Foundation MFC after: 1 week
* amd64 la57_trampoline: lgdt descriptor is always 10 bytes in long modeKonstantin Belousov2024-09-161-1/+1
| | | | | | | | | | Extend its storage to be compliant. This is currently nop due to padding and nul gdt descriptor right after the lgdt descriptor. Sponsored by: Advanced Micro Devices (AMD) Sponsored by: The FreeBSD Foundation MFC after: 1 week
* amd64 la57_trampoline: turn off global pages and PCID before turning off pagingKonstantin Belousov2024-09-161-0/+7
| | | | | | | | | SDM is explicit that having CR4.PCID=1 while toggling CR3.PG causes #GP. To be safe and to avoid some more effects, also turn off CR4.PGE. Sponsored by: Advanced Micro Devices (AMD) Sponsored by: The FreeBSD Foundation MFC after: 1 week
* amd64 la57_trampoline: disable EFER.LME around setting CR4.LA57Konstantin Belousov2024-09-161-0/+10
| | | | | | | | Changing paging mode while LME is set seems to be not allowed. Sponsored by: Advanced Micro Devices (AMD) Sponsored by: jThe FreeBSD Foundation MFC after: 1 week
* amd64 la57_trampoline: stop using %rdx to remember original %cr0Konstantin Belousov2024-09-161-3/+5
| | | | | | | | Store %cr0 in %ebp. %rdx is needed for MSR access. Sponsored by: Advanced Micro Devices (AMD) Sponsored by: The FreeBSD Foundation MFC after: 1 week
* amd64 la57_trampoline: jump immediately after re-enabling pagingKonstantin Belousov2024-09-161-1/+2
| | | | | | | | | Literally follow requirements from SDM and execute jmp right after %cr0 CR0_PG bit is toggled back. Sponsored by: Advanced Micro Devices (AMD) Sponsored by: The FreeBSD Foundation MFC after: 1 week
* amd64 pmap: flush whole TLB after LA57 trampoline is installedKonstantin Belousov2024-09-161-1/+1
| | | | | | Sponsored by: Advanced Micro Devices (AMD) Sponsored by: The FreeBSD Foundation MFC after: 1 week
* amd64 pmap: be more verbose around entering and leaving LA57 trampolineKonstantin Belousov2024-09-161-0/+7
| | | | | | Sponsored by: Advanced Micro Devices (AMD) Sponsored by: The FreeBSD Foundation MFC after: 1 week
* rangeset: speed up range traversalDoug Moore2024-09-091-13/+7
| | | | | | | | | | For rangeset-next search, use exact search rather than greater-than search. Move a bit of the testing logic from the pmap code to the common rangeset code. Reviewed by: kib (previous version) Tested by: pho (previous version) Differential Revision: https://reviews.freebsd.org/D46314
* amd64: move the right parenthesis to the right placeWuyang Chung2024-09-061-1/+1
| | | | | Reviewed by: imp, emaste Pull Request: https://github.com/freebsd/freebsd-src/pull/1356
generated by cgit v1.2.3 (git 2.25.1) at 2026-03-28 20:12:01 +0000