linux.git/arch/x86/virt/svm, branch master

x86/sev, crypto/ccp: Move HSAVE_PA setup to arch/x86/

2026-03-29T17:59:58+00:00

Now that there is snp_prepare() that indicates when the CCP driver wants to
prepare the architecture for SNP_INIT(_EX), move this architecture-specific
bit of code to a more sensible place.

Signed-off-by: Tycho Andersen (AMD) 
Signed-off-by: Borislav Petkov (AMD) 
Reviewed-by: Tom Lendacky 
Link: https://patch.msgid.link/20260324161301.1353976-6-tycho@kernel.org

x86/sev, crypto/ccp: Move SNP init to ccp driver

2026-03-29T10:32:09+00:00

Use the new snp_prepare() to initialize SNP from the ccp driver instead of at
boot time. This means that SNP is not enabled unless it is really going to be
used (i.e. kvm_amd loads the ccp driver automatically).

Signed-off-by: Tycho Andersen (AMD) 
Signed-off-by: Borislav Petkov (AMD) 
Reviewed-by: Tom Lendacky 
Link: https://patch.msgid.link/20260324161301.1353976-5-tycho@kernel.org

x86/sev: Create snp_shutdown()

2026-03-29T10:15:17+00:00

After SNP_SHUTDOWN, two things should be done:

1. clear the RMP table
2. disable MFDM to prevent the FW_WARN in k8_check_syscfg_dram_mod_en() in
   the event of a kexec

Create and export to the CCP driver a function that does them.

Also change the MFDM helper to allow for disabling the bit, since the SNP x86
shutdown path needs to disable MFDM.

The comment for k8_check_syscfg_dram_mod_en() notes, the "BIOS" is supposed
clear it, or the kernel in the case of module unload and shutdown followed by
kexec.

Signed-off-by: Tycho Andersen (AMD) 
Signed-off-by: Borislav Petkov (AMD) 
Reviewed-by: Tom Lendacky 
Link: https://patch.msgid.link/20260324161301.1353976-4-tycho@kernel.org

x86/sev: Create snp_prepare()

2026-03-28T21:16:03+00:00

In preparation for delayed SNP initialization, create a function snp_prepare()
that does the necessary architecture setup.  Export this function for the ccp
module to allow it to do the setup as necessary.

Introduce a cpu_read_lock/unlock() wrapper around the MFDM and SNP enable.
While CPU hotplug is not supported, this makes sure that the bit setting
happens on the same set of CPUs in both cases.

This improvement was suggested by Sashiko:

  https://sashiko.dev/#/patchset/20260324161301.1353976-1-tycho%40kernel.org

Also move {mfd,snp}_enable() out of the __init section, since these will be
called later.

Signed-off-by: Tycho Andersen (AMD) 
Signed-off-by: Borislav Petkov (AMD) 
Reviewed-by: Tom Lendacky 
Link: https://patch.msgid.link/20260326161110.1764303-3-tycho@kernel.org

x86/sev: Create a function to clear/zero the RMP

2026-03-28T21:10:56+00:00

In preparation for delayed SNP initialization and disablement on shutdown,
create a function, clear_rmp(), that clears the RMP bookkeeping area and the
RMP entries.

Signed-off-by: Tom Lendacky 
Signed-off-by: Tycho Andersen (AMD) 
Signed-off-by: Borislav Petkov (AMD) 
Link: https://patch.msgid.link/20260324161301.1353976-2-tycho@kernel.org

x86/virt/sev: Keep the RMP table bookkeeping area mapped

2026-03-09T20:49:18+00:00

In preparation for delayed SNP initialization and disablement on shutdown, the
RMP will need to be cleared each time SNP is disabled. Maintain the mapping to
the RMP bookkeeping area to avoid mapping and unmapping it each time and any
possible errors that may arise from that.

Signed-off-by: Tom Lendacky 
Signed-off-by: Tycho Andersen (AMD) 
Signed-off-by: Borislav Petkov (AMD) 
Link: https://patch.msgid.link/20260309180053.2389118-4-tycho@kernel.org

x86/virt/sev: Drop WBINVD before setting MSR_AMD64_SYSCFG_SNP_EN

2026-03-09T20:48:46+00:00

WBINVD is required before SNP_INIT(_EX), but not before setting
MSR_AMD64_SYSCFG_SNP_EN, since the ccp driver already does its own WBINVD
before SNP_INIT (and this one would be too early for that anyway...).

Signed-off-by: Tycho Andersen (AMD) 
Signed-off-by: Borislav Petkov (AMD) 
Reviewed-by: Tom Lendacky 
Link: https://patch.msgid.link/20260309180053.2389118-3-tycho@kernel.org

x86/virt/sev: Drop support for SNP hotplug

2026-03-09T20:48:33+00:00

During an SNP_INIT(_EX), the SEV firmware checks that all CPUs have the SNP
syscfg bit set, and fails if they do not. As such, it does not make
sense to have offline CPUs: the firmware will fail initialization because
of the offlined ones that the kernel did not initialize.

Further, there is a bug: during SNP_INIT(_EX) the firmware requires the MFDM
syscfg bit to be set in addition to having SNP enabled, which the previous
hotplug code did not do. Since k8_check_syscfg_dram_mod_en() enforces this
be cleared, hotplug wouldn't work.

Signed-off-by: Tycho Andersen (AMD) 
Signed-off-by: Borislav Petkov (AMD) 
Reviewed-by: Tom Lendacky 
Link: https://patch.msgid.link/20260309180053.2389118-2-tycho@kernel.org

Convert 'alloc_obj' family to use the new default GFP_KERNEL argument

2026-02-22T01:09:51+00:00

This was done entirely with mindless brute force, using

    git grep -l '\

treewide: Replace kmalloc with kmalloc_obj for non-scalar types

2026-02-21T09:02:28+00:00

This is the result of running the Coccinelle script from
scripts/coccinelle/api/kmalloc_objs.cocci. The script is designed to
avoid scalar types (which need careful case-by-case checking), and
instead replace kmalloc-family calls that allocate struct or union
object instances:

Single allocations:	kmalloc(sizeof(TYPE), ...)
are replaced with:	kmalloc_obj(TYPE, ...)

Array allocations:	kmalloc_array(COUNT, sizeof(TYPE), ...)
are replaced with:	kmalloc_objs(TYPE, COUNT, ...)

Flex array allocations:	kmalloc(struct_size(PTR, FAM, COUNT), ...)
are replaced with:	kmalloc_flex(*PTR, FAM, COUNT, ...)

(where TYPE may also be *VAR)

The resulting allocations no longer return "void *", instead returning
"TYPE *".

Signed-off-by: Kees Cook