* Re: [PATCH v3 0/3] tracing: Read user data from futex system call trace event
From: Peter Zijlstra @ 2026-04-01 20:13 UTC (permalink / raw)
To: Thomas Gleixner
Cc: Steven Rostedt, linux-kernel, linux-trace-kernel,
Masami Hiramatsu, Mark Rutland, Mathieu Desnoyers, Andrew Morton,
Brian Geffon, John Stultz, Ian Rogers, Suleiman Souhlal
In-Reply-To: <87zf3m7a0o.ffs@tglx>
On Wed, Apr 01, 2026 at 09:31:19PM +0200, Thomas Gleixner wrote:
> Stick 100 lines of python into tools/tracing and be done with it. I'm
> happy to contribute to that.
>
> Aside of that:
>
> Putting the decoder (futex_print_syscall) into the futex code itself
> is admittedly a smart move to offload the work of keeping that up to
> date to the people who are actually working on futexes.
>
> TBH, I'm not interested to deal with that at all. If you want this
> ftrace magic pretty printing, then stick it into kernel/trace or if
> there is a real technical reason (hint there is none) into
> kernel/futex/trace.c and take ownership of it. But please do not burden
> others with your fancy toy of the day.
Him putting it near futex is my fault. I really hate having this
description nonsense separate -- *IF* is has to exist at all.
That said, I was thinking of Sasha's syscall api endeavour, I feel that
if we do any of this, it should all come from the same one place.
^ permalink raw reply
* Re: [PATCH v3 0/3] tracing: Read user data from futex system call trace event
From: Steven Rostedt @ 2026-04-01 20:19 UTC (permalink / raw)
To: Thomas Gleixner
Cc: linux-kernel, linux-trace-kernel, Masami Hiramatsu, Mark Rutland,
Mathieu Desnoyers, Andrew Morton, Peter Zijlstra, Brian Geffon,
John Stultz, Ian Rogers, Suleiman Souhlal
In-Reply-To: <87zf3m7a0o.ffs@tglx>
On Wed, 01 Apr 2026 21:31:19 +0200
Thomas Gleixner <tglx@linutronix.de> wrote:
> On Tue, Mar 31 2026 at 14:13, Steven Rostedt wrote:
> > We are looking at the performance of futexes and require a bit more
> > information when tracing them.
> >
> > The two patches here extend the system call reading of user space to
>
> s/two/three/ :)
Ah v1 had only two patches and this was cut and pasted from there.
>
> I understand what you are trying to achieve, but do we really need all
> the complexity of decoding and pretty printing in the kernel?
You could say the same for most tracepoints. ;-)
>
> Isn't it sufficient to store and expose the raw data and use post
> processing to make it readable?
Yes this is possible, and will also work too, as libtraceevent will be
updated to parse the raw data.
>
> I've been doing complex futex analysis for two decades with a small set
> of python scripts which translate raw text or binary trace data into
> human readable information.
>
> I agree that it's useful to have the actual timeout value and other data
> which is missing today, but that still does not require all this
> customized printing.
>
> The initial idea of having at least some information about the data
> entry (type, meaning etc.) in $event/format and use that for kernel text
> output and for user space tools to analyze a binary trace has been
> definitely the right way to go.
>
> But that now deviates because $event/format cannot carry that
> information you translate to in the kernel. It will still describe raw
> event data, no?
It still shows a bit:
name: sys_enter_futex
ID: 592
format:
field:unsigned short common_type; offset:0; size:2; signed:0;
field:unsigned char common_flags; offset:2; size:1; signed:0;
field:unsigned char common_preempt_count; offset:3; size:1; signed:0;
field:int common_pid; offset:4; size:4; signed:1;
field:int __syscall_nr; offset:8; size:4; signed:1;
field:u32 * uaddr; offset:16; size:8; signed:0;
field:int op; offset:24; size:8; signed:0;
field:u32 val; offset:32; size:8; signed:0;
field:const struct __kernel_timespec * utime; offset:40; size:8; signed:0;
field:u32 * uaddr2; offset:48; size:8; signed:0;
field:u32 val3; offset:56; size:8; signed:0;
field:u32 __value; offset:64; size:4; signed:0;
field:u32 __value2; offset:68; size:4; signed:0;
field:unsigned long __ts1; offset:72; size:8; signed:0;
field:unsigned long __ts2; offset:80; size:8; signed:0;
print fmt: "uaddr: 0x%lx (0x%lx) cmd=%s%s%s val: 0x%x timeout/val2: 0x%llx (%lu.%lu) uaddr2: 0x%lx (0x%lx) val3: 0x%x", REC->uaddr, REC->__value, __print_symbolic(REC->op & 0xfffffe7f, {0, "FUTEX_WAIT"} ,{1, "FUTEX_WAKE"} ,{2, "FUTEX_FD"} ,{3, "FUTEX_REQUEUE"} ,{4, "FUTEX_CMP_REQUEUE"} ,{5, "FUTEX_WAKE_OP"} ,{6, "FUTEX_LOCK_PI"} ,{7, "FUTEX_UNLOCK_PI"} ,{8, "FUTEX_TRYLOCK_PI"} ,{9, "FUTEX_WAIT_BITSET"} ,{10, "FUTEX_WAKE_BITSET"} ,{11, "FUTEX_WAIT_REQUEUE_PI"} ,{12, "FUTEX_CMP_REQUEUE_PI"} ,{13, "FUTEX_LOCK_PI2"} ), (REC->op & 128) ? "|FUTEX_PRIVATE_FLAG" : "", (REC->op & 256) ? "|FUTEX_CLOCK_REALTIME" : "", REC->val, REC->utime, REC->__ts1, REC->__ts2, REC->uaddr, REC->__value2, REC->val3
>
> So why not keeping the well known and working solution of identifying
> the data in the format, print it raw and leave the post processing to
> user space tools in case there is a need.
>
> You actually make it harder to do development. Look at the patch series
> related to robust futexes:
>
> https://lore.kernel.org/lkml/20260330114212.927686587@kernel.org/
>
> So your decoding:
>
> > sys_futex(uaddr: 0x56196292e830 (0), FUTEX_WAKE|FUTEX_PRIVATE_FLAG)
>
> fails to decode the new flag and the usage of uaddr2 unless I go and add
> it in the first place _before_ working on the code. Right now it is just
> printing op as a hex value and it just works when a new bit is added.
>
> Stick 100 lines of python into tools/tracing and be done with it. I'm
> happy to contribute to that.
Well, it would be updated for trace-cmd not tools/tracing.
>
> Aside of that:
>
> Putting the decoder (futex_print_syscall) into the futex code itself
> is admittedly a smart move to offload the work of keeping that up to
> date to the people who are actually working on futexes.
>
> TBH, I'm not interested to deal with that at all. If you want this
> ftrace magic pretty printing, then stick it into kernel/trace or if
> there is a real technical reason (hint there is none) into
> kernel/futex/trace.c and take ownership of it. But please do not burden
> others with your fancy toy of the day.
v1 kept it all within the tracing subsystem, but Peter suggested that it be
closer to the syscall:
https://lore.kernel.org/all/20260304090748.GO606826@noisy.programming.kicks-ass.net/
I'm happy to put it back and maintain it separately.
Or I can just keep the simple bits (the reading of user space), and not do
all the more fancy formatting. Basically dropping patch 2 and 3.
I've been using trace-cmd start / show for testing. But I could also move
the logic to libtraceevent, which would require using trace-cmd record
instead.
How much are you against the full series? Are you OK with it if it stays
within the tracing subsystem? Or would you prefer just keeping with patch 1
and dropping the other patches and doing that work in libtraceevent?
-- Steve
^ permalink raw reply
* Re: [PATCH v3 0/3] tracing: Read user data from futex system call trace event
From: Peter Zijlstra @ 2026-04-01 20:25 UTC (permalink / raw)
To: Steven Rostedt
Cc: Thomas Gleixner, linux-kernel, linux-trace-kernel,
Masami Hiramatsu, Mark Rutland, Mathieu Desnoyers, Andrew Morton,
Brian Geffon, John Stultz, Ian Rogers, Suleiman Souhlal
In-Reply-To: <20260401161919.147355bf@gandalf.local.home>
On Wed, Apr 01, 2026 at 04:19:19PM -0400, Steven Rostedt wrote:
> Well, it would be updated for trace-cmd not tools/tracing.
So I tried using that a while ago:
https://lkml.kernel.org/r/20260116181524.GF831285@noisy.programming.kicks-ass.net
and I found it really rough.
^ permalink raw reply
* Re: [PATCH RFC v4 08/44] KVM: Introduce KVM_SET_MEMORY_ATTRIBUTES2
From: Sean Christopherson @ 2026-04-01 21:04 UTC (permalink / raw)
To: Michael Roth
Cc: Ackerley Tng, aik, andrew.jones, binbin.wu, brauner, chao.p.peng,
david, ira.weiny, jmattson, jroedel, jthoughton, oupton,
pankaj.gupta, qperret, rick.p.edgecombe, rientjes, shivankg,
steven.price, tabba, willy, wyihan, yan.y.zhao, forkloop,
pratyush, suzuki.poulose, aneesh.kumar, Paolo Bonzini,
Thomas Gleixner, Ingo Molnar, Borislav Petkov, Dave Hansen, x86,
H. Peter Anvin, Steven Rostedt, Masami Hiramatsu,
Mathieu Desnoyers, Jonathan Corbet, Shuah Khan, Shuah Khan,
Vishal Annapurve, Andrew Morton, Chris Li, Kairui Song,
Kemeng Shi, Nhat Pham, Baoquan He, Barry Song, Axel Rasmussen,
Yuanchu Xie, Wei Xu, Jason Gunthorpe, Vlastimil Babka, kvm,
linux-kernel, linux-trace-kernel, linux-doc, linux-kselftest,
linux-mm
In-Reply-To: <cxmtst7txfodp6lo4ipue3trohx2ge7nkagkfzfixfdnlf5qlo@e3jw3tmbe272>
On Tue, Mar 31, 2026, Michael Roth wrote:
> On Thu, Mar 26, 2026 at 03:24:17PM -0700, Ackerley Tng wrote:
> > Introduce a "version 2" of KVM_SET_MEMORY_ATTRIBUTES to support returning
> > information back to userspace.
>
> Hi Ackerley,
>
> Not trying to bikeshed below, but I'm working on getting related QEMU
> patches cleaned up to post soon and was working through some of the new
> uAPI bits, and plumbing some of these capabilities in seems a little
> awkward in a couple places so wondering if we should revisit how some of
> this API is defined...
>
> >
> > This new ioctl and structure will, in a later patch, be shared as a
> > guest_memfd ioctl, where the padding in the new kvm_memory_attributes2
> > structure will be for writing the response from the guest_memfd ioctl to
> > userspace.
> >
> > A new ioctl is necessary for these reasons:
> >
> > 1. KVM_SET_MEMORY_ATTRIBUTES is currently a write-only ioctl and does not
> > allow userspace to read fields. There's nothing in code (yet?) that
> > validates this, but using _IOWR for consistency would be prudent.
> >
> > 2. KVM_SET_MEMORY_ATTRIBUTES, when used as a guest_memfd ioctl, will need
> > an additional field to provide userspace with more error details.
> >
> > Alternatively, a completely new ioctl could be defined, unrelated to
> > KVM_SET_MEMORY_ATTRIBUTES, but using the same ioctl number and struct for
> > the vm and guest_memfd ioctls streamlines the interface for userspace. In
> > addition, any memory attributes, implemented on the vm or guest_memfd
> > ioctl, can be easily shared with the other.
> >
> > Add KVM_CAP_MEMORY_ATTRIBUTES2 to indicate that struct
> > kvm_memory_attributes2 exists and can be used either with
> > KVM_SET_MEMORY_ATTRIBUTES2 via the vm or guest_memfd ioctl.
>
> The guest_memfd support for the KVM_SET_MEMORY_ATTRIBUTES2 ioctl isn't
> added until patch #10, and to scan for it you sort of need to infer it
> via KVM_CAP_GUEST_MEMFD_MEMORY_ATTRIBUTES reporting non-zero (i.e.
> KVM_MEMORY_ATTRIBUTE_PRIVATE), so it's confusing to state that
> KVM_CAP_MEMORY_ATTRIBUTES2 means you can use the struct via a guest_memfd
> ioctl.
>
> I think the above is trying to simply say that the corresponding struct
> exists, and remain agnostic about how it can be used. But if that were
> the case, there would be no way to know when KVM_SET_MEMORY_ATTRIBUTES2 is
> available in the first place, so in the case of KVM ioctls at least,
> KVM_CAP_MEMORY_ATTRIBUTES2 is advertising both the struct and the ioctl,
> whereas for guest_memfd it's only advertising the struct and not saying
> anything about whether a similar gmem ioctl is available to use it.
+1 to everything Mike said.
> Instead, maybe they should both have the same semantics:
>
> KVM_CAP_MEMORY_ATTRIBUTES2: *SET_ATTRIBUTES* ioctl exists for KVM that utilizes
> struct kvm_memory_attributes2
>
> KVM_CAP_GUEST_MEMFD_MEMORY_ATTRIBUTES: *SET_ATTRIBUTES* ioctl exists for
> guest_memfd that utilizes struct kvm_memory_attributes2
>
> In which case you would leave out any mention of guest_memfd here as far as
> the documentation does, and then in patch #10 you could modify it to be
> something like:
>
> 4.145 KVM_SET_MEMORY_ATTRIBUTES2
> ---------------------------------
>
> -:Capability: KVM_CAP_MEMORY_ATTRIBUTES2
> +:Capability: KVM_CAP_MEMORY_ATTRIBUTES2, KVM_GUEST_MEMFD_CAP_MEMORY_ATTRIBUTES
> -:Architectures: x86
> +:Architectures: all
> -:Type: vm ioctl
> +:Type: vm, guest_memfd ioctl
> :Parameters: struct kvm_memory_attributes2 (in/out)
> :Returns: 0 on success, <0 on error
As discussed at PUCK, I think we should omit KVM_CAP_MEMORY_ATTRIBUTES2 and
vm-scoped support entirely until it's needed (which may be never).
> and *then* add in your mentions of how the usage/fields differ for
> guest_memfd/KVM_GUEST_MEMFD_CAP_MEMORY_ATTRIBUTES case vs. KVM ioctls.
>
> This avoids needing to issue 2 checks for the guest_memfd variant vs. 1
> for KVM, but more importantly avoids subtle differences in how these
> similarly-named capabilities are used/documented that might cause
> unecessary confusion.
^ permalink raw reply
* Re: [PATCH RFC v4 25/44] KVM: selftests: Test basic single-page conversion flow
From: Sean Christopherson @ 2026-04-01 21:08 UTC (permalink / raw)
To: Ackerley Tng
Cc: aik, andrew.jones, binbin.wu, brauner, chao.p.peng, david,
ira.weiny, jmattson, jroedel, jthoughton, michael.roth, oupton,
pankaj.gupta, qperret, rick.p.edgecombe, rientjes, shivankg,
steven.price, tabba, willy, wyihan, yan.y.zhao, forkloop,
pratyush, suzuki.poulose, aneesh.kumar, Paolo Bonzini,
Thomas Gleixner, Ingo Molnar, Borislav Petkov, Dave Hansen, x86,
H. Peter Anvin, Steven Rostedt, Masami Hiramatsu,
Mathieu Desnoyers, Jonathan Corbet, Shuah Khan, Shuah Khan,
Vishal Annapurve, Andrew Morton, Chris Li, Kairui Song,
Kemeng Shi, Nhat Pham, Baoquan He, Barry Song, Axel Rasmussen,
Yuanchu Xie, Wei Xu, Jason Gunthorpe, Vlastimil Babka, kvm,
linux-kernel, linux-trace-kernel, linux-doc, linux-kselftest,
linux-mm
In-Reply-To: <CAEvNRgE6Tn81Yddgbjqs-gs491NzpppjbDHKzpmdPCxgSPeUPQ@mail.gmail.com>
On Tue, Mar 31, 2026, Ackerley Tng wrote:
> Ackerley Tng <ackerleytng@google.com> writes:
Please trim your replies (even more, since you did trim a little).
> > +static void run_guest_do_rmw(struct kvm_vcpu *vcpu, loff_t pgoff,
> > + char expected_val, char write_val)
> > +{
> > + struct ucall uc;
> > + int r;
> > +
> > + guest_data.mem = (void *)GUEST_MEMFD_SHARING_TEST_GVA + pgoff * page_size;
> > + guest_data.expected_val = expected_val;
> > + guest_data.write_val = write_val;
> > + sync_global_to_guest(vcpu->vm, guest_data);
> > +
> > + do {
> > + r = __vcpu_run(vcpu);
> > + } while (r == -1 && errno == EINTR);
> > +
> > + TEST_ASSERT_EQ(r, 0);
>
> TEST_ASSERT_EQ() ends up calling exit() on failures, which skips
> FIXTURE_TEARDOWN().
>
> Other than the explicit assertions not working with the
> kselftest_harness, kvm selftest library functions like vm_mem_add() also
> call TEST_ASSERT, which doesn't play nice with kselftest_harness.
>
> Any suggestions for this? Should we use the kselftest framework with
> these tests?
>
> (I ran into this issue while trying to test something else, where I
> needed FIXTURE_TEARDOWN() to clean up system state.)
>
> Or is it "okay" in this case since FIXTURE_TEARDOWN() only cleans up
> stuff that would happen if the program exits anyway?
Can you see if any of the ideas in https://lore.kernel.org/all/ZjUwqEXPA5QVItyX@google.com
would help? Converting more tests to TAP+FIXTURE is still on my wish list, I've
just never been able to carve out cycles to see it through.
^ permalink raw reply
* Re: [PATCH RFC v4 10/44] KVM: guest_memfd: Add support for KVM_SET_MEMORY_ATTRIBUTES2
From: Sean Christopherson @ 2026-04-01 21:12 UTC (permalink / raw)
To: Michael Roth
Cc: Ackerley Tng, aik, andrew.jones, binbin.wu, brauner, chao.p.peng,
david, ira.weiny, jmattson, jroedel, jthoughton, oupton,
pankaj.gupta, qperret, rick.p.edgecombe, rientjes, shivankg,
steven.price, tabba, willy, wyihan, yan.y.zhao, forkloop,
pratyush, suzuki.poulose, aneesh.kumar, Paolo Bonzini,
Thomas Gleixner, Ingo Molnar, Borislav Petkov, Dave Hansen, x86,
H. Peter Anvin, Steven Rostedt, Masami Hiramatsu,
Mathieu Desnoyers, Jonathan Corbet, Shuah Khan, Shuah Khan,
Vishal Annapurve, Andrew Morton, Chris Li, Kairui Song,
Kemeng Shi, Nhat Pham, Baoquan He, Barry Song, Axel Rasmussen,
Yuanchu Xie, Wei Xu, Jason Gunthorpe, Vlastimil Babka, kvm,
linux-kernel, linux-trace-kernel, linux-doc, linux-kselftest,
linux-mm
In-Reply-To: <2r4mmfiuisw26qymahnbh2oxqkkrywqev477kc4rlkcyx7tels@c7ple7kdgpo3>
On Wed, Apr 01, 2026, Michael Roth wrote:
> On Thu, Mar 26, 2026 at 03:24:19PM -0700, Ackerley Tng wrote:
> > #ifdef CONFIG_KVM_VM_MEMORY_ATTRIBUTES
> > static unsigned long kvm_get_vm_memory_attributes(struct kvm *kvm, gfn_t gfn)
> > {
> > @@ -2635,6 +2625,8 @@ static int kvm_vm_ioctl_set_mem_attributes(struct kvm *kvm,
> > return -EINVAL;
> > if (!PAGE_ALIGNED(attrs->address) || !PAGE_ALIGNED(attrs->size))
> > return -EINVAL;
> > + if (attrs->error_offset)
> > + return -EINVAL;
> > for (i = 0; i < ARRAY_SIZE(attrs->reserved); i++) {
> > if (attrs->reserved[i])
> > return -EINVAL;
> > @@ -4983,6 +4975,11 @@ static int kvm_vm_ioctl_check_extension_generic(struct kvm *kvm, long arg)
> > return 1;
> > case KVM_CAP_GUEST_MEMFD_FLAGS:
> > return kvm_gmem_get_supported_flags(kvm);
> > + case KVM_CAP_GUEST_MEMFD_MEMORY_ATTRIBUTES:
> > + if (vm_memory_attributes)
> > + return 0;
> > +
> > + return kvm_supported_mem_attributes(kvm);
>
> Based on the discussion from the PUCK call this morning, it sounds like it
> would be a good idea to limit kvm_supported_mem_attributes() to only
> reporting KVM_MEMORY_ATTRIBUTE_PRIVATE if the underlying CoCo
> implementation has all the necessary enablement to support in-place
> conversion via guest_memfd. In the case of SNP, there is a
> documentation/parameter check in snp_launch_update() that needs to be
> relaxed in order for userspace to be able to pass in a NULL 'src'
> parameter (since, for in-place conversion, it would be initialized in place
> as shared memory prior to the call, since by the time kvm_gmem_poulate()
> it will have been set to private and therefore cannot be faulted in via
> GUP (and if it could, we'd be unecessarily copying the src back on top
> of itself since src/dst are the same).
>
> So maybe there should be an arch hook to check a whitelist of VM types
> that support KVM_MEMORY_ATTRIBUTE_PRIVATE when vm_memory_attributes=0,
> and if we decide to enable it for SNP as part of this series you could
> include the 1-2 patches needed there, or I could enable the SNP support
> separately as a small series and I guess that would then become a prereq
> for the SNP self-tests?
If it's trivial-ish, my preference would be to include SNP as part of this series,
_before_ KVM_CAP_GUEST_MEMFD_MEMORY_ATTRIBUTES is exposed to userspace. I think
that would avoid the need for pivoting on the VM type? I.e. don't advertise
support until all VM types play nice.
> Not sure if additional enablement is needed for TDX or not before
> KVM_MEMORY_ATTRIBUTE_PRIVATE would be advertised, but similar
> considerations there.
^ permalink raw reply
* Re: [PATCH RFC v4 07/44] KVM: guest_memfd: Only prepare folios for private pages
From: Ackerley Tng @ 2026-04-01 21:43 UTC (permalink / raw)
To: Michael Roth
Cc: aik, andrew.jones, binbin.wu, brauner, chao.p.peng, david,
ira.weiny, jmattson, jroedel, jthoughton, oupton, pankaj.gupta,
qperret, rick.p.edgecombe, rientjes, shivankg, steven.price,
tabba, willy, wyihan, yan.y.zhao, forkloop, pratyush,
suzuki.poulose, aneesh.kumar, Paolo Bonzini, Sean Christopherson,
Thomas Gleixner, Ingo Molnar, Borislav Petkov, Dave Hansen, x86,
H. Peter Anvin, Steven Rostedt, Masami Hiramatsu,
Mathieu Desnoyers, Jonathan Corbet, Shuah Khan, Shuah Khan,
Vishal Annapurve, Andrew Morton, Chris Li, Kairui Song,
Kemeng Shi, Nhat Pham, Baoquan He, Barry Song, Axel Rasmussen,
Yuanchu Xie, Wei Xu, Jason Gunthorpe, Vlastimil Babka, kvm,
linux-kernel, linux-trace-kernel, linux-doc, linux-kselftest,
linux-mm
In-Reply-To: <s4dbqrv2c6yzt4nsflfarnggtl25xlz6mzg74tfeg3eskceno6@6l5hpfmcbju3>
Michael Roth <michael.roth@amd.com> writes:
> On Wed, Apr 01, 2026 at 07:05:16AM -0700, Ackerley Tng wrote:
>> Ackerley Tng <ackerleytng@google.com> writes:
>>
>> > All-shared guest_memfd used to be only supported for non-CoCo VMs where
>> > preparation doesn't apply. INIT_SHARED is about to be supported for
>> > non-CoCo VMs in a later patch in this series.
>> >
>> > In addition, KVM_SET_MEMORY_ATTRIBUTES2 is about to be supported in
>> > guest_memfd in a later patch in this series.
>> >
>> > This means that the kvm fault handler may now call kvm_gmem_get_pfn() on a
>> > shared folio for a CoCo VM where preparation applies.
>> >
>> > Add a check to make sure that preparation is only performed for private
>> > folios.
>> >
>> > Preparation will be undone on freeing (see kvm_gmem_free_folio()) and on
>> > conversion to shared.
>> >
>> > Signed-off-by: Ackerley Tng <ackerleytng@google.com>
>> > ---
>> > virt/kvm/guest_memfd.c | 9 ++++++---
>> > 1 file changed, 6 insertions(+), 3 deletions(-)
>> >
>> > diff --git a/virt/kvm/guest_memfd.c b/virt/kvm/guest_memfd.c
>> > index b6ffa8734175d..d414ebfcb4c19 100644
>> > --- a/virt/kvm/guest_memfd.c
>> > +++ b/virt/kvm/guest_memfd.c
>> > @@ -900,6 +900,7 @@ int kvm_gmem_get_pfn(struct kvm *kvm, struct kvm_memory_slot *slot,
>> > int *max_order)
>> > {
>> > pgoff_t index = kvm_gmem_get_index(slot, gfn);
>> > + struct inode *inode;
>> > struct folio *folio;
>> > int r = 0;
>> >
>> > @@ -907,7 +908,8 @@ int kvm_gmem_get_pfn(struct kvm *kvm, struct kvm_memory_slot *slot,
>> > if (!file)
>> > return -EFAULT;
>> >
>> > - filemap_invalidate_lock_shared(file_inode(file)->i_mapping);
>> > + inode = file_inode(file);
>> > + filemap_invalidate_lock_shared(inode->i_mapping);
>> >
>> > folio = __kvm_gmem_get_pfn(file, slot, index, pfn, max_order);
>> > if (IS_ERR(folio)) {
>> > @@ -920,7 +922,8 @@ int kvm_gmem_get_pfn(struct kvm *kvm, struct kvm_memory_slot *slot,
>> > folio_mark_uptodate(folio);
>> > }
>> >
>> > - r = kvm_gmem_prepare_folio(kvm, slot, gfn, folio);
>> > + if (kvm_gmem_is_private_mem(inode, index))
>> > + r = kvm_gmem_prepare_folio(kvm, slot, gfn, folio);
>>
>> Michael, I might have misunderstood you at the last guest_memfd call:
>> sev_gmem_prepare() doesn't prepare a page for being a shared page,
>> right? Does this work? That prepare is only called to "make private"?
>
> Hmm, I guess your guest_memfd-inplace-conversion-v4 branch is out of sync with
> these patches?
>
My bad, it was. I just force-pushed to github to synchronize them with
this patch series.
> I have the below local patch based on top of that for SNP-specific enablement,
> which is basically identically, so suffice to say: yes, this should work
> for SNP :) If any architecture pops up that needs to do some prep in
> advance of mapping shared pages, then we could potentially plumb the
> shared/private flag through to the arch-specific prep hook, as was also
> suggested on the call, but it doesn't seem like that's needed by any
> users for now.
>
Thanks for checking :)
> -Mike
>
> Author: Michael Roth <michael.roth@amd.com>
> Date: Mon Oct 27 07:58:32 2025 -0500
>
> KVM: guest_memfd: Don't prepare shared folios
>
> In the current guest_memfd logic, "preparation" is only used currently
> to describe the additional work of putting a guest_memfd page into an
> architecturally-defined "private" state, such as updating RMP table
> entries for SEV-SNP guests. As such, there's no input to the
> corresponding kvm_arch_gmem_prepare() hooks as to whether a page is
> being prepared/accessed as shared or as private, so "preparation" will
> end up being erroneously done on pages that were supposed to remain in a
> shared state. Rather than plumb through the additional information
> needed to distinguish between shared vs. private preparation, just
> continue to only do preparation on private pages, as was the case prior
> to support for GUEST_MEMFD_FLAG_MMAP being introduced.
>
> Signed-off-by: Michael Roth <michael.roth@amd.com>
>
> diff --git a/virt/kvm/guest_memfd.c b/virt/kvm/guest_memfd.c
> index 3acc6d983449..4869e59e4fc5 100644
> --- a/virt/kvm/guest_memfd.c
> +++ b/virt/kvm/guest_memfd.c
> @@ -1249,7 +1249,8 @@ int kvm_gmem_get_pfn(struct kvm *kvm, struct kvm_memory_slot *slot,
> folio_mark_uptodate(folio);
> }
>
> - r = kvm_gmem_prepare_folio(kvm, slot, gfn, folio);
> + if (!kvm_gmem_is_shared_mem(file_inode(file), index))
> + r = kvm_gmem_prepare_folio(kvm, slot, gfn, folio);
>
> folio_unlock(folio);
>
>>
>> >
>> > folio_unlock(folio);
>> >
>> > @@ -930,7 +933,7 @@ int kvm_gmem_get_pfn(struct kvm *kvm, struct kvm_memory_slot *slot,
>> > folio_put(folio);
>> >
>> > out:
>> > - filemap_invalidate_unlock_shared(file_inode(file)->i_mapping);
>> > + filemap_invalidate_unlock_shared(inode->i_mapping);
>> > return r;
>> > }
>> > EXPORT_SYMBOL_FOR_KVM_INTERNAL(kvm_gmem_get_pfn);
>> >
>> > --
>> > 2.53.0.1018.g2bb0e51243-goog
^ permalink raw reply
* Re: [PATCH RFC v4 10/44] KVM: guest_memfd: Add support for KVM_SET_MEMORY_ATTRIBUTES2
From: Ackerley Tng @ 2026-04-01 22:38 UTC (permalink / raw)
To: Michael Roth
Cc: aik, andrew.jones, binbin.wu, brauner, chao.p.peng, david,
ira.weiny, jmattson, jroedel, jthoughton, oupton, pankaj.gupta,
qperret, rick.p.edgecombe, rientjes, shivankg, steven.price,
tabba, willy, wyihan, yan.y.zhao, forkloop, pratyush,
suzuki.poulose, aneesh.kumar, Paolo Bonzini, Sean Christopherson,
Thomas Gleixner, Ingo Molnar, Borislav Petkov, Dave Hansen, x86,
H. Peter Anvin, Steven Rostedt, Masami Hiramatsu,
Mathieu Desnoyers, Jonathan Corbet, Shuah Khan, Shuah Khan,
Vishal Annapurve, Andrew Morton, Chris Li, Kairui Song,
Kemeng Shi, Nhat Pham, Baoquan He, Barry Song, Axel Rasmussen,
Yuanchu Xie, Wei Xu, Jason Gunthorpe, Vlastimil Babka, kvm,
linux-kernel, linux-trace-kernel, linux-doc, linux-kselftest,
linux-mm
In-Reply-To: <2r4mmfiuisw26qymahnbh2oxqkkrywqev477kc4rlkcyx7tels@c7ple7kdgpo3>
Michael Roth <michael.roth@amd.com> writes:
>
> [...snip...]
>
>> static unsigned long kvm_get_vm_memory_attributes(struct kvm *kvm, gfn_t gfn)
>> {
>> @@ -2635,6 +2625,8 @@ static int kvm_vm_ioctl_set_mem_attributes(struct kvm *kvm,
>> return -EINVAL;
>> if (!PAGE_ALIGNED(attrs->address) || !PAGE_ALIGNED(attrs->size))
>> return -EINVAL;
>> + if (attrs->error_offset)
>> + return -EINVAL;
>> for (i = 0; i < ARRAY_SIZE(attrs->reserved); i++) {
>> if (attrs->reserved[i])
>> return -EINVAL;
>> @@ -4983,6 +4975,11 @@ static int kvm_vm_ioctl_check_extension_generic(struct kvm *kvm, long arg)
>> return 1;
>> case KVM_CAP_GUEST_MEMFD_FLAGS:
>> return kvm_gmem_get_supported_flags(kvm);
>> + case KVM_CAP_GUEST_MEMFD_MEMORY_ATTRIBUTES:
>> + if (vm_memory_attributes)
>> + return 0;
>> +
>> + return kvm_supported_mem_attributes(kvm);
>
> Based on the discussion from the PUCK call this morning,
Thanks for copying the discussion here, I'll start attending PUCK to
catch those discussions too :)
> it sounds like it
> would be a good idea to limit kvm_supported_mem_attributes() to only
> reporting KVM_MEMORY_ATTRIBUTE_PRIVATE if the underlying CoCo
> implementation has all the necessary enablement to support in-place
> conversion via guest_memfd. In the case of SNP, there is a
> documentation/parameter check in snp_launch_update() that needs to be
> relaxed in order for userspace to be able to pass in a NULL 'src'
> parameter (since, for in-place conversion, it would be initialized in place
> as shared memory prior to the call, since by the time kvm_gmem_poulate()
> it will have been set to private and therefore cannot be faulted in via
> GUP (and if it could, we'd be unecessarily copying the src back on top
> of itself since src/dst are the same).
Could this be a separate thing? If I'm understanding you correctly, it's
not strictly a requirement for snp_launch_update() to first support a
NULL 'src' parameter before this series lands.
Without this series, the startup procedure is to have memory set up in
non-guest_memfd shared memory, and then snp_launch_update()-ed into
guest_memfd private memory.
With this series, it is a little troublesome, but the startup procedure
can still set up memory in guest_memfd shared memory, then copy
everything out to some temporary memory, then set guest_memfd memory to
private, then snp_launch_update() the temporary memory into guest_memfd
private memory.
We would be unnecessarily copying the src (now in some temporary memory)
back onto itself. Can that be a separate patch series?
Btw, if snp_launch_update() is going to accept a NULL src parameter and
launch-update the src in-place:
+ Will userspace have to set that memory to private before calling launch
update?
+ If yes, then would we need some other mode of conversion that is
not ZERO and not quite PRESERVE (since PRESERVE is defined as that
the guest will see what the host wrote post-encryption, but it
sounds like launch update is doing the encryption)
+ Or should launch update be called when that memory is shared? Will
launch update then also set that memory to private in guest_memfd?
>
> So maybe there should be an arch hook to check a whitelist of VM types
> that support KVM_MEMORY_ATTRIBUTE_PRIVATE when vm_memory_attributes=0,
> and if we decide to enable it for SNP as part of this series you could
> include the 1-2 patches needed there, or I could enable the SNP support
> separately as a small series and I guess that would then become a prereq
> for the SNP self-tests?
>
> Not sure if additional enablement is needed for TDX or not before
> KVM_MEMORY_ATTRIBUTE_PRIVATE would be advertised, but similar
> considerations there.
>
> -Mike
>
>> #endif
>> default:
>> break;
>>
>> --
>> 2.53.0.1018.g2bb0e51243-goog
>>
^ permalink raw reply
* Re: [PATCH v15 4/5] ring-buffer: Add persistent ring buffer invalid-page inject test
From: Steven Rostedt @ 2026-04-01 22:40 UTC (permalink / raw)
To: Masami Hiramatsu (Google)
Cc: Catalin Marinas, Will Deacon, Mathieu Desnoyers, linux-kernel,
linux-trace-kernel, Ian Rogers, linux-arm-kernel
In-Reply-To: <177494619065.71933.9842685686800241005.stgit@mhiramat.tok.corp.google.com>
I replied with mostly grammar fixes and some rewrites of text.
On Tue, 31 Mar 2026 17:36:30 +0900
"Masami Hiramatsu (Google)" <mhiramat@kernel.org> wrote:
> From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
>
> Add a self-destractive test for the persistent ring buffer.
"self-destractive"? Do you mean "self-destructive"?
Probably better to call it a "self-corrupting test".
>
> This will inject erroneous value to some sub-buffer pages (where
inject an erroneous
> the index is even or multiples of 5) in the persistent ring buffer
> when kernel gets panic, and check whether the number of detected
when the kernel panics, and checks whether
> invalid pages and the total entry_bytes are the same as recorded
same as the recorded
> values after reboot.
>
> This can ensure the kernel correctly recover partially corrupted
This ensures that the kernel can correctly recover a partially corrupted
> persistent ring buffer when boot.
after a reboot or panic.
>
> The test only runs on the persistent ring buffer whose name is
> "ptracingtest". And user has to fill it up with events before
The user has to fill it with events before a kernel panic.
> kernel panics.
>
> To run the test, enable CONFIG_RING_BUFFER_PERSISTENT_INJECT
> and you have to setup the kernel cmdline;
and add the following kernel cmdline:
Note, it's more proper to use a colon (:) than a semi-colon (;) when
referencing an example on the next line.
>
> reserve_mem=20M:2M:trace trace_instance=ptracingtest^traceoff@trace
> panic=1
>
> And run following commands after the 1st boot;
Run the following commands after the 1st boot:
>
> cd /sys/kernel/tracing/instances/ptracingtest
> echo 1 > tracing_on
> echo 1 > events/enable
> sleep 3
> echo c > /proc/sysrq-trigger
>
> After panic message, the kernel will reboot and run the verification
> on the persistent ring buffer, e.g.
>
> Ring buffer meta [2] invalid buffer page detected
> Ring buffer meta [2] is from previous boot! (318 pages discarded)
> Ring buffer testing [2] invalid pages: PASSED (318/318)
> Ring buffer testing [2] entry_bytes: PASSED (1300476/1300476)
>
> Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
> ---
> Changes in v15:
> - Use pr_warn() for test result.
> - Inject errors on the page index is multiples of 5 so that
> this can reproduce contiguous empty pages.
> Changes in v14:
> - Rename config to CONFIG_RING_BUFFER_PERSISTENT_INJECT.
> - Clear meta->nr_invalid/entry_bytes after testing.
> - Add test commands in config comment.
> Changes in v10:
> - Add entry_bytes test.
> - Do not compile test code if CONFIG_RING_BUFFER_PERSISTENT_SELFTEST=n.
> Changes in v9:
> - Test also reader pages.
> ---
> include/linux/ring_buffer.h | 1 +
> kernel/trace/Kconfig | 31 ++++++++++++++++++
> kernel/trace/ring_buffer.c | 74 +++++++++++++++++++++++++++++++++++++++++++
> kernel/trace/trace.c | 4 ++
> 4 files changed, 110 insertions(+)
>
> diff --git a/include/linux/ring_buffer.h b/include/linux/ring_buffer.h
> index 994f52b34344..0670742b2d60 100644
> --- a/include/linux/ring_buffer.h
> +++ b/include/linux/ring_buffer.h
> @@ -238,6 +238,7 @@ int ring_buffer_subbuf_size_get(struct trace_buffer *buffer);
>
> enum ring_buffer_flags {
> RB_FL_OVERWRITE = 1 << 0,
> + RB_FL_TESTING = 1 << 1,
> };
>
> #ifdef CONFIG_RING_BUFFER
> diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig
> index e130da35808f..07305ed6d745 100644
> --- a/kernel/trace/Kconfig
> +++ b/kernel/trace/Kconfig
> @@ -1202,6 +1202,37 @@ config RING_BUFFER_VALIDATE_TIME_DELTAS
> Only say Y if you understand what this does, and you
> still want it enabled. Otherwise say N
>
> +config RING_BUFFER_PERSISTENT_INJECT
> + bool "Enable persistent ring buffer error injection test"
> + depends on RING_BUFFER
> + help
> + Run a selftest on the persistent ring buffer which names
> + "ptracingtest" (and its backup) when panic_on_reboot by
Does this do anything with the backup?
> + invalidating ring buffer pages.
This option will have the kernel check if the persistent ring
buffer is named "ptracingtest", and if so, it will corrupt some
of its pages on a kernel panic. This is used to test if the
persistent ring buffer can recover from some of its sub-buffers
being corrupted.
[space]
> + To use this, boot kernel with "ptracingtest" persistent
, boot a kernel with a "ptracingtest" persistent
> + ring buffer, e.g.
> +
> + reserve_mem=20M:2M:trace trace_instance=ptracingtest@trace panic=1
> +
> + And after the 1st boot, run test command, like;
, run the following commands:
> +
> + cd /sys/kernel/tracing/instances/ptracingtest
> + echo 1 > events/enable
> + echo 1 > tracing_on
> + sleep 3
> + echo c > /proc/sysrq-trigger
> +
> + After panic message, the kernel reboots and show test results
> + on the boot log.
After the panic message, the kernel will reboot and will show the
test results in the console output.
> +
> + Note that user has to enable events on the persistent ring
> + buffer manually to fill up ring buffers before rebooting.
Note that events for the ring buffer needs to be enabled prior to
crashing the kernel so that the ring buffer has content that the
test will corrupt.
> + Since this invalidates the data on test target ring buffer,
> + "ptracingtest" persistent ring buffer must not be used for
> + actual tracing, but only for testing.
As the test will corrupt events in the "ptracingtest" persistent
ring buffer, it should not be used for any other purpose other
than his test.
> +
> + If unsure, say N
> +
> config MMIOTRACE_TEST
> tristate "Test module for mmiotrace"
> depends on MMIOTRACE && m
> diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c
> index 5ff632ca3858..fb098b0b4505 100644
> --- a/kernel/trace/ring_buffer.c
> +++ b/kernel/trace/ring_buffer.c
> @@ -64,6 +64,10 @@ struct ring_buffer_cpu_meta {
> unsigned long commit_buffer;
> __u32 subbuf_size;
> __u32 nr_subbufs;
> +#ifdef CONFIG_RING_BUFFER_PERSISTENT_INJECT
> + __u32 nr_invalid;
> + __u32 entry_bytes;
> +#endif
> int buffers[];
> };
>
> @@ -2079,6 +2083,21 @@ static void rb_meta_validate_events(struct
> ring_buffer_per_cpu *cpu_buffer) if (discarded)
> pr_cont(" (%d pages discarded)", discarded);
> pr_cont("\n");
> +
> +#ifdef CONFIG_RING_BUFFER_PERSISTENT_INJECT
> + if (meta->nr_invalid)
> + pr_warn("Ring buffer testing [%d] invalid pages: %s (%d/%d)\n",
> + cpu_buffer->cpu,
> + (discarded == meta->nr_invalid) ? "PASSED" : "FAILED",
> + discarded, meta->nr_invalid);
> + if (meta->entry_bytes)
> + pr_warn("Ring buffer testing [%d] entry_bytes: %s (%ld/%ld)\n",
> + cpu_buffer->cpu,
> + (entry_bytes == meta->entry_bytes) ? "PASSED" : "FAILED",
> + (long)entry_bytes, (long)meta->entry_bytes);
> + meta->nr_invalid = 0;
> + meta->entry_bytes = 0;
> +#endif
> return;
>
> invalid:
> @@ -2559,12 +2578,67 @@ static void rb_free_cpu_buffer(struct
> ring_buffer_per_cpu *cpu_buffer) kfree(cpu_buffer);
> }
>
> +#ifdef CONFIG_RING_BUFFER_PERSISTENT_INJECT
> +static void rb_test_inject_invalid_pages(struct trace_buffer *buffer)
> +{
> + struct ring_buffer_per_cpu *cpu_buffer;
> + struct ring_buffer_cpu_meta *meta;
> + struct buffer_data_page *dpage;
> + u32 entry_bytes = 0;
> + unsigned long ptr;
> + int subbuf_size;
> + int invalid = 0;
> + int cpu;
> + int i;
> +
> + if (!(buffer->flags & RB_FL_TESTING))
> + return;
> +
> + guard(preempt)();
> + cpu = smp_processor_id();
> +
> + cpu_buffer = buffer->buffers[cpu];
> + meta = cpu_buffer->ring_meta;
> + ptr = (unsigned long)rb_subbufs_from_meta(meta);
> + subbuf_size = meta->subbuf_size;
> +
> + for (i = 0; i < meta->nr_subbufs; i++) {
> + int idx = meta->buffers[i];
> +
> + dpage = (void *)(ptr + idx * subbuf_size);
> + /* Skip unused pages */
> + if (!local_read(&dpage->commit))
> + continue;
> +
> + /*
> + * Invalidate even pages or multiples of 5. This will lead 3
This will cause 3
-- Steve
> + * contiguous invalidated(empty) pages.
> + */
> + if (!(i & 0x1) || !(i % 5)) {
> + local_add(subbuf_size + 1, &dpage->commit);
> + invalid++;
> + } else {
> + /* Count total commit bytes. */
> + entry_bytes += local_read(&dpage->commit);
> + }
> + }
> +
> + pr_info("Inject invalidated %d pages on CPU%d, total size: %ld\n",
> + invalid, cpu, (long)entry_bytes);
> + meta->nr_invalid = invalid;
> + meta->entry_bytes = entry_bytes;
> +}
^ permalink raw reply
* Re: [PATCH bpf v3 1/2] bpf: Reject sleepable kprobe_multi programs at attach time
From: Kumar Kartikeya Dwivedi @ 2026-04-01 22:45 UTC (permalink / raw)
To: Varun R Mallya, Jiri Olsa
Cc: bpf, ast, daniel, yonghong.song, rostedt, mhiramat, linux-kernel,
linux-trace-kernel
In-Reply-To: <20260401191126.440683-1-varunrmallya@gmail.com>
On Wed, 1 Apr 2026 at 21:11, Varun R Mallya <varunrmallya@gmail.com> wrote:
>
> kprobe.multi programs run in atomic/RCU context and cannot sleep.
> However, bpf_kprobe_multi_link_attach() did not validate whether the
> program being attached had the sleepable flag set, allowing sleepable
> helpers such as bpf_copy_from_user() to be invoked from a non-sleepable
> context.
>
> This causes a "sleeping function called from invalid context" splat:
>
> BUG: sleeping function called from invalid context at ./include/linux/uaccess.h:169
> in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 1787, name: sudo
> preempt_count: 1, expected: 0
> RCU nest depth: 2, expected: 0
>
> Fix this by rejecting sleepable programs early in
> bpf_kprobe_multi_link_attach(), before any further processing.
>
> Fixes: 0dcac2725406 ("bpf: Add multi kprobe link")
> Signed-off-by: Varun R Mallya <varunrmallya@gmail.com>
> ---
Acked-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Jiri, would be great if you can ack both patches too.
^ permalink raw reply
* Re: [PATCH RFC v4 10/44] KVM: guest_memfd: Add support for KVM_SET_MEMORY_ATTRIBUTES2
From: Ackerley Tng @ 2026-04-01 22:46 UTC (permalink / raw)
To: Michael Roth
Cc: aik, andrew.jones, binbin.wu, brauner, chao.p.peng, david,
ira.weiny, jmattson, jroedel, jthoughton, oupton, pankaj.gupta,
qperret, rick.p.edgecombe, rientjes, shivankg, steven.price,
tabba, willy, wyihan, yan.y.zhao, forkloop, pratyush,
suzuki.poulose, aneesh.kumar, Paolo Bonzini, Sean Christopherson,
Thomas Gleixner, Ingo Molnar, Borislav Petkov, Dave Hansen, x86,
H. Peter Anvin, Steven Rostedt, Masami Hiramatsu,
Mathieu Desnoyers, Jonathan Corbet, Shuah Khan, Shuah Khan,
Vishal Annapurve, Andrew Morton, Chris Li, Kairui Song,
Kemeng Shi, Nhat Pham, Baoquan He, Barry Song, Axel Rasmussen,
Yuanchu Xie, Wei Xu, Jason Gunthorpe, Vlastimil Babka, kvm,
linux-kernel, linux-trace-kernel, linux-doc, linux-kselftest,
linux-mm
In-Reply-To: <cqkyz4zxosmev6lnbasa32ed5jjfljrx3gr6plyjfhrtbysuwr@rg5noy6y6ayu>
Michael Roth <michael.roth@amd.com> writes:
> On Thu, Mar 26, 2026 at 03:24:19PM -0700, Ackerley Tng wrote:
>> For shared to private conversions, if refcounts on any of the folios
>> within the range are elevated, fail the conversion with -EAGAIN.
>>
>> At the point of shared to private conversion, all folios in range are
>> also unmapped. The filemap_invalidate_lock() is held, so no faulting
>> can occur. Hence, from that point on, only transient refcounts can be
>> taken on the folios associated with that guest_memfd.
>>
>> Hence, it is safe to do the conversion from shared to private.
>>
>> After conversion is complete, refcounts may become elevated, but that
>> is fine since users of transient refcounts don't actually access
>> memory.
>>
>> For private to shared conversions, there are no refcount checks, since
>> the guest is the only user of private pages, and guest_memfd will be the
>> only holder of refcounts on private pages.
>
> I think KVM_CAP_GUEST_MEMFD_MEMORY_ATTRIBUTES deserves some mention in
> the commit log.
>
Will update this in the next revision. Thanks!
>>
>>
>> [...snip...]
>>
>>
>> +Set attributes for a range of offsets within a guest_memfd to
>> +KVM_MEMORY_ATTRIBUTE_PRIVATE to limit the specified guest_memfd backed
>> +memory range for guest_use. Even if KVM_CAP_GUEST_MEMFD_MMAP is
>> +supported, after a successful call to set
>> +KVM_MEMORY_ATTRIBUTE_PRIVATE, the requested range will not be mappable
>> +into host userspace and will only be mappable by the guest.
>> +
>> +To allow the range to be mappable into host userspace again, call
>> +KVM_SET_MEMORY_ATTRIBUTES2 on the guest_memfd again with
>> +KVM_MEMORY_ATTRIBUTE_PRIVATE unset.
>> +
>> +If this ioctl returns -EAGAIN, the offset of the page with unexpected
>> +refcounts will be returned in `error_offset`. This can occur if there
>> +are transient refcounts on the pages, taken by other parts of the
>> +kernel.
>
> That's only true for the guest_memfd ioctl, for KVM ioctl these new
> fields and r/w behavior are basically ignored. So you might need to be
> clearer on which fields/behavior are specific to guest_memfd like in
> the preceeding paragraphs..
>
Yes, will update in the next revision, thanks!
> ..or maybe it's better to do the opposite and just have a blanket 'for
> now, all newly-described behavior pertains only to usage via a
> guest_memfd ioctl, and for KVM ioctls only the fields/behaviors common
> with KVM_SET_MEMORY_ATTRIBUTES are applicable.', since it doesn't seem
> like vm_memory_attributes=1 is long for this world and that's the only
> case where KVM memory attribute ioctls seem relevant.
>
> But then it makes me wonder, if we adopt the semantics I mentioned
> earlier and have KVM_CAP_GUEST_MEMFD_MEMORY_ATTRIBUTES advertise both
> the gmem ioctl support as well as the struct kvm_memory_attributes2
> support, if we should even advertise KVM_CAP_MEMORY_ATTRIBUTES2 at all
> as part of this series.
>
Read your other email as well, thanks for reviewing!
It makes sense, hope this captures what you suggested. In v5,
If vm_memory_attributes == 1:
(KVM_CAP_MEMORY_ATTRIBUTES2 will be removed (will return 0))
If vm_memory_attributes == 0 aka attributes are tracked by guest_memfd:
KVM_CAP_GUEST_MEMFD_MEMORY_ATTRIBUTES2 will return valid attributes
(KVM_CAP_MEMORY_ATTRIBUTES2 will be removed (will return 0))
So yup, KVM_CAP_MEMORY_ATTRIBUTES2 will not even be #defined at all.
>> +
>> +Userspace is expected to figure out how to remove all known refcounts
>> +on the shared pages, such as refcounts taken by get_user_pages(), and
>> +try the ioctl again. A possible source of these long term refcounts is
>> +if the guest_memfd memory was pinned in IOMMU page tables.
>
> One might read this to mean error_offset is used purely for the EAGAIN
> case, so it might be worth touching on the other cases as well.
>
Will update this in the next revision.
> -Mike
>
>>
>> [...snip...]
>>
^ permalink raw reply
* Re: [PATCH v15 5/5] ring-buffer: Show commit numbers in buffer_meta file
From: Steven Rostedt @ 2026-04-01 22:49 UTC (permalink / raw)
To: Masami Hiramatsu (Google)
Cc: Catalin Marinas, Will Deacon, Mathieu Desnoyers, linux-kernel,
linux-trace-kernel, Ian Rogers, linux-arm-kernel
In-Reply-To: <177494619878.71933.15471023049227398684.stgit@mhiramat.tok.corp.google.com>
On Tue, 31 Mar 2026 17:36:38 +0900
"Masami Hiramatsu (Google)" <mhiramat@kernel.org> wrote:
> From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
>
> In addition to the index number, show the commit numbers of
> each data page in per_cpu buffer_meta file.
in the per_cpu buffer_meta file.
-- Steve
> This is useful for understanding the current status of the
> persistent ring buffer. (Note that this file is shown
> only for persistent ring buffer and its backup instance)
>
^ permalink raw reply
* Re: [PATCH bpf v3 2/2] selftests/bpf: Add test to ensure kprobe_multi is not sleepable
From: Kumar Kartikeya Dwivedi @ 2026-04-01 22:50 UTC (permalink / raw)
To: Varun R Mallya
Cc: bpf, ast, daniel, yonghong.song, jolsa, rostedt, mhiramat,
linux-kernel, linux-trace-kernel
In-Reply-To: <20260401191126.440683-2-varunrmallya@gmail.com>
On Wed, 1 Apr 2026 at 21:11, Varun R Mallya <varunrmallya@gmail.com> wrote:
>
> Add a selftest to ensure that kprobe_multi programs cannot be attached
> using the BPF_F_SLEEPABLE flag. This test succeeds when the kernel
> rejects attachment of kprobe_multi when the BPF_F_SLEEPABLE flag is set.
>
> Signed-off-by: Varun R Mallya <varunrmallya@gmail.com>
> ---
Acked-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
> .../bpf/prog_tests/kprobe_multi_test.c | 41 +++++++++++++++++++
> .../bpf/progs/kprobe_multi_sleepable.c | 13 ++++++
> 2 files changed, 54 insertions(+)
> create mode 100644 tools/testing/selftests/bpf/progs/kprobe_multi_sleepable.c
>
> diff --git a/tools/testing/selftests/bpf/prog_tests/kprobe_multi_test.c b/tools/testing/selftests/bpf/prog_tests/kprobe_multi_test.c
> index 78c974d4ea33..f02fec2b6fda 100644
> --- a/tools/testing/selftests/bpf/prog_tests/kprobe_multi_test.c
> +++ b/tools/testing/selftests/bpf/prog_tests/kprobe_multi_test.c
> @@ -10,6 +10,7 @@
> #include "kprobe_multi_session_cookie.skel.h"
> #include "kprobe_multi_verifier.skel.h"
> #include "kprobe_write_ctx.skel.h"
> +#include "kprobe_multi_sleepable.skel.h"
> #include "bpf/libbpf_internal.h"
> #include "bpf/hashmap.h"
>
> @@ -633,6 +634,44 @@ static void test_attach_write_ctx(void)
> }
> #endif
>
> +static void test_attach_multi_sleepable(void)
> +{
> + struct kprobe_multi_sleepable *skel;
> + int err;
> +
> + skel = kprobe_multi_sleepable__open();
> + if (!ASSERT_OK_PTR(skel, "kprobe_multi_sleepable__open"))
> + return;
> +
> + err = bpf_program__set_flags(skel->progs.handle_kprobe_multi_sleepable,
> + BPF_F_SLEEPABLE);
> + if (!ASSERT_OK(err, "bpf_program__set_flags"))
> + goto cleanup;
> +
> + /* Load should succeed even with BPF_F_SLEEPABLE for KPROBE types */
> + err = kprobe_multi_sleepable__load(skel);
> + if (!ASSERT_OK(err, "kprobe_multi_sleepable__load"))
> + goto cleanup;
> +
> + /* Attachment must fail for kprobe.multi + BPF_F_SLEEPABLE.
> + * Also chosen a stable symbol to send into opts
> + */
> + LIBBPF_OPTS(bpf_kprobe_multi_opts, opts);
> + const char *sym = "vfs_read";
> +
> + opts.syms = &sym;
> + opts.cnt = 1;
> +
> + skel->links.handle_kprobe_multi_sleepable =
> + bpf_program__attach_kprobe_multi_opts(skel->progs.handle_kprobe_multi_sleepable,
> + NULL, &opts);
> + ASSERT_ERR_PTR(skel->links.handle_kprobe_multi_sleepable,
> + "bpf_program__attach_kprobe_multi_opts");
Nit: While vfs_read will likely remain stable, the check could
probably be stronger to distinguish an attach error from -EINVAL?
I added a typo to vfs_read and it still passed, because it failed to
attach instead of getting rejected on unfixed kernel.
May not be a big deal since vfs_read is unlikely to break.
I verified it works by adding bpf_copy_from_user to the program and
attaching to SYS_PREFIX sys_getpid and invoking the splat though, so
LGTM otherwise.
> +
> +cleanup:
> + kprobe_multi_sleepable__destroy(skel);
> +}
> +
> [...]
^ permalink raw reply
* Re: [PATCH v2 00/11] Add spi-hid transport driver
From: Jingyuan Liang @ 2026-04-02 1:54 UTC (permalink / raw)
To: Krzysztof Kozlowski
Cc: Jiri Kosina, Benjamin Tissoires, Jonathan Corbet, Mark Brown,
Steven Rostedt, Masami Hiramatsu, Mathieu Desnoyers,
Dmitry Torokhov, Rob Herring, Krzysztof Kozlowski, Conor Dooley,
linux-input, linux-doc, linux-kernel, linux-spi,
linux-trace-kernel, devicetree, hbarnor, tfiga, Jarrett Schultz,
Dmitry Antipov, Angela Czubak
In-Reply-To: <20260325-naughty-hungry-wapiti-658e83@quoll>
On Wed, Mar 25, 2026 at 1:49 AM Krzysztof Kozlowski <krzk@kernel.org> wrote:
>
> On Tue, Mar 24, 2026 at 06:39:33AM +0000, Jingyuan Liang wrote:
> > This series picks up the spi-hid driver work originally started by
> > Microsoft. The patch breakdown has been modified and the implementation
> > has been refactored to address upstream feedback and testing issues. We
> > are submitting this as a new series while keeping the original sign-off
> > chain to reflect the history.
> >
> > Same as the original series, there is a change to HID documentation, some
> > HID core changes to support a SPI device, the SPI HID transport driver,
> > and HID over SPI Device Tree binding. We have added the HID over SPI ACPI
> > support, power management, panel follower, and quirks for Ilitek touch
> > controllers.
> >
> > Original authors: Jarrett Schultz <jaschultz@microsoft.com>,
> > Dmitry Antipov <dmanti@microsoft.com>
> > Link: https://lore.kernel.org/r/86b63b7b-afda-d7f4-7bfa-175085d5a8ef@gmail.com
> >
> > Signed-off-by: Jingyuan Liang <jingyliang@chromium.org>
> > ---
> > Changes in v2:
> > - Fix style problems and remove unnecessary fields from the DT binding file
>
> Style and removal? So other comments were skipped?
>
> Please write detailed changelogs, otherwise it feels you just ignore
> parts of the feedback.
>
> Best regards,
> Krzysztof
>
Comments are either resolved or awaiting further confirmation. I will
add more details
to v2 changelog in v3.
^ permalink raw reply
* [PATCH v3 00/11] Add spi-hid transport driver
From: Jingyuan Liang @ 2026-04-02 1:59 UTC (permalink / raw)
To: Jiri Kosina, Benjamin Tissoires, Jonathan Corbet, Mark Brown,
Steven Rostedt, Masami Hiramatsu, Mathieu Desnoyers,
Dmitry Torokhov, Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: linux-input, linux-doc, linux-kernel, linux-spi,
linux-trace-kernel, devicetree, hbarnor, tfiga, Jingyuan Liang,
Jarrett Schultz, Dmitry Antipov, Angela Czubak
This series picks up the spi-hid driver work originally started by
Microsoft. The patch breakdown has been modified and the implementation
has been refactored to address upstream feedback and testing issues. We
are submitting this as a new series while keeping the original sign-off
chain to reflect the history.
Same as the original series, there is a change to HID documentation, some
HID core changes to support a SPI device, the SPI HID transport driver,
and HID over SPI Device Tree binding. We have added the HID over SPI ACPI
support, power management, panel follower, and quirks for Ilitek touch
controllers.
Original authors: Jarrett Schultz <jaschultz@microsoft.com>,
Dmitry Antipov <dmanti@microsoft.com>
Link: https://lore.kernel.org/r/86b63b7b-afda-d7f4-7bfa-175085d5a8ef@gmail.com
Signed-off-by: Jingyuan Liang <jingyliang@chromium.org>
---
Changes in v3:
- Add io_lock init
- Relocate tracepoints to drivers/hid/spi-hid/ and fix tracepoint macros
- Add tracepoints for sync, error handling, reset, and report processing
- Clean up internal includes and fix Makefile CFLAGS
- Add more details in v2 changelog
- Link to v2: https://lore.kernel.org/r/20260324-send-upstream-v2-0-521ce8afff86@chromium.org
Changes in v2:
- Clean up DT bindings: fix formatting and remove timing and flags properties
- Update DT binding example: use a device-specific compatible and drop
reset_assert
- Simplify ACPI/OF match tables by removing ACPI_PTR/of_match_ptr
- Refactor OF driver to use match data for timing parameters instead
of DT properties
- Switch to fsleep() for delays in ACPI and OF drivers
- Drop patch 12 as it is vendor specific
- Add a lock to fix input/output concurrency race
- Link to v1: https://lore.kernel.org/r/20260303-send-upstream-v1-0-1515ba218f3d@chromium.org
---
Angela Czubak (2):
HID: spi-hid: add transport driver skeleton for HID over SPI bus
HID: spi_hid: add ACPI support for SPI over HID
Jarrett Schultz (3):
Documentation: Correction in HID output_report callback description.
HID: Add BUS_SPI support and define HID_SPI_DEVICE macro
HID: spi_hid: add device tree support for SPI over HID
Jingyuan Liang (6):
HID: spi-hid: add spi-hid driver HID layer
HID: spi-hid: add HID SPI protocol implementation
HID: spi_hid: add spi_hid traces
dt-bindings: input: Document hid-over-spi DT schema
HID: spi-hid: add power management implementation
HID: spi-hid: add panel follower support
.../devicetree/bindings/input/hid-over-spi.yaml | 126 ++
Documentation/hid/hid-transport.rst | 4 +-
drivers/hid/Kconfig | 2 +
drivers/hid/Makefile | 2 +
drivers/hid/hid-core.c | 3 +
drivers/hid/spi-hid/Kconfig | 45 +
drivers/hid/spi-hid/Makefile | 12 +
drivers/hid/spi-hid/spi-hid-acpi.c | 254 ++++
drivers/hid/spi-hid/spi-hid-core.c | 1456 ++++++++++++++++++++
drivers/hid/spi-hid/spi-hid-core.h | 98 ++
drivers/hid/spi-hid/spi-hid-of.c | 244 ++++
drivers/hid/spi-hid/spi-hid-trace.h | 169 +++
drivers/hid/spi-hid/spi-hid.h | 46 +
include/linux/hid.h | 2 +
14 files changed, 2461 insertions(+), 2 deletions(-)
---
base-commit: 05f7e89ab9731565d8a62e3b5d1ec206485eeb0b
change-id: 20260212-send-upstream-75f6fd9ed92e
Best regards,
--
Jingyuan Liang <jingyliang@chromium.org>
^ permalink raw reply
* [PATCH v3 01/11] Documentation: Correction in HID output_report callback description.
From: Jingyuan Liang @ 2026-04-02 1:59 UTC (permalink / raw)
To: Jiri Kosina, Benjamin Tissoires, Jonathan Corbet, Mark Brown,
Steven Rostedt, Masami Hiramatsu, Mathieu Desnoyers,
Dmitry Torokhov, Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: linux-input, linux-doc, linux-kernel, linux-spi,
linux-trace-kernel, devicetree, hbarnor, tfiga, Jingyuan Liang,
Jarrett Schultz, Dmitry Antipov
In-Reply-To: <20260402-send-upstream-v3-0-6091c458d357@chromium.org>
From: Jarrett Schultz <jaschultz@microsoft.com>
Originally output_report callback was described as must-be asynchronous,
but that is not the case in some implementations, namely i2c-hid.
Correct the documentation to say that it may be asynchronous.
Signed-off-by: Dmitry Antipov <dmanti@microsoft.com>
Reviewed-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Jingyuan Liang <jingyliang@chromium.org>
---
Documentation/hid/hid-transport.rst | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/Documentation/hid/hid-transport.rst b/Documentation/hid/hid-transport.rst
index 6f1692da296c..2008cf432af1 100644
--- a/Documentation/hid/hid-transport.rst
+++ b/Documentation/hid/hid-transport.rst
@@ -327,8 +327,8 @@ The available HID callbacks are:
Send raw output report via intr channel. Used by some HID device drivers
which require high throughput for outgoing requests on the intr channel. This
- must not cause SET_REPORT calls! This must be implemented as asynchronous
- output report on the intr channel!
+ must not cause SET_REPORT calls! This call might be asynchronous, so the
+ caller should not expect an immediate response!
::
--
2.53.0.1185.g05d4b7b318-goog
^ permalink raw reply related
* [PATCH v3 02/11] HID: Add BUS_SPI support and define HID_SPI_DEVICE macro
From: Jingyuan Liang @ 2026-04-02 1:59 UTC (permalink / raw)
To: Jiri Kosina, Benjamin Tissoires, Jonathan Corbet, Mark Brown,
Steven Rostedt, Masami Hiramatsu, Mathieu Desnoyers,
Dmitry Torokhov, Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: linux-input, linux-doc, linux-kernel, linux-spi,
linux-trace-kernel, devicetree, hbarnor, tfiga, Jingyuan Liang,
Jarrett Schultz, Dmitry Antipov
In-Reply-To: <20260402-send-upstream-v3-0-6091c458d357@chromium.org>
From: Jarrett Schultz <jaschultz@microsoft.com>
If connecting a hid_device with bus field indicating BUS_SPI print out
"SPI" in the debug print.
Macro sets the bus field to BUS_SPI and uses arguments to set vendor
product fields.
Signed-off-by: Dmitry Antipov <dmanti@microsoft.com>
Reviewed-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Jingyuan Liang <jingyliang@chromium.org>
---
drivers/hid/hid-core.c | 3 +++
include/linux/hid.h | 2 ++
2 files changed, 5 insertions(+)
diff --git a/drivers/hid/hid-core.c b/drivers/hid/hid-core.c
index a5b3a8ca2fcb..813c9c743ccd 100644
--- a/drivers/hid/hid-core.c
+++ b/drivers/hid/hid-core.c
@@ -2316,6 +2316,9 @@ int hid_connect(struct hid_device *hdev, unsigned int connect_mask)
case BUS_I2C:
bus = "I2C";
break;
+ case BUS_SPI:
+ bus = "SPI";
+ break;
case BUS_SDW:
bus = "SOUNDWIRE";
break;
diff --git a/include/linux/hid.h b/include/linux/hid.h
index dce862cafbbd..957f322a0ebd 100644
--- a/include/linux/hid.h
+++ b/include/linux/hid.h
@@ -786,6 +786,8 @@ struct hid_descriptor {
.bus = BUS_BLUETOOTH, .vendor = (ven), .product = (prod)
#define HID_I2C_DEVICE(ven, prod) \
.bus = BUS_I2C, .vendor = (ven), .product = (prod)
+#define HID_SPI_DEVICE(ven, prod) \
+ .bus = BUS_SPI, .vendor = (ven), .product = (prod)
#define HID_REPORT_ID(rep) \
.report_type = (rep)
--
2.53.0.1185.g05d4b7b318-goog
^ permalink raw reply related
* [PATCH v3 03/11] HID: spi-hid: add transport driver skeleton for HID over SPI bus
From: Jingyuan Liang @ 2026-04-02 1:59 UTC (permalink / raw)
To: Jiri Kosina, Benjamin Tissoires, Jonathan Corbet, Mark Brown,
Steven Rostedt, Masami Hiramatsu, Mathieu Desnoyers,
Dmitry Torokhov, Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: linux-input, linux-doc, linux-kernel, linux-spi,
linux-trace-kernel, devicetree, hbarnor, tfiga, Jingyuan Liang,
Angela Czubak, Dmitry Antipov
In-Reply-To: <20260402-send-upstream-v3-0-6091c458d357@chromium.org>
From: Angela Czubak <acz@semihalf.com>
Create spi-hid folder and add Kconfig and Makefile for spi-hid driver.
Add basic device structure, definitions, and probe/remove functions.
Signed-off-by: Dmitry Antipov <dmanti@microsoft.com>
Signed-off-by: Angela Czubak <acz@semihalf.com>
Signed-off-by: Jingyuan Liang <jingyliang@chromium.org>
---
drivers/hid/Kconfig | 2 +
drivers/hid/Makefile | 2 +
drivers/hid/spi-hid/Kconfig | 15 +++
drivers/hid/spi-hid/Makefile | 9 ++
drivers/hid/spi-hid/spi-hid-core.c | 213 +++++++++++++++++++++++++++++++++++++
5 files changed, 241 insertions(+)
diff --git a/drivers/hid/Kconfig b/drivers/hid/Kconfig
index 920a64b66b25..c6ae23bfb75d 100644
--- a/drivers/hid/Kconfig
+++ b/drivers/hid/Kconfig
@@ -1434,6 +1434,8 @@ source "drivers/hid/bpf/Kconfig"
source "drivers/hid/i2c-hid/Kconfig"
+source "drivers/hid/spi-hid/Kconfig"
+
source "drivers/hid/intel-ish-hid/Kconfig"
source "drivers/hid/amd-sfh-hid/Kconfig"
diff --git a/drivers/hid/Makefile b/drivers/hid/Makefile
index 361a7daedeb8..6b43e789b39a 100644
--- a/drivers/hid/Makefile
+++ b/drivers/hid/Makefile
@@ -169,6 +169,8 @@ obj-$(CONFIG_USB_KBD) += usbhid/
obj-$(CONFIG_I2C_HID_CORE) += i2c-hid/
+obj-$(CONFIG_SPI_HID_CORE) += spi-hid/
+
obj-$(CONFIG_INTEL_ISH_HID) += intel-ish-hid/
obj-$(CONFIG_AMD_SFH_HID) += amd-sfh-hid/
diff --git a/drivers/hid/spi-hid/Kconfig b/drivers/hid/spi-hid/Kconfig
new file mode 100644
index 000000000000..836fdefe8345
--- /dev/null
+++ b/drivers/hid/spi-hid/Kconfig
@@ -0,0 +1,15 @@
+# SPDX-License-Identifier: GPL-2.0-only
+#
+# Copyright (c) 2021 Microsoft Corporation
+#
+
+menuconfig SPI_HID
+ tristate "SPI HID support"
+ default y
+ depends on SPI
+
+if SPI_HID
+
+config SPI_HID_CORE
+ tristate
+endif
diff --git a/drivers/hid/spi-hid/Makefile b/drivers/hid/spi-hid/Makefile
new file mode 100644
index 000000000000..92e24cddbfc2
--- /dev/null
+++ b/drivers/hid/spi-hid/Makefile
@@ -0,0 +1,9 @@
+# SPDX-License-Identifier: GPL-2.0-only
+#
+# Makefile for the SPI HID input drivers
+#
+# Copyright (c) 2021 Microsoft Corporation
+#
+
+obj-$(CONFIG_SPI_HID_CORE) += spi-hid.o
+spi-hid-objs = spi-hid-core.o
diff --git a/drivers/hid/spi-hid/spi-hid-core.c b/drivers/hid/spi-hid/spi-hid-core.c
new file mode 100644
index 000000000000..d7b4d4adad95
--- /dev/null
+++ b/drivers/hid/spi-hid/spi-hid-core.c
@@ -0,0 +1,213 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * HID over SPI protocol implementation
+ *
+ * Copyright (c) 2021 Microsoft Corporation
+ * Copyright (c) 2026 Google LLC
+ *
+ * This code is partly based on "HID over I2C protocol implementation:
+ *
+ * Copyright (c) 2012 Benjamin Tissoires <benjamin.tissoires@gmail.com>
+ * Copyright (c) 2012 Ecole Nationale de l'Aviation Civile, France
+ * Copyright (c) 2012 Red Hat, Inc
+ *
+ * which in turn is partly based on "USB HID support for Linux":
+ *
+ * Copyright (c) 1999 Andreas Gal
+ * Copyright (c) 2000-2005 Vojtech Pavlik <vojtech@suse.cz>
+ * Copyright (c) 2005 Michael Haboustak <mike-@cinci.rr.com> for Concept2, Inc
+ * Copyright (c) 2007-2008 Oliver Neukum
+ * Copyright (c) 2006-2010 Jiri Kosina
+ */
+
+#include <linux/device.h>
+#include <linux/hid.h>
+#include <linux/hid-over-spi.h>
+#include <linux/interrupt.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/spi/spi.h>
+
+/* struct spi_hid_conf - Conf provided to the core */
+struct spi_hid_conf {
+ u32 input_report_header_address;
+ u32 input_report_body_address;
+ u32 output_report_address;
+ u8 read_opcode;
+ u8 write_opcode;
+};
+
+/**
+ * struct spihid_ops - Ops provided to the core
+ * @power_up: do sequencing to power up the device
+ * @power_down: do sequencing to power down the device
+ * @assert_reset: do sequencing to assert the reset line
+ * @deassert_reset: do sequencing to deassert the reset line
+ * @sleep_minimal_reset_delay: minimal sleep delay during reset
+ */
+struct spihid_ops {
+ int (*power_up)(struct spihid_ops *ops);
+ int (*power_down)(struct spihid_ops *ops);
+ int (*assert_reset)(struct spihid_ops *ops);
+ int (*deassert_reset)(struct spihid_ops *ops);
+ void (*sleep_minimal_reset_delay)(struct spihid_ops *ops);
+};
+
+/* Driver context */
+struct spi_hid {
+ struct spi_device *spi; /* spi device. */
+ struct hid_device *hid; /* pointer to corresponding HID dev. */
+
+ struct spihid_ops *ops;
+ struct spi_hid_conf *conf;
+
+ enum hidspi_power_state power_state;
+
+ u32 regulator_error_count;
+ int regulator_last_error;
+ u32 bus_error_count;
+ int bus_last_error;
+ u32 dir_count; /* device initiated reset count. */
+};
+
+static const char *spi_hid_power_mode_string(enum hidspi_power_state power_state)
+{
+ switch (power_state) {
+ case HIDSPI_ON:
+ return "d0";
+ case HIDSPI_SLEEP:
+ return "d2";
+ case HIDSPI_OFF:
+ return "d3";
+ default:
+ return "unknown";
+ }
+}
+
+static irqreturn_t spi_hid_dev_irq(int irq, void *_shid)
+{
+ return IRQ_HANDLED;
+}
+
+static ssize_t bus_error_count_show(struct device *dev,
+ struct device_attribute *attr, char *buf)
+{
+ struct spi_hid *shid = dev_get_drvdata(dev);
+
+ return sysfs_emit(buf, "%d (%d)\n",
+ shid->bus_error_count, shid->bus_last_error);
+}
+static DEVICE_ATTR_RO(bus_error_count);
+
+static ssize_t regulator_error_count_show(struct device *dev,
+ struct device_attribute *attr,
+ char *buf)
+{
+ struct spi_hid *shid = dev_get_drvdata(dev);
+
+ return sysfs_emit(buf, "%d (%d)\n",
+ shid->regulator_error_count,
+ shid->regulator_last_error);
+}
+static DEVICE_ATTR_RO(regulator_error_count);
+
+static ssize_t device_initiated_reset_count_show(struct device *dev,
+ struct device_attribute *attr,
+ char *buf)
+{
+ struct spi_hid *shid = dev_get_drvdata(dev);
+
+ return sysfs_emit(buf, "%d\n", shid->dir_count);
+}
+static DEVICE_ATTR_RO(device_initiated_reset_count);
+
+static struct attribute *spi_hid_attrs[] = {
+ &dev_attr_bus_error_count.attr,
+ &dev_attr_regulator_error_count.attr,
+ &dev_attr_device_initiated_reset_count.attr,
+ NULL /* Terminator */
+};
+
+static const struct attribute_group spi_hid_group = {
+ .attrs = spi_hid_attrs,
+};
+
+const struct attribute_group *spi_hid_groups[] = {
+ &spi_hid_group,
+ NULL
+};
+EXPORT_SYMBOL_GPL(spi_hid_groups);
+
+int spi_hid_core_probe(struct spi_device *spi, struct spihid_ops *ops,
+ struct spi_hid_conf *conf)
+{
+ struct device *dev = &spi->dev;
+ struct spi_hid *shid;
+ int error;
+
+ if (spi->irq <= 0)
+ return dev_err_probe(dev, spi->irq ?: -EINVAL, "Missing IRQ\n");
+
+ shid = devm_kzalloc(dev, sizeof(*shid), GFP_KERNEL);
+ if (!shid)
+ return -ENOMEM;
+
+ shid->spi = spi;
+ shid->power_state = HIDSPI_ON;
+ shid->ops = ops;
+ shid->conf = conf;
+
+ spi_set_drvdata(spi, shid);
+
+ /*
+ * At the end of probe we initialize the device:
+ * 0) assert reset, bias the interrupt line
+ * 1) sleep minimal reset delay
+ * 2) request IRQ
+ * 3) power up the device
+ * 4) deassert reset (high)
+ * After this we expect an IRQ with a reset response.
+ */
+
+ shid->ops->assert_reset(shid->ops);
+
+ shid->ops->sleep_minimal_reset_delay(shid->ops);
+
+ error = devm_request_threaded_irq(dev, spi->irq, NULL, spi_hid_dev_irq,
+ IRQF_ONESHOT, dev_name(&spi->dev), shid);
+ if (error) {
+ dev_err(dev, "%s: unable to request threaded IRQ.", __func__);
+ return error;
+ }
+
+ error = shid->ops->power_up(shid->ops);
+ if (error) {
+ dev_err(dev, "%s: could not power up.", __func__);
+ return error;
+ }
+
+ shid->ops->deassert_reset(shid->ops);
+
+ dev_dbg(dev, "%s: d3 -> %s.", __func__,
+ spi_hid_power_mode_string(shid->power_state));
+
+ return 0;
+}
+EXPORT_SYMBOL_GPL(spi_hid_core_probe);
+
+void spi_hid_core_remove(struct spi_device *spi)
+{
+ struct spi_hid *shid = spi_get_drvdata(spi);
+ struct device *dev = &spi->dev;
+ int error;
+
+ shid->ops->assert_reset(shid->ops);
+ error = shid->ops->power_down(shid->ops);
+ if (error)
+ dev_err(dev, "failed to disable regulator.");
+}
+EXPORT_SYMBOL_GPL(spi_hid_core_remove);
+
+MODULE_DESCRIPTION("HID over SPI transport driver");
+MODULE_AUTHOR("Dmitry Antipov <dmanti@microsoft.com>");
+MODULE_LICENSE("GPL");
--
2.53.0.1185.g05d4b7b318-goog
^ permalink raw reply related
* [PATCH v3 04/11] HID: spi-hid: add spi-hid driver HID layer
From: Jingyuan Liang @ 2026-04-02 1:59 UTC (permalink / raw)
To: Jiri Kosina, Benjamin Tissoires, Jonathan Corbet, Mark Brown,
Steven Rostedt, Masami Hiramatsu, Mathieu Desnoyers,
Dmitry Torokhov, Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: linux-input, linux-doc, linux-kernel, linux-spi,
linux-trace-kernel, devicetree, hbarnor, tfiga, Jingyuan Liang,
Dmitry Antipov, Angela Czubak
In-Reply-To: <20260402-send-upstream-v3-0-6091c458d357@chromium.org>
Add HID low level driver callbacks to register SPI as a HID driver, and
an external touch device as a HID device.
Signed-off-by: Dmitry Antipov <dmanti@microsoft.com>
Signed-off-by: Angela Czubak <acz@semihalf.com>
Signed-off-by: Jingyuan Liang <jingyliang@chromium.org>
---
drivers/hid/spi-hid/spi-hid-core.c | 519 +++++++++++++++++++++++++++++++++++++
1 file changed, 519 insertions(+)
diff --git a/drivers/hid/spi-hid/spi-hid-core.c b/drivers/hid/spi-hid/spi-hid-core.c
index d7b4d4adad95..4723b87346d4 100644
--- a/drivers/hid/spi-hid/spi-hid-core.c
+++ b/drivers/hid/spi-hid/spi-hid-core.c
@@ -20,13 +20,69 @@
* Copyright (c) 2006-2010 Jiri Kosina
*/
+#include <linux/completion.h>
+#include <linux/crc32.h>
#include <linux/device.h>
+#include <linux/err.h>
#include <linux/hid.h>
#include <linux/hid-over-spi.h>
#include <linux/interrupt.h>
+#include <linux/jiffies.h>
#include <linux/module.h>
+#include <linux/mutex.h>
#include <linux/slab.h>
#include <linux/spi/spi.h>
+#include <linux/string.h>
+#include <linux/sysfs.h>
+#include <linux/unaligned.h>
+
+#define SPI_HID_OUTPUT_REPORT_CONTENT_ID_DESC_REQUEST 0x00
+
+#define SPI_HID_RESP_TIMEOUT 1000
+
+/* Protocol message size constants */
+#define SPI_HID_OUTPUT_HEADER_LEN 8
+
+/* flags */
+/*
+ * ready flag indicates that the FW is ready to accept commands and
+ * requests. The FW becomes ready after sending the report descriptor.
+ */
+#define SPI_HID_READY 0
+
+/* Raw input buffer with data from the bus */
+struct spi_hid_input_buf {
+ u8 header[HIDSPI_INPUT_HEADER_SIZE];
+ u8 body[HIDSPI_INPUT_BODY_HEADER_SIZE];
+ u8 content[];
+};
+
+/* Raw output report buffer to be put on the bus */
+struct spi_hid_output_buf {
+ u8 header[SPI_HID_OUTPUT_HEADER_LEN];
+ u8 content[];
+};
+
+/* Data necessary to send an output report */
+struct spi_hid_output_report {
+ u8 report_type;
+ u16 content_length;
+ u8 content_id;
+ u8 *content;
+};
+
+/* Processed data from a device descriptor */
+struct spi_hid_device_descriptor {
+ u16 hid_version;
+ u16 report_descriptor_length;
+ u16 max_input_length;
+ u16 max_output_length;
+ u16 max_fragment_length;
+ u16 vendor_id;
+ u16 product_id;
+ u16 version_id;
+ u8 no_output_report_ack;
+};
/* struct spi_hid_conf - Conf provided to the core */
struct spi_hid_conf {
@@ -61,8 +117,26 @@ struct spi_hid {
struct spihid_ops *ops;
struct spi_hid_conf *conf;
+ struct spi_hid_device_descriptor desc; /* HID device descriptor. */
+ struct spi_hid_output_buf *output; /* Output buffer. */
+ struct spi_hid_input_buf *input; /* Input buffer. */
+ struct spi_hid_input_buf *response; /* Response buffer. */
+
+ u16 response_length;
+ u16 bufsize;
+
enum hidspi_power_state power_state;
+ u8 reset_attempts; /* The number of reset attempts. */
+
+ unsigned long flags; /* device flags. */
+
+ /* Control lock to make sure one output transaction at a time. */
+ struct mutex output_lock;
+ struct completion output_done;
+
+ u32 report_descriptor_crc32; /* HID report descriptor crc32 checksum. */
+
u32 regulator_error_count;
int regulator_last_error;
u32 bus_error_count;
@@ -70,6 +144,33 @@ struct spi_hid {
u32 dir_count; /* device initiated reset count. */
};
+static struct hid_ll_driver spi_hid_ll_driver;
+
+static void spi_hid_populate_output_header(u8 *buf,
+ const struct spi_hid_conf *conf,
+ const struct spi_hid_output_report *report)
+{
+ buf[0] = conf->write_opcode;
+ put_unaligned_be24(conf->output_report_address, &buf[1]);
+ buf[4] = report->report_type;
+ put_unaligned_le16(report->content_length, &buf[5]);
+ buf[7] = report->content_id;
+}
+
+static int spi_hid_output(struct spi_hid *shid, const void *buf, u16 length)
+{
+ int error;
+
+ error = spi_write(shid->spi, buf, length);
+
+ if (error) {
+ shid->bus_error_count++;
+ shid->bus_last_error = error;
+ }
+
+ return error;
+}
+
static const char *spi_hid_power_mode_string(enum hidspi_power_state power_state)
{
switch (power_state) {
@@ -84,11 +185,416 @@ static const char *spi_hid_power_mode_string(enum hidspi_power_state power_state
}
}
+static void spi_hid_stop_hid(struct spi_hid *shid)
+{
+ struct hid_device *hid = shid->hid;
+
+ shid->hid = NULL;
+ clear_bit(SPI_HID_READY, &shid->flags);
+
+ if (hid)
+ hid_destroy_device(hid);
+}
+
+static int spi_hid_send_output_report(struct spi_hid *shid,
+ struct spi_hid_output_report *report)
+{
+ struct spi_hid_output_buf *buf = shid->output;
+ struct device *dev = &shid->spi->dev;
+ u16 report_length;
+ u16 padded_length;
+ u8 padding;
+ int error;
+
+ guard(mutex)(&shid->output_lock);
+ if (report->content_length > shid->desc.max_output_length) {
+ dev_err(dev, "Output report too big, content_length 0x%x.",
+ report->content_length);
+ return -E2BIG;
+ }
+
+ spi_hid_populate_output_header(buf->header, shid->conf, report);
+
+ if (report->content_length)
+ memcpy(&buf->content, report->content, report->content_length);
+
+ report_length = sizeof(buf->header) + report->content_length;
+ padded_length = round_up(report_length, 4);
+ padding = padded_length - report_length;
+ memset(&buf->content[report->content_length], 0, padding);
+
+ error = spi_hid_output(shid, buf, padded_length);
+ if (error)
+ dev_err(dev, "Failed output transfer: %d.", error);
+
+ return error;
+}
+
+static int spi_hid_sync_request(struct spi_hid *shid,
+ struct spi_hid_output_report *report)
+{
+ struct device *dev = &shid->spi->dev;
+ int error;
+
+ error = spi_hid_send_output_report(shid, report);
+ if (error)
+ return error;
+
+ error = wait_for_completion_interruptible_timeout(&shid->output_done,
+ msecs_to_jiffies(SPI_HID_RESP_TIMEOUT));
+ if (error == 0) {
+ dev_err(dev, "Response timed out.");
+ return -ETIMEDOUT;
+ }
+
+ return 0;
+}
+
+/*
+ * This function returns the length of the report descriptor, or a negative
+ * error code if something went wrong.
+ */
+static int spi_hid_report_descriptor_request(struct spi_hid *shid)
+{
+ struct device *dev = &shid->spi->dev;
+ struct spi_hid_output_report report = {
+ .report_type = REPORT_DESCRIPTOR,
+ .content_length = 0,
+ .content_id = SPI_HID_OUTPUT_REPORT_CONTENT_ID_DESC_REQUEST,
+ .content = NULL,
+ };
+ int ret;
+
+ ret = spi_hid_sync_request(shid, &report);
+ if (ret) {
+ dev_err(dev,
+ "Expected report descriptor not received: %d.", ret);
+ return ret;
+ }
+
+ ret = shid->response_length;
+ if (ret != shid->desc.report_descriptor_length) {
+ ret = min_t(unsigned int, ret, shid->desc.report_descriptor_length);
+ dev_err(dev, "Received report descriptor length doesn't match device descriptor field, using min of the two: %d.",
+ ret);
+ }
+
+ return ret;
+}
+
+static int spi_hid_create_device(struct spi_hid *shid)
+{
+ struct hid_device *hid;
+ struct device *dev = &shid->spi->dev;
+ int error;
+
+ hid = hid_allocate_device();
+ error = PTR_ERR_OR_ZERO(hid);
+ if (error) {
+ dev_err(dev, "Failed to allocate hid device: %d.", error);
+ return error;
+ }
+
+ hid->driver_data = shid->spi;
+ hid->ll_driver = &spi_hid_ll_driver;
+ hid->dev.parent = &shid->spi->dev;
+ hid->bus = BUS_SPI;
+ hid->version = shid->desc.hid_version;
+ hid->vendor = shid->desc.vendor_id;
+ hid->product = shid->desc.product_id;
+
+ snprintf(hid->name, sizeof(hid->name), "spi %04X:%04X",
+ hid->vendor, hid->product);
+ strscpy(hid->phys, dev_name(&shid->spi->dev), sizeof(hid->phys));
+
+ shid->hid = hid;
+
+ error = hid_add_device(hid);
+ if (error) {
+ dev_err(dev, "Failed to add hid device: %d.", error);
+ /*
+ * We likely got here because report descriptor request timed
+ * out. Let's disconnect and destroy the hid_device structure.
+ */
+ spi_hid_stop_hid(shid);
+ return error;
+ }
+
+ return 0;
+}
+
+static int spi_hid_get_request(struct spi_hid *shid, u8 content_id)
+{
+ struct device *dev = &shid->spi->dev;
+ struct spi_hid_output_report report = {
+ .report_type = GET_FEATURE,
+ .content_length = 0,
+ .content_id = content_id,
+ .content = NULL,
+ };
+ int error;
+
+ error = spi_hid_sync_request(shid, &report);
+ if (error) {
+ dev_err(dev,
+ "Expected get request response not received! Error %d.",
+ error);
+ return error;
+ }
+
+ return 0;
+}
+
+static int spi_hid_set_request(struct spi_hid *shid, u8 *arg_buf, u16 arg_len,
+ u8 content_id)
+{
+ struct spi_hid_output_report report = {
+ .report_type = SET_FEATURE,
+ .content_length = arg_len,
+ .content_id = content_id,
+ .content = arg_buf,
+ };
+
+ return spi_hid_sync_request(shid, &report);
+}
+
+/* This is a placeholder. Will be implemented in the next patch. */
static irqreturn_t spi_hid_dev_irq(int irq, void *_shid)
{
return IRQ_HANDLED;
}
+static int spi_hid_alloc_buffers(struct spi_hid *shid, size_t report_size)
+{
+ struct device *dev = &shid->spi->dev;
+ int inbufsize = sizeof(shid->input->header) + sizeof(shid->input->body) + report_size;
+ int outbufsize = sizeof(shid->output->header) + report_size;
+
+ // devm_krealloc with __GFP_ZERO ensures the new memory is initialized
+ shid->output = devm_krealloc(dev, shid->output, outbufsize, GFP_KERNEL | __GFP_ZERO);
+ shid->input = devm_krealloc(dev, shid->input, inbufsize, GFP_KERNEL | __GFP_ZERO);
+ shid->response = devm_krealloc(dev, shid->response, inbufsize, GFP_KERNEL | __GFP_ZERO);
+
+ if (!shid->output || !shid->input || !shid->response)
+ return -ENOMEM;
+
+ shid->bufsize = report_size;
+
+ return 0;
+}
+
+static int spi_hid_get_report_length(struct hid_report *report)
+{
+ return ((report->size - 1) >> 3) + 1 +
+ report->device->report_enum[report->type].numbered + 2;
+}
+
+/*
+ * Traverse the supplied list of reports and find the longest
+ */
+static void spi_hid_find_max_report(struct hid_device *hid, u32 type,
+ u16 *max)
+{
+ struct hid_report *report;
+ u16 size;
+
+ /*
+ * We should not rely on wMaxInputLength, as some devices may set it to
+ * a wrong length.
+ */
+ list_for_each_entry(report, &hid->report_enum[type].report_list, list) {
+ size = spi_hid_get_report_length(report);
+ if (*max < size)
+ *max = size;
+ }
+}
+
+/* hid_ll_driver interface functions */
+
+static int spi_hid_ll_start(struct hid_device *hid)
+{
+ struct spi_device *spi = hid->driver_data;
+ struct spi_hid *shid = spi_get_drvdata(spi);
+ int error = 0;
+ u16 bufsize = 0;
+
+ spi_hid_find_max_report(hid, HID_INPUT_REPORT, &bufsize);
+ spi_hid_find_max_report(hid, HID_OUTPUT_REPORT, &bufsize);
+ spi_hid_find_max_report(hid, HID_FEATURE_REPORT, &bufsize);
+
+ if (bufsize < HID_MIN_BUFFER_SIZE) {
+ dev_err(&spi->dev,
+ "HID_MIN_BUFFER_SIZE > max_input_length (%d).",
+ bufsize);
+ return -EINVAL;
+ }
+
+ if (bufsize > shid->bufsize) {
+ guard(disable_irq)(&shid->spi->irq);
+
+ error = spi_hid_alloc_buffers(shid, bufsize);
+ if (error)
+ return error;
+ }
+
+ return 0;
+}
+
+static void spi_hid_ll_stop(struct hid_device *hid)
+{
+ hid->claimed = 0;
+}
+
+static int spi_hid_ll_open(struct hid_device *hid)
+{
+ struct spi_device *spi = hid->driver_data;
+ struct spi_hid *shid = spi_get_drvdata(spi);
+
+ set_bit(SPI_HID_READY, &shid->flags);
+ return 0;
+}
+
+static void spi_hid_ll_close(struct hid_device *hid)
+{
+ struct spi_device *spi = hid->driver_data;
+ struct spi_hid *shid = spi_get_drvdata(spi);
+
+ clear_bit(SPI_HID_READY, &shid->flags);
+ shid->reset_attempts = 0;
+}
+
+static int spi_hid_ll_power(struct hid_device *hid, int level)
+{
+ struct spi_device *spi = hid->driver_data;
+ struct spi_hid *shid = spi_get_drvdata(spi);
+ int error = 0;
+
+ guard(mutex)(&shid->output_lock);
+ if (!shid->hid)
+ error = -ENODEV;
+
+ return error;
+}
+
+static int spi_hid_ll_parse(struct hid_device *hid)
+{
+ struct spi_device *spi = hid->driver_data;
+ struct spi_hid *shid = spi_get_drvdata(spi);
+ struct device *dev = &spi->dev;
+ int error, len;
+
+ len = spi_hid_report_descriptor_request(shid);
+ if (len < 0) {
+ dev_err(dev, "Report descriptor request failed, %d.", len);
+ return len;
+ }
+
+ /*
+ * FIXME: below call returning 0 doesn't mean that the report descriptor
+ * is good. We might be caching a crc32 of a corrupted r. d. or who
+ * knows what the FW sent. Need to have a feedback loop about r. d.
+ * being ok and only then cache it.
+ */
+ error = hid_parse_report(hid, (u8 *)shid->response->content, len);
+ if (error) {
+ dev_err(dev, "failed parsing report: %d.", error);
+ return error;
+ }
+ shid->report_descriptor_crc32 = crc32_le(0,
+ (unsigned char const *)shid->response->content,
+ len);
+
+ return 0;
+}
+
+static int spi_hid_ll_raw_request(struct hid_device *hid,
+ unsigned char reportnum, __u8 *buf,
+ size_t len, unsigned char rtype, int reqtype)
+{
+ struct spi_device *spi = hid->driver_data;
+ struct spi_hid *shid = spi_get_drvdata(spi);
+ struct device *dev = &spi->dev;
+ int ret;
+
+ switch (reqtype) {
+ case HID_REQ_SET_REPORT:
+ if (buf[0] != reportnum) {
+ dev_err(dev, "report id mismatch.");
+ return -EINVAL;
+ }
+
+ ret = spi_hid_set_request(shid, &buf[1], len - 1,
+ reportnum);
+ if (ret) {
+ dev_err(dev, "failed to set report.");
+ return ret;
+ }
+
+ ret = len;
+ break;
+ case HID_REQ_GET_REPORT:
+ ret = spi_hid_get_request(shid, reportnum);
+ if (ret) {
+ dev_err(dev, "failed to get report.");
+ return ret;
+ }
+
+ ret = min_t(size_t, len,
+ (shid->response->body[1] | (shid->response->body[2] << 8)) + 1);
+ buf[0] = shid->response->body[3];
+ memcpy(&buf[1], &shid->response->content, ret);
+ break;
+ default:
+ dev_err(dev, "invalid request type.");
+ return -EIO;
+ }
+
+ return ret;
+}
+
+static int spi_hid_ll_output_report(struct hid_device *hid, __u8 *buf,
+ size_t len)
+{
+ struct spi_device *spi = hid->driver_data;
+ struct spi_hid *shid = spi_get_drvdata(spi);
+ struct device *dev = &spi->dev;
+ struct spi_hid_output_report report = {
+ .report_type = OUTPUT_REPORT,
+ .content_length = len - 1,
+ .content_id = buf[0],
+ .content = &buf[1],
+ };
+ int error;
+
+ if (!test_bit(SPI_HID_READY, &shid->flags)) {
+ dev_err(dev, "%s called in unready state", __func__);
+ return -ENODEV;
+ }
+
+ if (shid->desc.no_output_report_ack)
+ error = spi_hid_send_output_report(shid, &report);
+ else
+ error = spi_hid_sync_request(shid, &report);
+
+ if (error) {
+ dev_err(dev, "failed to send output report.");
+ return error;
+ }
+
+ return len;
+}
+
+static struct hid_ll_driver spi_hid_ll_driver = {
+ .start = spi_hid_ll_start,
+ .stop = spi_hid_ll_stop,
+ .open = spi_hid_ll_open,
+ .close = spi_hid_ll_close,
+ .power = spi_hid_ll_power,
+ .parse = spi_hid_ll_parse,
+ .output_report = spi_hid_ll_output_report,
+ .raw_request = spi_hid_ll_raw_request,
+};
+
static ssize_t bus_error_count_show(struct device *dev,
struct device_attribute *attr, char *buf)
{
@@ -159,6 +665,15 @@ int spi_hid_core_probe(struct spi_device *spi, struct spihid_ops *ops,
spi_set_drvdata(spi, shid);
+ /*
+ * we need to allocate the buffer without knowing the maximum
+ * size of the reports. Let's use SZ_2K, then we do the
+ * real computation later.
+ */
+ error = spi_hid_alloc_buffers(shid, SZ_2K);
+ if (error)
+ return error;
+
/*
* At the end of probe we initialize the device:
* 0) assert reset, bias the interrupt line
@@ -191,6 +706,8 @@ int spi_hid_core_probe(struct spi_device *spi, struct spihid_ops *ops,
dev_dbg(dev, "%s: d3 -> %s.", __func__,
spi_hid_power_mode_string(shid->power_state));
+ spi_hid_create_device(shid);
+
return 0;
}
EXPORT_SYMBOL_GPL(spi_hid_core_probe);
@@ -201,6 +718,8 @@ void spi_hid_core_remove(struct spi_device *spi)
struct device *dev = &spi->dev;
int error;
+ spi_hid_stop_hid(shid);
+
shid->ops->assert_reset(shid->ops);
error = shid->ops->power_down(shid->ops);
if (error)
--
2.53.0.1185.g05d4b7b318-goog
^ permalink raw reply related
* [PATCH v3 05/11] HID: spi-hid: add HID SPI protocol implementation
From: Jingyuan Liang @ 2026-04-02 1:59 UTC (permalink / raw)
To: Jiri Kosina, Benjamin Tissoires, Jonathan Corbet, Mark Brown,
Steven Rostedt, Masami Hiramatsu, Mathieu Desnoyers,
Dmitry Torokhov, Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: linux-input, linux-doc, linux-kernel, linux-spi,
linux-trace-kernel, devicetree, hbarnor, tfiga, Jingyuan Liang,
Dmitry Antipov, Angela Czubak
In-Reply-To: <20260402-send-upstream-v3-0-6091c458d357@chromium.org>
This driver follows HID Over SPI Protocol Specification 1.0 available at
https://www.microsoft.com/en-us/download/details.aspx?id=103325. The
initial version of the driver does not support: 1) multi-fragment input
reports, 2) sending GET_INPUT and COMMAND output report types and
processing their respective acknowledge input reports, and 3) device
sleep power state.
Signed-off-by: Dmitry Antipov <dmanti@microsoft.com>
Signed-off-by: Angela Czubak <acz@semihalf.com>
Signed-off-by: Jingyuan Liang <jingyliang@chromium.org>
---
drivers/hid/spi-hid/spi-hid-core.c | 582 ++++++++++++++++++++++++++++++++++++-
1 file changed, 572 insertions(+), 10 deletions(-)
diff --git a/drivers/hid/spi-hid/spi-hid-core.c b/drivers/hid/spi-hid/spi-hid-core.c
index 4723b87346d4..00b9718ba2c3 100644
--- a/drivers/hid/spi-hid/spi-hid-core.c
+++ b/drivers/hid/spi-hid/spi-hid-core.c
@@ -23,11 +23,16 @@
#include <linux/completion.h>
#include <linux/crc32.h>
#include <linux/device.h>
+#include <linux/dma-mapping.h>
#include <linux/err.h>
#include <linux/hid.h>
#include <linux/hid-over-spi.h>
+#include <linux/input.h>
#include <linux/interrupt.h>
+#include <linux/irq.h>
#include <linux/jiffies.h>
+#include <linux/kernel.h>
+#include <linux/list.h>
#include <linux/module.h>
#include <linux/mutex.h>
#include <linux/slab.h>
@@ -35,12 +40,22 @@
#include <linux/string.h>
#include <linux/sysfs.h>
#include <linux/unaligned.h>
+#include <linux/wait.h>
+#include <linux/workqueue.h>
+
+/* Protocol constants */
+#define SPI_HID_READ_APPROVAL_CONSTANT 0xff
+#define SPI_HID_INPUT_HEADER_SYNC_BYTE 0x5a
+#define SPI_HID_INPUT_HEADER_VERSION 0x03
+#define SPI_HID_SUPPORTED_VERSION 0x0300
#define SPI_HID_OUTPUT_REPORT_CONTENT_ID_DESC_REQUEST 0x00
-#define SPI_HID_RESP_TIMEOUT 1000
+#define SPI_HID_MAX_RESET_ATTEMPTS 3
+#define SPI_HID_RESP_TIMEOUT 1000
/* Protocol message size constants */
+#define SPI_HID_READ_APPROVAL_LEN 5
#define SPI_HID_OUTPUT_HEADER_LEN 8
/* flags */
@@ -49,6 +64,22 @@
* requests. The FW becomes ready after sending the report descriptor.
*/
#define SPI_HID_READY 0
+/*
+ * refresh_in_progress is set to true while the refresh_device worker
+ * thread is destroying and recreating the hidraw device. When this flag
+ * is set to true, the ll_close and ll_open functions will not cause
+ * power state changes.
+ */
+#define SPI_HID_REFRESH_IN_PROGRESS 1
+/*
+ * reset_pending indicates that the device is being reset. When this flag
+ * is set to true, garbage interrupts triggered during reset will be
+ * dropped and will not cause error handling.
+ */
+#define SPI_HID_RESET_PENDING 2
+#define SPI_HID_RESET_RESPONSE 3
+#define SPI_HID_CREATE_DEVICE 4
+#define SPI_HID_ERROR 5
/* Raw input buffer with data from the bus */
struct spi_hid_input_buf {
@@ -57,6 +88,22 @@ struct spi_hid_input_buf {
u8 content[];
};
+/* Processed data from input report header */
+struct spi_hid_input_header {
+ u8 version;
+ u16 report_length;
+ u8 last_fragment_flag;
+ u8 sync_const;
+};
+
+/* Processed data from an input report */
+struct spi_hid_input_report {
+ u8 report_type;
+ u16 content_length;
+ u8 content_id;
+ u8 *content;
+};
+
/* Raw output report buffer to be put on the bus */
struct spi_hid_output_buf {
u8 header[SPI_HID_OUTPUT_HEADER_LEN];
@@ -114,6 +161,9 @@ struct spi_hid {
struct spi_device *spi; /* spi device. */
struct hid_device *hid; /* pointer to corresponding HID dev. */
+ struct spi_transfer input_transfer[2]; /* Transfer buffer for read and write. */
+ struct spi_message input_message; /* used to execute a sequence of spi transfers. */
+
struct spihid_ops *ops;
struct spi_hid_conf *conf;
@@ -131,10 +181,20 @@ struct spi_hid {
unsigned long flags; /* device flags. */
- /* Control lock to make sure one output transaction at a time. */
+ struct work_struct reset_work;
+
+ /* Control lock to ensure complete output transaction. */
struct mutex output_lock;
+ /* Power lock to make sure one power state change at a time. */
+ struct mutex power_lock;
+ /* I/O lock to prevent concurrent output writes during the input read. */
+ struct mutex io_lock;
+
struct completion output_done;
+ u8 read_approval_header[SPI_HID_READ_APPROVAL_LEN];
+ u8 read_approval_body[SPI_HID_READ_APPROVAL_LEN];
+
u32 report_descriptor_crc32; /* HID report descriptor crc32 checksum. */
u32 regulator_error_count;
@@ -146,6 +206,66 @@ struct spi_hid {
static struct hid_ll_driver spi_hid_ll_driver;
+static void spi_hid_populate_read_approvals(const struct spi_hid_conf *conf,
+ u8 *header_buf, u8 *body_buf)
+{
+ header_buf[0] = conf->read_opcode;
+ put_unaligned_be24(conf->input_report_header_address, &header_buf[1]);
+ header_buf[4] = SPI_HID_READ_APPROVAL_CONSTANT;
+
+ body_buf[0] = conf->read_opcode;
+ put_unaligned_be24(conf->input_report_body_address, &body_buf[1]);
+ body_buf[4] = SPI_HID_READ_APPROVAL_CONSTANT;
+}
+
+static void spi_hid_parse_dev_desc(const struct hidspi_dev_descriptor *raw,
+ struct spi_hid_device_descriptor *desc)
+{
+ desc->hid_version = le16_to_cpu(raw->bcd_ver);
+ desc->report_descriptor_length = le16_to_cpu(raw->rep_desc_len);
+ desc->max_input_length = le16_to_cpu(raw->max_input_len);
+ desc->max_output_length = le16_to_cpu(raw->max_output_len);
+
+ /* FIXME: multi-fragment not supported, field below not used */
+ desc->max_fragment_length = le16_to_cpu(raw->max_frag_len);
+
+ desc->vendor_id = le16_to_cpu(raw->vendor_id);
+ desc->product_id = le16_to_cpu(raw->product_id);
+ desc->version_id = le16_to_cpu(raw->version_id);
+ desc->no_output_report_ack = le16_to_cpu(raw->flags) & BIT(0);
+}
+
+static void spi_hid_populate_input_header(const u8 *buf,
+ struct spi_hid_input_header *header)
+{
+ header->version = buf[0] & 0xf;
+ header->report_length = (get_unaligned_le16(&buf[1]) & 0x3fff) * 4;
+ header->last_fragment_flag = (buf[2] & 0x40) >> 6;
+ header->sync_const = buf[3];
+}
+
+static void spi_hid_populate_input_body(const u8 *buf,
+ struct input_report_body_header *body)
+{
+ body->input_report_type = buf[0];
+ body->content_len = get_unaligned_le16(&buf[1]);
+ body->content_id = buf[3];
+}
+
+static void spi_hid_input_report_prepare(struct spi_hid_input_buf *buf,
+ struct spi_hid_input_report *report)
+{
+ struct spi_hid_input_header header;
+ struct input_report_body_header body;
+
+ spi_hid_populate_input_header(buf->header, &header);
+ spi_hid_populate_input_body(buf->body, &body);
+ report->report_type = body.input_report_type;
+ report->content_length = body.content_len;
+ report->content_id = body.content_id;
+ report->content = buf->content;
+}
+
static void spi_hid_populate_output_header(u8 *buf,
const struct spi_hid_conf *conf,
const struct spi_hid_output_report *report)
@@ -157,6 +277,33 @@ static void spi_hid_populate_output_header(u8 *buf,
buf[7] = report->content_id;
}
+static int spi_hid_input_sync(struct spi_hid *shid, void *buf, u16 length,
+ bool is_header)
+{
+ int error;
+
+ shid->input_transfer[0].tx_buf = is_header ?
+ shid->read_approval_header :
+ shid->read_approval_body;
+ shid->input_transfer[0].len = SPI_HID_READ_APPROVAL_LEN;
+
+ shid->input_transfer[1].rx_buf = buf;
+ shid->input_transfer[1].len = length;
+
+ spi_message_init_with_transfers(&shid->input_message,
+ shid->input_transfer, 2);
+
+ error = spi_sync(shid->spi, &shid->input_message);
+ if (error) {
+ dev_err(&shid->spi->dev, "Error starting sync transfer: %d.", error);
+ shid->bus_error_count++;
+ shid->bus_last_error = error;
+ return error;
+ }
+
+ return 0;
+}
+
static int spi_hid_output(struct spi_hid *shid, const void *buf, u16 length)
{
int error;
@@ -196,6 +343,50 @@ static void spi_hid_stop_hid(struct spi_hid *shid)
hid_destroy_device(hid);
}
+static void spi_hid_error(struct spi_hid *shid)
+{
+ struct device *dev = &shid->spi->dev;
+ int error;
+
+ guard(mutex)(&shid->power_lock);
+ if (shid->power_state == HIDSPI_OFF)
+ return;
+
+ if (shid->reset_attempts++ >= SPI_HID_MAX_RESET_ATTEMPTS) {
+ dev_err(dev, "unresponsive device, aborting.");
+ spi_hid_stop_hid(shid);
+ shid->ops->assert_reset(shid->ops);
+ error = shid->ops->power_down(shid->ops);
+ if (error) {
+ dev_err(dev, "failed to disable regulator.");
+ shid->regulator_error_count++;
+ shid->regulator_last_error = error;
+ }
+ return;
+ }
+
+ clear_bit(SPI_HID_READY, &shid->flags);
+ set_bit(SPI_HID_RESET_PENDING, &shid->flags);
+
+ shid->ops->assert_reset(shid->ops);
+
+ shid->power_state = HIDSPI_OFF;
+
+ /*
+ * We want to cancel pending reset work as the device is being reset
+ * to recover from an error. cancel_work_sync will put us in a deadlock
+ * because this function is scheduled in 'reset_work' and we should
+ * avoid waiting for itself.
+ */
+ cancel_work(&shid->reset_work);
+
+ shid->ops->sleep_minimal_reset_delay(shid->ops);
+
+ shid->power_state = HIDSPI_ON;
+
+ shid->ops->deassert_reset(shid->ops);
+}
+
static int spi_hid_send_output_report(struct spi_hid *shid,
struct spi_hid_output_report *report)
{
@@ -206,13 +397,13 @@ static int spi_hid_send_output_report(struct spi_hid *shid,
u8 padding;
int error;
- guard(mutex)(&shid->output_lock);
if (report->content_length > shid->desc.max_output_length) {
dev_err(dev, "Output report too big, content_length 0x%x.",
report->content_length);
return -E2BIG;
}
+ guard(mutex)(&shid->io_lock);
spi_hid_populate_output_header(buf->header, shid->conf, report);
if (report->content_length)
@@ -236,6 +427,7 @@ static int spi_hid_sync_request(struct spi_hid *shid,
struct device *dev = &shid->spi->dev;
int error;
+ guard(mutex)(&shid->output_lock);
error = spi_hid_send_output_report(shid, report);
if (error)
return error;
@@ -250,6 +442,86 @@ static int spi_hid_sync_request(struct spi_hid *shid,
return 0;
}
+/*
+ * Handle the reset response from the FW by sending a request for the device
+ * descriptor.
+ */
+static void spi_hid_reset_response(struct spi_hid *shid)
+{
+ struct device *dev = &shid->spi->dev;
+ struct spi_hid_output_report report = {
+ .report_type = DEVICE_DESCRIPTOR,
+ .content_length = 0x0,
+ .content_id = SPI_HID_OUTPUT_REPORT_CONTENT_ID_DESC_REQUEST,
+ .content = NULL,
+ };
+ int error;
+
+ if (test_bit(SPI_HID_READY, &shid->flags)) {
+ dev_err(dev, "Spontaneous FW reset!");
+ clear_bit(SPI_HID_READY, &shid->flags);
+ shid->dir_count++;
+ }
+
+ if (shid->power_state == HIDSPI_OFF)
+ return;
+
+ error = spi_hid_sync_request(shid, &report);
+ if (error) {
+ dev_WARN_ONCE(dev, true,
+ "Failed to send device descriptor request: %d.", error);
+ set_bit(SPI_HID_ERROR, &shid->flags);
+ schedule_work(&shid->reset_work);
+ }
+}
+
+static int spi_hid_input_report_handler(struct spi_hid *shid,
+ struct spi_hid_input_buf *buf)
+{
+ struct device *dev = &shid->spi->dev;
+ struct spi_hid_input_report r;
+ int error = 0;
+
+ if (!test_bit(SPI_HID_READY, &shid->flags) ||
+ test_bit(SPI_HID_REFRESH_IN_PROGRESS, &shid->flags) || !shid->hid) {
+ dev_err(dev, "HID not ready");
+ return 0;
+ }
+
+ spi_hid_input_report_prepare(buf, &r);
+
+ error = hid_input_report(shid->hid, HID_INPUT_REPORT,
+ r.content - 1, r.content_length + 1, 1);
+
+ if (error == -ENODEV || error == -EBUSY) {
+ dev_err(dev, "ignoring report --> %d.", error);
+ return 0;
+ } else if (error) {
+ dev_err(dev, "Bad input report: %d.", error);
+ }
+
+ return error;
+}
+
+static void spi_hid_response_handler(struct spi_hid *shid,
+ struct input_report_body_header *body)
+{
+ shid->response_length = body->content_len;
+ /* completion_done returns 0 if there are waiters, otherwise 1 */
+ if (completion_done(&shid->output_done)) {
+ dev_err(&shid->spi->dev, "Unexpected response report.");
+ } else {
+ if (body->input_report_type == REPORT_DESCRIPTOR_RESPONSE ||
+ body->input_report_type == GET_FEATURE_RESPONSE) {
+ memcpy(shid->response->body, shid->input->body,
+ sizeof(shid->input->body));
+ memcpy(shid->response->content, shid->input->content,
+ body->content_len);
+ }
+ complete(&shid->output_done);
+ }
+}
+
/*
* This function returns the length of the report descriptor, or a negative
* error code if something went wrong.
@@ -269,6 +541,8 @@ static int spi_hid_report_descriptor_request(struct spi_hid *shid)
if (ret) {
dev_err(dev,
"Expected report descriptor not received: %d.", ret);
+ set_bit(SPI_HID_ERROR, &shid->flags);
+ schedule_work(&shid->reset_work);
return ret;
}
@@ -323,6 +597,205 @@ static int spi_hid_create_device(struct spi_hid *shid)
return 0;
}
+static void spi_hid_refresh_device(struct spi_hid *shid)
+{
+ struct device *dev = &shid->spi->dev;
+ u32 new_crc32 = 0;
+ int error = 0;
+
+ error = spi_hid_report_descriptor_request(shid);
+ if (error < 0) {
+ dev_err(dev,
+ "%s: failed report descriptor request: %d",
+ __func__, error);
+ return;
+ }
+ new_crc32 = crc32_le(0, (unsigned char const *)shid->response->content,
+ (size_t)error);
+
+ /* Same report descriptor, so no need to create a new hid device. */
+ if (new_crc32 == shid->report_descriptor_crc32) {
+ set_bit(SPI_HID_READY, &shid->flags);
+ return;
+ }
+
+ shid->report_descriptor_crc32 = new_crc32;
+
+ set_bit(SPI_HID_REFRESH_IN_PROGRESS, &shid->flags);
+
+ spi_hid_stop_hid(shid);
+
+ error = spi_hid_create_device(shid);
+ if (error) {
+ dev_err(dev, "%s: Failed to create hid device: %d.", __func__, error);
+ return;
+ }
+
+ clear_bit(SPI_HID_REFRESH_IN_PROGRESS, &shid->flags);
+}
+
+static void spi_hid_reset_work(struct work_struct *work)
+{
+ struct spi_hid *shid =
+ container_of(work, struct spi_hid, reset_work);
+ struct device *dev = &shid->spi->dev;
+ int error = 0;
+
+ if (test_and_clear_bit(SPI_HID_RESET_RESPONSE, &shid->flags)) {
+ spi_hid_reset_response(shid);
+ return;
+ }
+
+ if (test_and_clear_bit(SPI_HID_CREATE_DEVICE, &shid->flags)) {
+ guard(mutex)(&shid->power_lock);
+ if (shid->power_state == HIDSPI_OFF) {
+ dev_err(dev, "%s: Powered off, returning", __func__);
+ return;
+ }
+
+ if (!shid->hid) {
+ error = spi_hid_create_device(shid);
+ if (error) {
+ dev_err(dev, "%s: Failed to create hid device: %d.",
+ __func__, error);
+ return;
+ }
+ } else {
+ spi_hid_refresh_device(shid);
+ }
+
+ return;
+ }
+
+ if (test_and_clear_bit(SPI_HID_ERROR, &shid->flags)) {
+ spi_hid_error(shid);
+ return;
+ }
+}
+
+static int spi_hid_process_input_report(struct spi_hid *shid,
+ struct spi_hid_input_buf *buf)
+{
+ struct spi_hid_input_header header;
+ struct input_report_body_header body;
+ struct device *dev = &shid->spi->dev;
+ struct hidspi_dev_descriptor *raw;
+
+ spi_hid_populate_input_header(buf->header, &header);
+ spi_hid_populate_input_body(buf->body, &body);
+
+ if (body.content_len > header.report_length) {
+ dev_err(dev, "Bad body length %d > %d.", body.content_len,
+ header.report_length);
+ return -EPROTO;
+ }
+
+ switch (body.input_report_type) {
+ case DATA:
+ return spi_hid_input_report_handler(shid, buf);
+ case RESET_RESPONSE:
+ clear_bit(SPI_HID_RESET_PENDING, &shid->flags);
+ set_bit(SPI_HID_RESET_RESPONSE, &shid->flags);
+ schedule_work(&shid->reset_work);
+ break;
+ case DEVICE_DESCRIPTOR_RESPONSE:
+ /* Mark the completion done to avoid timeout */
+ spi_hid_response_handler(shid, &body);
+
+ /* Reset attempts at every device descriptor fetch */
+ shid->reset_attempts = 0;
+ raw = (struct hidspi_dev_descriptor *)buf->content;
+
+ /* Validate device descriptor length before parsing */
+ if (body.content_len != HIDSPI_DEVICE_DESCRIPTOR_SIZE) {
+ dev_err(dev, "Invalid content length %d, expected %lu.",
+ body.content_len,
+ HIDSPI_DEVICE_DESCRIPTOR_SIZE);
+ return -EPROTO;
+ }
+
+ if (le16_to_cpu(raw->dev_desc_len) !=
+ HIDSPI_DEVICE_DESCRIPTOR_SIZE) {
+ dev_err(dev,
+ "Invalid wDeviceDescLength %d, expected %lu.",
+ raw->dev_desc_len,
+ HIDSPI_DEVICE_DESCRIPTOR_SIZE);
+ return -EPROTO;
+ }
+
+ spi_hid_parse_dev_desc(raw, &shid->desc);
+
+ if (shid->desc.hid_version != SPI_HID_SUPPORTED_VERSION) {
+ dev_err(dev,
+ "Unsupported device descriptor version %4x.",
+ shid->desc.hid_version);
+ return -EPROTONOSUPPORT;
+ }
+
+ set_bit(SPI_HID_CREATE_DEVICE, &shid->flags);
+ schedule_work(&shid->reset_work);
+
+ break;
+ case OUTPUT_REPORT_RESPONSE:
+ if (shid->desc.no_output_report_ack) {
+ dev_err(dev, "Unexpected output report response.");
+ break;
+ }
+ fallthrough;
+ case GET_FEATURE_RESPONSE:
+ case SET_FEATURE_RESPONSE:
+ case REPORT_DESCRIPTOR_RESPONSE:
+ spi_hid_response_handler(shid, &body);
+ break;
+ /*
+ * FIXME: sending GET_INPUT and COMMAND reports not supported, thus
+ * throw away responses to those, they should never come.
+ */
+ case GET_INPUT_REPORT_RESPONSE:
+ case COMMAND_RESPONSE:
+ dev_err(dev, "Not a supported report type: 0x%x.",
+ body.input_report_type);
+ break;
+ default:
+ dev_err(dev, "Unknown input report: 0x%x.", body.input_report_type);
+ return -EPROTO;
+ }
+
+ return 0;
+}
+
+static int spi_hid_bus_validate_header(struct spi_hid *shid,
+ struct spi_hid_input_header *header)
+{
+ struct device *dev = &shid->spi->dev;
+
+ if (header->version != SPI_HID_INPUT_HEADER_VERSION) {
+ dev_err(dev, "Unknown input report version (v 0x%x).",
+ header->version);
+ return -EINVAL;
+ }
+
+ if (shid->desc.max_input_length != 0 &&
+ header->report_length > shid->desc.max_input_length) {
+ dev_err(dev, "Input report body size %u > max expected of %u.",
+ header->report_length, shid->desc.max_input_length);
+ return -EMSGSIZE;
+ }
+
+ if (header->last_fragment_flag != 1) {
+ dev_err(dev, "Multi-fragment reports not supported.");
+ return -EOPNOTSUPP;
+ }
+
+ if (header->sync_const != SPI_HID_INPUT_HEADER_SYNC_BYTE) {
+ dev_err(dev, "Invalid input report sync constant (0x%x).",
+ header->sync_const);
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
static int spi_hid_get_request(struct spi_hid *shid, u8 content_id)
{
struct device *dev = &shid->spi->dev;
@@ -339,6 +812,8 @@ static int spi_hid_get_request(struct spi_hid *shid, u8 content_id)
dev_err(dev,
"Expected get request response not received! Error %d.",
error);
+ set_bit(SPI_HID_ERROR, &shid->flags);
+ schedule_work(&shid->reset_work);
return error;
}
@@ -358,9 +833,83 @@ static int spi_hid_set_request(struct spi_hid *shid, u8 *arg_buf, u16 arg_len,
return spi_hid_sync_request(shid, &report);
}
-/* This is a placeholder. Will be implemented in the next patch. */
static irqreturn_t spi_hid_dev_irq(int irq, void *_shid)
{
+ struct spi_hid *shid = _shid;
+ struct device *dev = &shid->spi->dev;
+ struct spi_hid_input_header header;
+ int error = 0;
+
+ scoped_guard(mutex, &shid->io_lock) {
+ error = spi_hid_input_sync(shid, shid->input->header,
+ sizeof(shid->input->header), true);
+ if (error) {
+ dev_err(dev, "Failed to transfer header: %d.", error);
+ goto err;
+ }
+
+ if (shid->power_state == HIDSPI_OFF) {
+ dev_warn(dev, "Device is off after header was received.");
+ goto out;
+ }
+
+ if (shid->input_message.status < 0) {
+ dev_warn(dev, "Error reading header: %d.",
+ shid->input_message.status);
+ shid->bus_error_count++;
+ shid->bus_last_error = shid->input_message.status;
+ goto err;
+ }
+
+ spi_hid_populate_input_header(shid->input->header, &header);
+
+ error = spi_hid_bus_validate_header(shid, &header);
+ if (error) {
+ if (!test_bit(SPI_HID_RESET_PENDING, &shid->flags)) {
+ dev_err(dev, "Failed to validate header: %d.", error);
+ print_hex_dump(KERN_ERR, "spi_hid: header buffer: ",
+ DUMP_PREFIX_NONE, 16, 1, shid->input->header,
+ sizeof(shid->input->header), false);
+ shid->bus_error_count++;
+ shid->bus_last_error = error;
+ goto err;
+ }
+ goto out;
+ }
+
+ error = spi_hid_input_sync(shid, shid->input->body, header.report_length,
+ false);
+ if (error) {
+ dev_err(dev, "Failed to transfer body: %d.", error);
+ goto err;
+ }
+
+ if (shid->power_state == HIDSPI_OFF) {
+ dev_warn(dev, "Device is off after body was received.");
+ goto out;
+ }
+
+ if (shid->input_message.status < 0) {
+ dev_warn(dev, "Error reading body: %d.",
+ shid->input_message.status);
+ shid->bus_error_count++;
+ shid->bus_last_error = shid->input_message.status;
+ goto err;
+ }
+ }
+
+ error = spi_hid_process_input_report(shid, shid->input);
+ if (error) {
+ dev_err(dev, "Failed to process input report: %d.", error);
+ goto err;
+ }
+
+out:
+ return IRQ_HANDLED;
+
+err:
+ set_bit(SPI_HID_ERROR, &shid->flags);
+ schedule_work(&shid->reset_work);
return IRQ_HANDLED;
}
@@ -571,10 +1120,13 @@ static int spi_hid_ll_output_report(struct hid_device *hid, __u8 *buf,
return -ENODEV;
}
- if (shid->desc.no_output_report_ack)
- error = spi_hid_send_output_report(shid, &report);
- else
+ if (shid->desc.no_output_report_ack) {
+ scoped_guard(mutex, &shid->output_lock) {
+ error = spi_hid_send_output_report(shid, &report);
+ }
+ } else {
error = spi_hid_sync_request(shid, &report);
+ }
if (error) {
dev_err(dev, "failed to send output report.");
@@ -662,11 +1214,23 @@ int spi_hid_core_probe(struct spi_device *spi, struct spihid_ops *ops,
shid->power_state = HIDSPI_ON;
shid->ops = ops;
shid->conf = conf;
+ set_bit(SPI_HID_RESET_PENDING, &shid->flags);
spi_set_drvdata(spi, shid);
+ /* Using now populated conf let's pre-calculate the read approvals */
+ spi_hid_populate_read_approvals(shid->conf, shid->read_approval_header,
+ shid->read_approval_body);
+
+ mutex_init(&shid->output_lock);
+ mutex_init(&shid->power_lock);
+ mutex_init(&shid->io_lock);
+ init_completion(&shid->output_done);
+
+ INIT_WORK(&shid->reset_work, spi_hid_reset_work);
+
/*
- * we need to allocate the buffer without knowing the maximum
+ * We need to allocate the buffer without knowing the maximum
* size of the reports. Let's use SZ_2K, then we do the
* real computation later.
*/
@@ -706,8 +1270,6 @@ int spi_hid_core_probe(struct spi_device *spi, struct spihid_ops *ops,
dev_dbg(dev, "%s: d3 -> %s.", __func__,
spi_hid_power_mode_string(shid->power_state));
- spi_hid_create_device(shid);
-
return 0;
}
EXPORT_SYMBOL_GPL(spi_hid_core_probe);
--
2.53.0.1185.g05d4b7b318-goog
^ permalink raw reply related
* [PATCH v3 06/11] HID: spi_hid: add spi_hid traces
From: Jingyuan Liang @ 2026-04-02 1:59 UTC (permalink / raw)
To: Jiri Kosina, Benjamin Tissoires, Jonathan Corbet, Mark Brown,
Steven Rostedt, Masami Hiramatsu, Mathieu Desnoyers,
Dmitry Torokhov, Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: linux-input, linux-doc, linux-kernel, linux-spi,
linux-trace-kernel, devicetree, hbarnor, tfiga, Jingyuan Liang,
Dmitry Antipov, Angela Czubak
In-Reply-To: <20260402-send-upstream-v3-0-6091c458d357@chromium.org>
Add traces for purposed of debugging spi_hid driver.
Signed-off-by: Dmitry Antipov <dmanti@microsoft.com>
Signed-off-by: Angela Czubak <acz@semihalf.com>
Signed-off-by: Jingyuan Liang <jingyliang@chromium.org>
---
drivers/hid/spi-hid/Makefile | 1 +
drivers/hid/spi-hid/spi-hid-core.c | 118 +++++++++----------------
drivers/hid/spi-hid/spi-hid-core.h | 91 +++++++++++++++++++
drivers/hid/spi-hid/spi-hid-trace.h | 169 ++++++++++++++++++++++++++++++++++++
4 files changed, 303 insertions(+), 76 deletions(-)
diff --git a/drivers/hid/spi-hid/Makefile b/drivers/hid/spi-hid/Makefile
index 92e24cddbfc2..733e006df56e 100644
--- a/drivers/hid/spi-hid/Makefile
+++ b/drivers/hid/spi-hid/Makefile
@@ -7,3 +7,4 @@
obj-$(CONFIG_SPI_HID_CORE) += spi-hid.o
spi-hid-objs = spi-hid-core.o
+CFLAGS_spi-hid-core.o := -I$(src)
diff --git a/drivers/hid/spi-hid/spi-hid-core.c b/drivers/hid/spi-hid/spi-hid-core.c
index 00b9718ba2c3..802615565541 100644
--- a/drivers/hid/spi-hid/spi-hid-core.c
+++ b/drivers/hid/spi-hid/spi-hid-core.c
@@ -43,6 +43,11 @@
#include <linux/wait.h>
#include <linux/workqueue.h>
+#include "spi-hid-core.h"
+
+#define CREATE_TRACE_POINTS
+#include "spi-hid-trace.h"
+
/* Protocol constants */
#define SPI_HID_READ_APPROVAL_CONSTANT 0xff
#define SPI_HID_INPUT_HEADER_SYNC_BYTE 0x5a
@@ -81,13 +86,6 @@
#define SPI_HID_CREATE_DEVICE 4
#define SPI_HID_ERROR 5
-/* Raw input buffer with data from the bus */
-struct spi_hid_input_buf {
- u8 header[HIDSPI_INPUT_HEADER_SIZE];
- u8 body[HIDSPI_INPUT_BODY_HEADER_SIZE];
- u8 content[];
-};
-
/* Processed data from input report header */
struct spi_hid_input_header {
u8 version;
@@ -104,12 +102,6 @@ struct spi_hid_input_report {
u8 *content;
};
-/* Raw output report buffer to be put on the bus */
-struct spi_hid_output_buf {
- u8 header[SPI_HID_OUTPUT_HEADER_LEN];
- u8 content[];
-};
-
/* Data necessary to send an output report */
struct spi_hid_output_report {
u8 report_type;
@@ -118,19 +110,6 @@ struct spi_hid_output_report {
u8 *content;
};
-/* Processed data from a device descriptor */
-struct spi_hid_device_descriptor {
- u16 hid_version;
- u16 report_descriptor_length;
- u16 max_input_length;
- u16 max_output_length;
- u16 max_fragment_length;
- u16 vendor_id;
- u16 product_id;
- u16 version_id;
- u8 no_output_report_ack;
-};
-
/* struct spi_hid_conf - Conf provided to the core */
struct spi_hid_conf {
u32 input_report_header_address;
@@ -156,54 +135,6 @@ struct spihid_ops {
void (*sleep_minimal_reset_delay)(struct spihid_ops *ops);
};
-/* Driver context */
-struct spi_hid {
- struct spi_device *spi; /* spi device. */
- struct hid_device *hid; /* pointer to corresponding HID dev. */
-
- struct spi_transfer input_transfer[2]; /* Transfer buffer for read and write. */
- struct spi_message input_message; /* used to execute a sequence of spi transfers. */
-
- struct spihid_ops *ops;
- struct spi_hid_conf *conf;
-
- struct spi_hid_device_descriptor desc; /* HID device descriptor. */
- struct spi_hid_output_buf *output; /* Output buffer. */
- struct spi_hid_input_buf *input; /* Input buffer. */
- struct spi_hid_input_buf *response; /* Response buffer. */
-
- u16 response_length;
- u16 bufsize;
-
- enum hidspi_power_state power_state;
-
- u8 reset_attempts; /* The number of reset attempts. */
-
- unsigned long flags; /* device flags. */
-
- struct work_struct reset_work;
-
- /* Control lock to ensure complete output transaction. */
- struct mutex output_lock;
- /* Power lock to make sure one power state change at a time. */
- struct mutex power_lock;
- /* I/O lock to prevent concurrent output writes during the input read. */
- struct mutex io_lock;
-
- struct completion output_done;
-
- u8 read_approval_header[SPI_HID_READ_APPROVAL_LEN];
- u8 read_approval_body[SPI_HID_READ_APPROVAL_LEN];
-
- u32 report_descriptor_crc32; /* HID report descriptor crc32 checksum. */
-
- u32 regulator_error_count;
- int regulator_last_error;
- u32 bus_error_count;
- int bus_last_error;
- u32 dir_count; /* device initiated reset count. */
-};
-
static struct hid_ll_driver spi_hid_ll_driver;
static void spi_hid_populate_read_approvals(const struct spi_hid_conf *conf,
@@ -293,6 +224,11 @@ static int spi_hid_input_sync(struct spi_hid *shid, void *buf, u16 length,
spi_message_init_with_transfers(&shid->input_message,
shid->input_transfer, 2);
+ trace_spi_hid_input_sync(shid, shid->input_transfer[0].tx_buf,
+ shid->input_transfer[0].len,
+ shid->input_transfer[1].rx_buf,
+ shid->input_transfer[1].len, 0);
+
error = spi_sync(shid->spi, &shid->input_message);
if (error) {
dev_err(&shid->spi->dev, "Error starting sync transfer: %d.", error);
@@ -343,11 +279,13 @@ static void spi_hid_stop_hid(struct spi_hid *shid)
hid_destroy_device(hid);
}
-static void spi_hid_error(struct spi_hid *shid)
+static void spi_hid_error_handler(struct spi_hid *shid)
{
struct device *dev = &shid->spi->dev;
int error;
+ trace_spi_hid_error_handler(shid);
+
guard(mutex)(&shid->power_lock);
if (shid->power_state == HIDSPI_OFF)
return;
@@ -457,6 +395,8 @@ static void spi_hid_reset_response(struct spi_hid *shid)
};
int error;
+ trace_spi_hid_reset_response(shid);
+
if (test_bit(SPI_HID_READY, &shid->flags)) {
dev_err(dev, "Spontaneous FW reset!");
clear_bit(SPI_HID_READY, &shid->flags);
@@ -482,6 +422,8 @@ static int spi_hid_input_report_handler(struct spi_hid *shid,
struct spi_hid_input_report r;
int error = 0;
+ trace_spi_hid_input_report_handler(shid);
+
if (!test_bit(SPI_HID_READY, &shid->flags) ||
test_bit(SPI_HID_REFRESH_IN_PROGRESS, &shid->flags) || !shid->hid) {
dev_err(dev, "HID not ready");
@@ -506,6 +448,8 @@ static int spi_hid_input_report_handler(struct spi_hid *shid,
static void spi_hid_response_handler(struct spi_hid *shid,
struct input_report_body_header *body)
{
+ trace_spi_hid_response_handler(shid);
+
shid->response_length = body->content_len;
/* completion_done returns 0 if there are waiters, otherwise 1 */
if (completion_done(&shid->output_done)) {
@@ -562,6 +506,8 @@ static int spi_hid_create_device(struct spi_hid *shid)
struct device *dev = &shid->spi->dev;
int error;
+ trace_spi_hid_create_device(shid);
+
hid = hid_allocate_device();
error = PTR_ERR_OR_ZERO(hid);
if (error) {
@@ -603,6 +549,8 @@ static void spi_hid_refresh_device(struct spi_hid *shid)
u32 new_crc32 = 0;
int error = 0;
+ trace_spi_hid_refresh_device(shid);
+
error = spi_hid_report_descriptor_request(shid);
if (error < 0) {
dev_err(dev,
@@ -668,7 +616,7 @@ static void spi_hid_reset_work(struct work_struct *work)
}
if (test_and_clear_bit(SPI_HID_ERROR, &shid->flags)) {
- spi_hid_error(shid);
+ spi_hid_error_handler(shid);
return;
}
}
@@ -681,6 +629,8 @@ static int spi_hid_process_input_report(struct spi_hid *shid,
struct device *dev = &shid->spi->dev;
struct hidspi_dev_descriptor *raw;
+ trace_spi_hid_process_input_report(shid);
+
spi_hid_populate_input_header(buf->header, &header);
spi_hid_populate_input_body(buf->body, &body);
@@ -840,6 +790,9 @@ static irqreturn_t spi_hid_dev_irq(int irq, void *_shid)
struct spi_hid_input_header header;
int error = 0;
+ trace_spi_hid_dev_irq(shid, irq);
+ trace_spi_hid_header_transfer(shid);
+
scoped_guard(mutex, &shid->io_lock) {
error = spi_hid_input_sync(shid, shid->input->header,
sizeof(shid->input->header), true);
@@ -853,6 +806,13 @@ static irqreturn_t spi_hid_dev_irq(int irq, void *_shid)
goto out;
}
+ trace_spi_hid_input_header_complete(shid,
+ shid->input_transfer[0].tx_buf,
+ shid->input_transfer[0].len,
+ shid->input_transfer[1].rx_buf,
+ shid->input_transfer[1].len,
+ shid->input_message.status);
+
if (shid->input_message.status < 0) {
dev_warn(dev, "Error reading header: %d.",
shid->input_message.status);
@@ -889,6 +849,12 @@ static irqreturn_t spi_hid_dev_irq(int irq, void *_shid)
goto out;
}
+ trace_spi_hid_input_body_complete(shid, shid->input_transfer[0].tx_buf,
+ shid->input_transfer[0].len,
+ shid->input_transfer[1].rx_buf,
+ shid->input_transfer[1].len,
+ shid->input_message.status);
+
if (shid->input_message.status < 0) {
dev_warn(dev, "Error reading body: %d.",
shid->input_message.status);
diff --git a/drivers/hid/spi-hid/spi-hid-core.h b/drivers/hid/spi-hid/spi-hid-core.h
new file mode 100644
index 000000000000..293e2cfcfbf7
--- /dev/null
+++ b/drivers/hid/spi-hid/spi-hid-core.h
@@ -0,0 +1,91 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (c) 2021 Microsoft Corporation
+ * Copyright (c) 2026 Google LLC
+ */
+
+#ifndef SPI_HID_CORE_H
+#define SPI_HID_CORE_H
+
+#include <linux/hid-over-spi.h>
+#include <linux/spi/spi.h>
+
+/* Protocol message size constants */
+#define SPI_HID_READ_APPROVAL_LEN 5
+#define SPI_HID_OUTPUT_HEADER_LEN 8
+
+/* Raw input buffer with data from the bus */
+struct spi_hid_input_buf {
+ u8 header[HIDSPI_INPUT_HEADER_SIZE];
+ u8 body[HIDSPI_INPUT_BODY_HEADER_SIZE];
+ u8 content[];
+};
+
+/* Raw output report buffer to be put on the bus */
+struct spi_hid_output_buf {
+ u8 header[SPI_HID_OUTPUT_HEADER_LEN];
+ u8 content[];
+};
+
+/* Processed data from a device descriptor */
+struct spi_hid_device_descriptor {
+ u16 hid_version;
+ u16 report_descriptor_length;
+ u16 max_input_length;
+ u16 max_output_length;
+ u16 max_fragment_length;
+ u16 vendor_id;
+ u16 product_id;
+ u16 version_id;
+ u8 no_output_report_ack;
+};
+
+/* Driver context */
+struct spi_hid {
+ struct spi_device *spi; /* spi device. */
+ struct hid_device *hid; /* pointer to corresponding HID dev. */
+
+ struct spi_transfer input_transfer[2]; /* Transfer buffer for read and write. */
+ struct spi_message input_message; /* used to execute a sequence of spi transfers. */
+
+ struct spihid_ops *ops;
+ struct spi_hid_conf *conf;
+
+ struct spi_hid_device_descriptor desc; /* HID device descriptor. */
+ struct spi_hid_output_buf *output; /* Output buffer. */
+ struct spi_hid_input_buf *input; /* Input buffer. */
+ struct spi_hid_input_buf *response; /* Response buffer. */
+
+ u16 response_length;
+ u16 bufsize;
+
+ enum hidspi_power_state power_state;
+
+ u8 reset_attempts; /* The number of reset attempts. */
+
+ unsigned long flags; /* device flags. */
+
+ struct work_struct reset_work;
+
+ /* Control lock to ensure complete output transaction. */
+ struct mutex output_lock;
+ /* Power lock to make sure one power state change at a time. */
+ struct mutex power_lock;
+ /* I/O lock to prevent concurrent output writes during the input read. */
+ struct mutex io_lock;
+
+ struct completion output_done;
+
+ u8 read_approval_header[SPI_HID_READ_APPROVAL_LEN];
+ u8 read_approval_body[SPI_HID_READ_APPROVAL_LEN];
+
+ u32 report_descriptor_crc32; /* HID report descriptor crc32 checksum. */
+
+ u32 regulator_error_count;
+ int regulator_last_error;
+ u32 bus_error_count;
+ int bus_last_error;
+ u32 dir_count; /* device initiated reset count. */
+};
+
+#endif /* SPI_HID_CORE_H */
diff --git a/drivers/hid/spi-hid/spi-hid-trace.h b/drivers/hid/spi-hid/spi-hid-trace.h
new file mode 100644
index 000000000000..cc13d71a14de
--- /dev/null
+++ b/drivers/hid/spi-hid/spi-hid-trace.h
@@ -0,0 +1,169 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (c) 2021 Microsoft Corporation
+ */
+
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM spi_hid
+
+#if !defined(_SPI_HID_TRACE_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _SPI_HID_TRACE_H
+
+#include <linux/types.h>
+#include <linux/tracepoint.h>
+#include "spi-hid-core.h"
+
+DECLARE_EVENT_CLASS(spi_hid_transfer,
+ TP_PROTO(struct spi_hid *shid, const void *tx_buf, int tx_len,
+ const void *rx_buf, u16 rx_len, int ret),
+
+ TP_ARGS(shid, tx_buf, tx_len, rx_buf, rx_len, ret),
+
+ TP_STRUCT__entry(
+ __field(int, bus_num)
+ __field(int, chip_select)
+ __field(int, ret)
+ __dynamic_array(u8, rx_buf, rx_len)
+ __dynamic_array(u8, tx_buf, tx_len)
+ ),
+
+ TP_fast_assign(
+ __entry->bus_num = shid->spi->controller->bus_num;
+ __entry->chip_select = shid->spi->chip_select;
+ __entry->ret = ret;
+
+ memcpy(__get_dynamic_array(tx_buf), tx_buf, tx_len);
+ memcpy(__get_dynamic_array(rx_buf), rx_buf, rx_len);
+ ),
+
+ TP_printk("spi%d.%d: len=%d tx=[%*phD] rx=[%*phD] --> %d",
+ __entry->bus_num, __entry->chip_select,
+ __get_dynamic_array_len(tx_buf) + __get_dynamic_array_len(rx_buf),
+ __get_dynamic_array_len(tx_buf), __get_dynamic_array(tx_buf),
+ __get_dynamic_array_len(rx_buf), __get_dynamic_array(rx_buf),
+ __entry->ret)
+);
+
+DEFINE_EVENT(spi_hid_transfer, spi_hid_input_sync,
+ TP_PROTO(struct spi_hid *shid, const void *tx_buf, int tx_len,
+ const void *rx_buf, u16 rx_len, int ret),
+ TP_ARGS(shid, tx_buf, tx_len, rx_buf, rx_len, ret));
+
+DEFINE_EVENT(spi_hid_transfer, spi_hid_input_header_complete,
+ TP_PROTO(struct spi_hid *shid, const void *tx_buf, int tx_len,
+ const void *rx_buf, u16 rx_len, int ret),
+ TP_ARGS(shid, tx_buf, tx_len, rx_buf, rx_len, ret));
+
+DEFINE_EVENT(spi_hid_transfer, spi_hid_input_body_complete,
+ TP_PROTO(struct spi_hid *shid, const void *tx_buf, int tx_len,
+ const void *rx_buf, u16 rx_len, int ret),
+ TP_ARGS(shid, tx_buf, tx_len, rx_buf, rx_len, ret));
+
+DECLARE_EVENT_CLASS(spi_hid_irq,
+ TP_PROTO(struct spi_hid *shid, int irq),
+
+ TP_ARGS(shid, irq),
+
+ TP_STRUCT__entry(
+ __field(int, bus_num)
+ __field(int, chip_select)
+ __field(int, irq)
+ ),
+
+ TP_fast_assign(
+ __entry->bus_num = shid->spi->controller->bus_num;
+ __entry->chip_select = shid->spi->chip_select;
+ __entry->irq = irq;
+ ),
+
+ TP_printk("spi%d.%d: IRQ %d",
+ __entry->bus_num, __entry->chip_select, __entry->irq)
+);
+
+DEFINE_EVENT(spi_hid_irq, spi_hid_dev_irq,
+ TP_PROTO(struct spi_hid *shid, int irq), TP_ARGS(shid, irq));
+
+DECLARE_EVENT_CLASS(spi_hid,
+ TP_PROTO(struct spi_hid *shid),
+
+ TP_ARGS(shid),
+
+ TP_STRUCT__entry(
+ __field(int, bus_num)
+ __field(int, chip_select)
+ __field(int, power_state)
+ __field(u32, flags)
+
+ __field(int, vendor_id)
+ __field(int, product_id)
+ __field(int, max_input_length)
+ __field(int, max_output_length)
+ __field(u16, hid_version)
+ __field(u16, report_descriptor_length)
+ __field(u16, version_id)
+ ),
+
+ TP_fast_assign(
+ __entry->bus_num = shid->spi->controller->bus_num;
+ __entry->chip_select = shid->spi->chip_select;
+ __entry->power_state = shid->power_state;
+ __entry->flags = shid->flags;
+
+ __entry->vendor_id = shid->desc.vendor_id;
+ __entry->product_id = shid->desc.product_id;
+ __entry->max_input_length = shid->desc.max_input_length;
+ __entry->max_output_length = shid->desc.max_output_length;
+ __entry->hid_version = shid->desc.hid_version;
+ __entry->report_descriptor_length =
+ shid->desc.report_descriptor_length;
+ __entry->version_id = shid->desc.version_id;
+ ),
+
+ TP_printk("spi%d.%d: (%04x:%04x v%d) HID v%d.%d state p:%d len i:%d o:%d r:%d flags 0x%08x",
+ __entry->bus_num, __entry->chip_select,
+ __entry->vendor_id, __entry->product_id, __entry->version_id,
+ __entry->hid_version >> 8, __entry->hid_version & 0xff,
+ __entry->power_state, __entry->max_input_length,
+ __entry->max_output_length, __entry->report_descriptor_length,
+ __entry->flags)
+);
+
+DEFINE_EVENT(spi_hid, spi_hid_header_transfer, TP_PROTO(struct spi_hid *shid),
+ TP_ARGS(shid));
+
+DEFINE_EVENT(spi_hid, spi_hid_process_input_report,
+ TP_PROTO(struct spi_hid *shid), TP_ARGS(shid));
+
+DEFINE_EVENT(spi_hid, spi_hid_input_report_handler,
+ TP_PROTO(struct spi_hid *shid), TP_ARGS(shid));
+
+DEFINE_EVENT(spi_hid, spi_hid_reset_response, TP_PROTO(struct spi_hid *shid),
+ TP_ARGS(shid));
+
+DEFINE_EVENT(spi_hid, spi_hid_create_device, TP_PROTO(struct spi_hid *shid),
+ TP_ARGS(shid));
+
+DEFINE_EVENT(spi_hid, spi_hid_refresh_device, TP_PROTO(struct spi_hid *shid),
+ TP_ARGS(shid));
+
+DEFINE_EVENT(spi_hid, spi_hid_response_handler, TP_PROTO(struct spi_hid *shid),
+ TP_ARGS(shid));
+
+DEFINE_EVENT(spi_hid, spi_hid_error_handler, TP_PROTO(struct spi_hid *shid),
+ TP_ARGS(shid));
+
+#endif /* _SPI_HID_TRACE_H */
+
+/*
+ * The following must be outside the protection of the above #if block.
+ */
+#undef TRACE_INCLUDE_PATH
+#undef TRACE_INCLUDE_FILE
+#define TRACE_INCLUDE_PATH .
+
+/*
+ * It is required that the TRACE_INCLUDE_FILE be the same
+ * as this file without the ".h".
+ */
+#define TRACE_INCLUDE_FILE spi-hid-trace
+#include <trace/define_trace.h>
--
2.53.0.1185.g05d4b7b318-goog
^ permalink raw reply related
* [PATCH v3 07/11] HID: spi_hid: add ACPI support for SPI over HID
From: Jingyuan Liang @ 2026-04-02 1:59 UTC (permalink / raw)
To: Jiri Kosina, Benjamin Tissoires, Jonathan Corbet, Mark Brown,
Steven Rostedt, Masami Hiramatsu, Mathieu Desnoyers,
Dmitry Torokhov, Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: linux-input, linux-doc, linux-kernel, linux-spi,
linux-trace-kernel, devicetree, hbarnor, tfiga, Jingyuan Liang,
Angela Czubak
In-Reply-To: <20260402-send-upstream-v3-0-6091c458d357@chromium.org>
From: Angela Czubak <acz@semihalf.com>
Detect SPI HID devices described in ACPI.
Signed-off-by: Angela Czubak <acz@semihalf.com>
Reviewed-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Jingyuan Liang <jingyliang@chromium.org>
---
drivers/hid/spi-hid/Kconfig | 15 +++
drivers/hid/spi-hid/Makefile | 1 +
drivers/hid/spi-hid/spi-hid-acpi.c | 253 +++++++++++++++++++++++++++++++++++++
drivers/hid/spi-hid/spi-hid-core.c | 26 +---
drivers/hid/spi-hid/spi-hid.h | 45 +++++++
5 files changed, 315 insertions(+), 25 deletions(-)
diff --git a/drivers/hid/spi-hid/Kconfig b/drivers/hid/spi-hid/Kconfig
index 836fdefe8345..114b1e00da39 100644
--- a/drivers/hid/spi-hid/Kconfig
+++ b/drivers/hid/spi-hid/Kconfig
@@ -10,6 +10,21 @@ menuconfig SPI_HID
if SPI_HID
+config SPI_HID_ACPI
+ tristate "HID over SPI transport layer ACPI driver"
+ depends on ACPI
+ select SPI_HID_CORE
+ help
+ Say Y here if you use a keyboard, a touchpad, a touchscreen, or any
+ other HID based devices which are connected to your computer via SPI.
+ This driver supports ACPI-based systems.
+
+ If unsure, say N.
+
+ This support is also available as a module. If so, the module
+ will be called spi-hid-acpi. It will also build/depend on the
+ module spi-hid.
+
config SPI_HID_CORE
tristate
endif
diff --git a/drivers/hid/spi-hid/Makefile b/drivers/hid/spi-hid/Makefile
index 733e006df56e..3ca326602643 100644
--- a/drivers/hid/spi-hid/Makefile
+++ b/drivers/hid/spi-hid/Makefile
@@ -8,3 +8,4 @@
obj-$(CONFIG_SPI_HID_CORE) += spi-hid.o
spi-hid-objs = spi-hid-core.o
CFLAGS_spi-hid-core.o := -I$(src)
+obj-$(CONFIG_SPI_HID_ACPI) += spi-hid-acpi.o
diff --git a/drivers/hid/spi-hid/spi-hid-acpi.c b/drivers/hid/spi-hid/spi-hid-acpi.c
new file mode 100644
index 000000000000..298e3ba44d8a
--- /dev/null
+++ b/drivers/hid/spi-hid/spi-hid-acpi.c
@@ -0,0 +1,253 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * HID over SPI protocol, ACPI related code
+ *
+ * Copyright (c) 2021 Microsoft Corporation
+ * Copyright (c) 2026 Google LLC
+ *
+ * This code was forked out of the HID over SPI core code, which is partially
+ * based on "HID over I2C protocol implementation:
+ *
+ * Copyright (c) 2012 Benjamin Tissoires <benjamin.tissoires@gmail.com>
+ * Copyright (c) 2012 Ecole Nationale de l'Aviation Civile, France
+ * Copyright (c) 2012 Red Hat, Inc
+ *
+ * which in turn is partially based on "USB HID support for Linux":
+ *
+ * Copyright (c) 1999 Andreas Gal
+ * Copyright (c) 2000-2005 Vojtech Pavlik <vojtech@suse.cz>
+ * Copyright (c) 2005 Michael Haboustak <mike-@cinci.rr.com> for Concept2, Inc
+ * Copyright (c) 2007-2008 Oliver Neukum
+ * Copyright (c) 2006-2010 Jiri Kosina
+ */
+
+#include <linux/acpi.h>
+#include <linux/delay.h>
+#include <linux/device.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/reset.h>
+#include <linux/uuid.h>
+
+#include "spi-hid.h"
+
+/* Config structure is filled with data from ACPI */
+struct spi_hid_acpi_config {
+ struct spihid_ops ops;
+
+ struct spi_hid_conf property_conf;
+ u32 post_power_on_delay_ms;
+ u32 minimal_reset_delay_ms;
+ struct acpi_device *adev;
+};
+
+/* HID SPI Device: 6e2ac436-0fcf41af-a265-b32a220dcfab */
+static guid_t spi_hid_guid =
+ GUID_INIT(0x6E2AC436, 0x0FCF, 0x41AF,
+ 0xA2, 0x65, 0xB3, 0x2A, 0x22, 0x0D, 0xCF, 0xAB);
+
+static int spi_hid_acpi_populate_config(struct spi_hid_acpi_config *conf,
+ struct acpi_device *adev)
+{
+ acpi_handle handle = acpi_device_handle(adev);
+ union acpi_object *obj;
+
+ conf->adev = adev;
+
+ /* Revision 3 for HID over SPI V1, see specification. */
+ obj = acpi_evaluate_dsm_typed(handle, &spi_hid_guid, 3, 1, NULL,
+ ACPI_TYPE_INTEGER);
+ if (!obj) {
+ acpi_handle_err(handle,
+ "Error _DSM call to get HID input report header address failed");
+ return -ENODEV;
+ }
+ conf->property_conf.input_report_header_address = obj->integer.value;
+ ACPI_FREE(obj);
+
+ obj = acpi_evaluate_dsm_typed(handle, &spi_hid_guid, 3, 2, NULL,
+ ACPI_TYPE_INTEGER);
+ if (!obj) {
+ acpi_handle_err(handle,
+ "Error _DSM call to get HID input report body address failed");
+ return -ENODEV;
+ }
+ conf->property_conf.input_report_body_address = obj->integer.value;
+ ACPI_FREE(obj);
+
+ obj = acpi_evaluate_dsm_typed(handle, &spi_hid_guid, 3, 3, NULL,
+ ACPI_TYPE_INTEGER);
+ if (!obj) {
+ acpi_handle_err(handle,
+ "Error _DSM call to get HID output report header address failed");
+ return -ENODEV;
+ }
+ conf->property_conf.output_report_address = obj->integer.value;
+ ACPI_FREE(obj);
+
+ obj = acpi_evaluate_dsm_typed(handle, &spi_hid_guid, 3, 4, NULL,
+ ACPI_TYPE_BUFFER);
+ if (!obj) {
+ acpi_handle_err(handle,
+ "Error _DSM call to get HID read opcode failed");
+ return -ENODEV;
+ }
+ if (obj->buffer.length == 1) {
+ conf->property_conf.read_opcode = obj->buffer.pointer[0];
+ } else {
+ acpi_handle_err(handle,
+ "Error _DSM call to get HID read opcode, too long buffer");
+ ACPI_FREE(obj);
+ return -ENODEV;
+ }
+ ACPI_FREE(obj);
+
+ obj = acpi_evaluate_dsm_typed(handle, &spi_hid_guid, 3, 5, NULL,
+ ACPI_TYPE_BUFFER);
+ if (!obj) {
+ acpi_handle_err(handle,
+ "Error _DSM call to get HID write opcode failed");
+ return -ENODEV;
+ }
+ if (obj->buffer.length == 1) {
+ conf->property_conf.write_opcode = obj->buffer.pointer[0];
+ } else {
+ acpi_handle_err(handle,
+ "Error _DSM call to get HID write opcode, too long buffer");
+ ACPI_FREE(obj);
+ return -ENODEV;
+ }
+ ACPI_FREE(obj);
+
+ /* Value not provided in ACPI,*/
+ conf->post_power_on_delay_ms = 5;
+ conf->minimal_reset_delay_ms = 150;
+
+ if (!acpi_has_method(handle, "_RST")) {
+ acpi_handle_err(handle, "No reset method for acpi handle");
+ return -EINVAL;
+ }
+
+ /* FIXME: not reading hid-over-spi-flags, multi-SPI not supported */
+
+ return 0;
+}
+
+static int spi_hid_acpi_power_none(struct spihid_ops *ops)
+{
+ return 0;
+}
+
+static int spi_hid_acpi_power_down(struct spihid_ops *ops)
+{
+ struct spi_hid_acpi_config *conf = container_of(ops,
+ struct spi_hid_acpi_config,
+ ops);
+
+ return acpi_device_set_power(conf->adev, ACPI_STATE_D3);
+}
+
+static int spi_hid_acpi_power_up(struct spihid_ops *ops)
+{
+ struct spi_hid_acpi_config *conf = container_of(ops,
+ struct spi_hid_acpi_config,
+ ops);
+ int error;
+
+ error = acpi_device_set_power(conf->adev, ACPI_STATE_D0);
+ if (error) {
+ dev_err(&conf->adev->dev, "Error could not power up ACPI device: %d.", error);
+ return error;
+ }
+
+ if (conf->post_power_on_delay_ms)
+ msleep(conf->post_power_on_delay_ms);
+
+ return 0;
+}
+
+static int spi_hid_acpi_assert_reset(struct spihid_ops *ops)
+{
+ return 0;
+}
+
+static int spi_hid_acpi_deassert_reset(struct spihid_ops *ops)
+{
+ struct spi_hid_acpi_config *conf = container_of(ops,
+ struct spi_hid_acpi_config,
+ ops);
+
+ return device_reset(&conf->adev->dev);
+}
+
+static void spi_hid_acpi_sleep_minimal_reset_delay(struct spihid_ops *ops)
+{
+ struct spi_hid_acpi_config *conf = container_of(ops,
+ struct spi_hid_acpi_config,
+ ops);
+ fsleep(1000 * conf->minimal_reset_delay_ms);
+}
+
+static int spi_hid_acpi_probe(struct spi_device *spi)
+{
+ struct device *dev = &spi->dev;
+ struct acpi_device *adev;
+ struct spi_hid_acpi_config *config;
+ int error;
+
+ adev = ACPI_COMPANION(dev);
+ if (!adev) {
+ dev_err(dev, "Error could not get ACPI device.");
+ return -ENODEV;
+ }
+
+ config = devm_kzalloc(dev, sizeof(struct spi_hid_acpi_config),
+ GFP_KERNEL);
+ if (!config)
+ return -ENOMEM;
+
+ if (acpi_device_power_manageable(adev)) {
+ config->ops.power_up = spi_hid_acpi_power_up;
+ config->ops.power_down = spi_hid_acpi_power_down;
+ } else {
+ config->ops.power_up = spi_hid_acpi_power_none;
+ config->ops.power_down = spi_hid_acpi_power_none;
+ }
+ config->ops.assert_reset = spi_hid_acpi_assert_reset;
+ config->ops.deassert_reset = spi_hid_acpi_deassert_reset;
+ config->ops.sleep_minimal_reset_delay =
+ spi_hid_acpi_sleep_minimal_reset_delay;
+
+ error = spi_hid_acpi_populate_config(config, adev);
+ if (error) {
+ dev_err(dev, "%s: unable to populate config data.", __func__);
+ return error;
+ }
+
+ return spi_hid_core_probe(spi, &config->ops, &config->property_conf);
+}
+
+static const struct acpi_device_id spi_hid_acpi_match[] = {
+ { "ACPI0C51", 0 },
+ { "PNP0C51", 0 },
+ { }
+};
+MODULE_DEVICE_TABLE(acpi, spi_hid_acpi_match);
+
+static struct spi_driver spi_hid_acpi_driver = {
+ .driver = {
+ .name = "spi_hid_acpi",
+ .owner = THIS_MODULE,
+ .acpi_match_table = spi_hid_acpi_match,
+ .probe_type = PROBE_PREFER_ASYNCHRONOUS,
+ .dev_groups = spi_hid_groups,
+ },
+ .probe = spi_hid_acpi_probe,
+ .remove = spi_hid_core_remove,
+};
+
+module_spi_driver(spi_hid_acpi_driver);
+
+MODULE_DESCRIPTION("HID over SPI ACPI transport driver");
+MODULE_AUTHOR("Angela Czubak <aczubak@google.com>");
+MODULE_LICENSE("GPL");
diff --git a/drivers/hid/spi-hid/spi-hid-core.c b/drivers/hid/spi-hid/spi-hid-core.c
index 802615565541..d48175c764b9 100644
--- a/drivers/hid/spi-hid/spi-hid-core.c
+++ b/drivers/hid/spi-hid/spi-hid-core.c
@@ -43,6 +43,7 @@
#include <linux/wait.h>
#include <linux/workqueue.h>
+#include "spi-hid.h"
#include "spi-hid-core.h"
#define CREATE_TRACE_POINTS
@@ -110,31 +111,6 @@ struct spi_hid_output_report {
u8 *content;
};
-/* struct spi_hid_conf - Conf provided to the core */
-struct spi_hid_conf {
- u32 input_report_header_address;
- u32 input_report_body_address;
- u32 output_report_address;
- u8 read_opcode;
- u8 write_opcode;
-};
-
-/**
- * struct spihid_ops - Ops provided to the core
- * @power_up: do sequencing to power up the device
- * @power_down: do sequencing to power down the device
- * @assert_reset: do sequencing to assert the reset line
- * @deassert_reset: do sequencing to deassert the reset line
- * @sleep_minimal_reset_delay: minimal sleep delay during reset
- */
-struct spihid_ops {
- int (*power_up)(struct spihid_ops *ops);
- int (*power_down)(struct spihid_ops *ops);
- int (*assert_reset)(struct spihid_ops *ops);
- int (*deassert_reset)(struct spihid_ops *ops);
- void (*sleep_minimal_reset_delay)(struct spihid_ops *ops);
-};
-
static struct hid_ll_driver spi_hid_ll_driver;
static void spi_hid_populate_read_approvals(const struct spi_hid_conf *conf,
diff --git a/drivers/hid/spi-hid/spi-hid.h b/drivers/hid/spi-hid/spi-hid.h
new file mode 100644
index 000000000000..f5a5f4d54beb
--- /dev/null
+++ b/drivers/hid/spi-hid/spi-hid.h
@@ -0,0 +1,45 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (c) 2021 Microsoft Corporation
+ * Copyright (c) 2026 Google LLC
+ */
+
+#ifndef SPI_HID_H
+#define SPI_HID_H
+
+#include <linux/spi/spi.h>
+#include <linux/sysfs.h>
+
+/* struct spi_hid_conf - Conf provided to the core */
+struct spi_hid_conf {
+ u32 input_report_header_address;
+ u32 input_report_body_address;
+ u32 output_report_address;
+ u8 read_opcode;
+ u8 write_opcode;
+};
+
+/**
+ * struct spihid_ops - Ops provided to the core
+ * @power_up: do sequencing to power up the device
+ * @power_down: do sequencing to power down the device
+ * @assert_reset: do sequencing to assert the reset line
+ * @deassert_reset: do sequencing to deassert the reset line
+ * @sleep_minimal_reset_delay: minimal sleep delay during reset
+ */
+struct spihid_ops {
+ int (*power_up)(struct spihid_ops *ops);
+ int (*power_down)(struct spihid_ops *ops);
+ int (*assert_reset)(struct spihid_ops *ops);
+ int (*deassert_reset)(struct spihid_ops *ops);
+ void (*sleep_minimal_reset_delay)(struct spihid_ops *ops);
+};
+
+int spi_hid_core_probe(struct spi_device *spi, struct spihid_ops *ops,
+ struct spi_hid_conf *conf);
+
+void spi_hid_core_remove(struct spi_device *spi);
+
+extern const struct attribute_group *spi_hid_groups[];
+
+#endif /* SPI_HID_H */
--
2.53.0.1185.g05d4b7b318-goog
^ permalink raw reply related
* [PATCH v3 08/11] HID: spi_hid: add device tree support for SPI over HID
From: Jingyuan Liang @ 2026-04-02 1:59 UTC (permalink / raw)
To: Jiri Kosina, Benjamin Tissoires, Jonathan Corbet, Mark Brown,
Steven Rostedt, Masami Hiramatsu, Mathieu Desnoyers,
Dmitry Torokhov, Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: linux-input, linux-doc, linux-kernel, linux-spi,
linux-trace-kernel, devicetree, hbarnor, tfiga, Jingyuan Liang,
Jarrett Schultz, Dmitry Antipov
In-Reply-To: <20260402-send-upstream-v3-0-6091c458d357@chromium.org>
From: Jarrett Schultz <jaschultz@microsoft.com>
Signed-off-by: Dmitry Antipov <dmanti@microsoft.com>
Signed-off-by: Jingyuan Liang <jingyliang@chromium.org>
---
drivers/hid/spi-hid/Kconfig | 15 +++
drivers/hid/spi-hid/Makefile | 1 +
drivers/hid/spi-hid/spi-hid-of.c | 243 +++++++++++++++++++++++++++++++++++++++
3 files changed, 259 insertions(+)
diff --git a/drivers/hid/spi-hid/Kconfig b/drivers/hid/spi-hid/Kconfig
index 114b1e00da39..76a2cd587a3e 100644
--- a/drivers/hid/spi-hid/Kconfig
+++ b/drivers/hid/spi-hid/Kconfig
@@ -25,6 +25,21 @@ config SPI_HID_ACPI
will be called spi-hid-acpi. It will also build/depend on the
module spi-hid.
+config SPI_HID_OF
+ tristate "HID over SPI transport layer Open Firmware driver"
+ depends on OF
+ select SPI_HID_CORE
+ help
+ Say Y here if you use a keyboard, a touchpad, a touchscreen, or any
+ other HID based devices which are connected to your computer via SPI.
+ This driver supports Open Firmware (Device Tree)-based systems.
+
+ If unsure, say N.
+
+ This support is also available as a module. If so, the module
+ will be called spi-hid-of. It will also build/depend on the
+ module spi-hid.
+
config SPI_HID_CORE
tristate
endif
diff --git a/drivers/hid/spi-hid/Makefile b/drivers/hid/spi-hid/Makefile
index 3ca326602643..31192e71edae 100644
--- a/drivers/hid/spi-hid/Makefile
+++ b/drivers/hid/spi-hid/Makefile
@@ -9,3 +9,4 @@ obj-$(CONFIG_SPI_HID_CORE) += spi-hid.o
spi-hid-objs = spi-hid-core.o
CFLAGS_spi-hid-core.o := -I$(src)
obj-$(CONFIG_SPI_HID_ACPI) += spi-hid-acpi.o
+obj-$(CONFIG_SPI_HID_OF) += spi-hid-of.o
diff --git a/drivers/hid/spi-hid/spi-hid-of.c b/drivers/hid/spi-hid/spi-hid-of.c
new file mode 100644
index 000000000000..651456b6906d
--- /dev/null
+++ b/drivers/hid/spi-hid/spi-hid-of.c
@@ -0,0 +1,243 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * HID over SPI protocol, Open Firmware related code
+ *
+ * Copyright (c) 2021 Microsoft Corporation
+ *
+ * This code was forked out of the HID over SPI core code, which is partially
+ * based on "HID over I2C protocol implementation:
+ *
+ * Copyright (c) 2012 Benjamin Tissoires <benjamin.tissoires@gmail.com>
+ * Copyright (c) 2012 Ecole Nationale de l'Aviation Civile, France
+ * Copyright (c) 2012 Red Hat, Inc
+ *
+ * which in turn is partially based on "USB HID support for Linux":
+ *
+ * Copyright (c) 1999 Andreas Gal
+ * Copyright (c) 2000-2005 Vojtech Pavlik <vojtech@suse.cz>
+ * Copyright (c) 2005 Michael Haboustak <mike-@cinci.rr.com> for Concept2, Inc
+ * Copyright (c) 2007-2008 Oliver Neukum
+ * Copyright (c) 2006-2010 Jiri Kosina
+ */
+
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/regulator/consumer.h>
+#include <linux/gpio/consumer.h>
+#include <linux/delay.h>
+
+#include "spi-hid.h"
+
+struct spi_hid_timing_data {
+ u32 post_power_on_delay_ms;
+ u32 minimal_reset_delay_ms;
+};
+
+/* Config structure is filled with data from Device Tree */
+struct spi_hid_of_config {
+ struct spihid_ops ops;
+
+ struct spi_hid_conf property_conf;
+ const struct spi_hid_timing_data *timing_data;
+
+ struct gpio_desc *reset_gpio;
+ struct regulator *supply;
+ bool supply_enabled;
+ u16 hid_over_spi_flags;
+};
+
+static int spi_hid_of_populate_config(struct spi_hid_of_config *conf,
+ struct device *dev)
+{
+ int error;
+ u32 val;
+
+ error = device_property_read_u32(dev, "input-report-header-address",
+ &val);
+ if (error) {
+ dev_err(dev, "Input report header address not provided.");
+ return -ENODEV;
+ }
+ conf->property_conf.input_report_header_address = val;
+
+ error = device_property_read_u32(dev, "input-report-body-address", &val);
+ if (error) {
+ dev_err(dev, "Input report body address not provided.");
+ return -ENODEV;
+ }
+ conf->property_conf.input_report_body_address = val;
+
+ error = device_property_read_u32(dev, "output-report-address", &val);
+ if (error) {
+ dev_err(dev, "Output report address not provided.");
+ return -ENODEV;
+ }
+ conf->property_conf.output_report_address = val;
+
+ error = device_property_read_u32(dev, "read-opcode", &val);
+ if (error) {
+ dev_err(dev, "Read opcode not provided.");
+ return -ENODEV;
+ }
+ conf->property_conf.read_opcode = val;
+
+ error = device_property_read_u32(dev, "write-opcode", &val);
+ if (error) {
+ dev_err(dev, "Write opcode not provided.");
+ return -ENODEV;
+ }
+ conf->property_conf.write_opcode = val;
+
+ conf->supply = devm_regulator_get(dev, "vdd");
+ if (IS_ERR(conf->supply)) {
+ if (PTR_ERR(conf->supply) != -EPROBE_DEFER)
+ dev_err(dev, "Failed to get regulator: %ld.",
+ PTR_ERR(conf->supply));
+ return PTR_ERR(conf->supply);
+ }
+ conf->supply_enabled = false;
+
+ conf->reset_gpio = devm_gpiod_get(dev, "reset", GPIOD_OUT_LOW);
+ if (IS_ERR(conf->reset_gpio)) {
+ dev_err(dev, "%s: error getting reset GPIO.", __func__);
+ return PTR_ERR(conf->reset_gpio);
+ }
+
+ return 0;
+}
+
+static int spi_hid_of_power_down(struct spihid_ops *ops)
+{
+ struct spi_hid_of_config *conf = container_of(ops,
+ struct spi_hid_of_config,
+ ops);
+ int error;
+
+ if (!conf->supply_enabled)
+ return 0;
+
+ error = regulator_disable(conf->supply);
+ if (error == 0)
+ conf->supply_enabled = false;
+
+ return error;
+}
+
+static int spi_hid_of_power_up(struct spihid_ops *ops)
+{
+ struct spi_hid_of_config *conf = container_of(ops,
+ struct spi_hid_of_config,
+ ops);
+ int error;
+
+ if (conf->supply_enabled)
+ return 0;
+
+ error = regulator_enable(conf->supply);
+
+ if (error == 0) {
+ conf->supply_enabled = true;
+ fsleep(1000 * conf->timing_data->post_power_on_delay_ms);
+ }
+
+ return error;
+}
+
+static int spi_hid_of_assert_reset(struct spihid_ops *ops)
+{
+ struct spi_hid_of_config *conf = container_of(ops,
+ struct spi_hid_of_config,
+ ops);
+
+ gpiod_set_value(conf->reset_gpio, 1);
+ return 0;
+}
+
+static int spi_hid_of_deassert_reset(struct spihid_ops *ops)
+{
+ struct spi_hid_of_config *conf = container_of(ops,
+ struct spi_hid_of_config,
+ ops);
+
+ gpiod_set_value(conf->reset_gpio, 0);
+ return 0;
+}
+
+static void spi_hid_of_sleep_minimal_reset_delay(struct spihid_ops *ops)
+{
+ struct spi_hid_of_config *conf = container_of(ops,
+ struct spi_hid_of_config,
+ ops);
+ fsleep(1000 * conf->timing_data->minimal_reset_delay_ms);
+}
+
+static int spi_hid_of_probe(struct spi_device *spi)
+{
+ struct device *dev = &spi->dev;
+ struct spi_hid_of_config *config;
+ int error;
+
+ config = devm_kzalloc(dev, sizeof(struct spi_hid_of_config),
+ GFP_KERNEL);
+ if (!config)
+ return -ENOMEM;
+
+ config->ops.power_up = spi_hid_of_power_up;
+ config->ops.power_down = spi_hid_of_power_down;
+ config->ops.assert_reset = spi_hid_of_assert_reset;
+ config->ops.deassert_reset = spi_hid_of_deassert_reset;
+ config->ops.sleep_minimal_reset_delay =
+ spi_hid_of_sleep_minimal_reset_delay;
+
+ config->timing_data = device_get_match_data(dev);
+ /*
+ * FIXME: hid_over_spi_flags could be retrieved from spi mode.
+ * It is always 0 because multi-SPI not supported.
+ */
+ config->hid_over_spi_flags = 0;
+
+ error = spi_hid_of_populate_config(config, dev);
+ if (error) {
+ dev_err(dev, "%s: unable to populate config data.", __func__);
+ return error;
+ }
+
+ return spi_hid_core_probe(spi, &config->ops, &config->property_conf);
+}
+
+const struct spi_hid_timing_data timing_data = {
+ .post_power_on_delay_ms = 10,
+ .minimal_reset_delay_ms = 100,
+};
+
+const struct of_device_id spi_hid_of_match[] = {
+ { .compatible = "hid-over-spi", .data = &timing_data },
+ {}
+};
+MODULE_DEVICE_TABLE(of, spi_hid_of_match);
+
+static const struct spi_device_id spi_hid_of_id_table[] = {
+ { "hid", 0 },
+ { "hid-over-spi", 0 },
+ { }
+};
+MODULE_DEVICE_TABLE(spi, spi_hid_of_id_table);
+
+static struct spi_driver spi_hid_of_driver = {
+ .driver = {
+ .name = "spi_hid_of",
+ .owner = THIS_MODULE,
+ .of_match_table = spi_hid_of_match,
+ .probe_type = PROBE_PREFER_ASYNCHRONOUS,
+ .dev_groups = spi_hid_groups,
+ },
+ .probe = spi_hid_of_probe,
+ .remove = spi_hid_core_remove,
+ .id_table = spi_hid_of_id_table,
+};
+
+module_spi_driver(spi_hid_of_driver);
+
+MODULE_DESCRIPTION("HID over SPI OF transport driver");
+MODULE_AUTHOR("Dmitry Antipov <dmanti@microsoft.com>");
+MODULE_LICENSE("GPL");
--
2.53.0.1185.g05d4b7b318-goog
^ permalink raw reply related
* [PATCH v3 09/11] dt-bindings: input: Document hid-over-spi DT schema
From: Jingyuan Liang @ 2026-04-02 1:59 UTC (permalink / raw)
To: Jiri Kosina, Benjamin Tissoires, Jonathan Corbet, Mark Brown,
Steven Rostedt, Masami Hiramatsu, Mathieu Desnoyers,
Dmitry Torokhov, Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: linux-input, linux-doc, linux-kernel, linux-spi,
linux-trace-kernel, devicetree, hbarnor, tfiga, Jingyuan Liang,
Dmitry Antipov, Jarrett Schultz
In-Reply-To: <20260402-send-upstream-v3-0-6091c458d357@chromium.org>
Documentation describes the required and optional properties for
implementing Device Tree for a Microsoft G6 Touch Digitizer that
supports HID over SPI Protocol 1.0 specification.
The properties are common to HID over SPI.
Signed-off-by: Dmitry Antipov <dmanti@microsoft.com>
Signed-off-by: Jarrett Schultz <jaschultz@microsoft.com>
Signed-off-by: Jingyuan Liang <jingyliang@chromium.org>
---
.../devicetree/bindings/input/hid-over-spi.yaml | 126 +++++++++++++++++++++
1 file changed, 126 insertions(+)
diff --git a/Documentation/devicetree/bindings/input/hid-over-spi.yaml b/Documentation/devicetree/bindings/input/hid-over-spi.yaml
new file mode 100644
index 000000000000..d1b0a2e26c32
--- /dev/null
+++ b/Documentation/devicetree/bindings/input/hid-over-spi.yaml
@@ -0,0 +1,126 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/input/hid-over-spi.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: HID over SPI Devices
+
+maintainers:
+ - Benjamin Tissoires <benjamin.tissoires@redhat.com>
+ - Jiri Kosina <jkosina@suse.cz>
+
+description: |+
+ HID over SPI provides support for various Human Interface Devices over the
+ SPI bus. These devices can be for example touchpads, keyboards, touch screens
+ or sensors.
+
+ The specification has been written by Microsoft and is currently available
+ here: https://www.microsoft.com/en-us/download/details.aspx?id=103325
+
+ If this binding is used, the kernel module spi-hid will handle the
+ communication with the device and the generic hid core layer will handle the
+ protocol.
+
+allOf:
+ - $ref: /schemas/input/touchscreen/touchscreen.yaml#
+
+properties:
+ compatible:
+ oneOf:
+ - items:
+ - enum:
+ - microsoft,g6-touch-digitizer
+ - const: hid-over-spi
+ - description: Just "hid-over-spi" alone is allowed, but not recommended.
+ const: hid-over-spi
+
+ reg:
+ maxItems: 1
+
+ interrupts:
+ maxItems: 1
+
+ reset-gpios:
+ maxItems: 1
+ description:
+ GPIO specifier for the digitizer's reset pin (active low). The line must
+ be flagged with GPIO_ACTIVE_LOW.
+
+ vdd-supply:
+ description:
+ Regulator for the VDD supply voltage.
+
+ input-report-header-address:
+ $ref: /schemas/types.yaml#/definitions/uint32
+ minimum: 0
+ maximum: 0xffffff
+ description:
+ A value to be included in the Read Approval packet, listing an address of
+ the input report header to be put on the SPI bus. This address has 24
+ bits.
+
+ input-report-body-address:
+ $ref: /schemas/types.yaml#/definitions/uint32
+ minimum: 0
+ maximum: 0xffffff
+ description:
+ A value to be included in the Read Approval packet, listing an address of
+ the input report body to be put on the SPI bus. This address has 24 bits.
+
+ output-report-address:
+ $ref: /schemas/types.yaml#/definitions/uint32
+ minimum: 0
+ maximum: 0xffffff
+ description:
+ A value to be included in the Output Report sent by the host, listing an
+ address where the output report on the SPI bus is to be written to. This
+ address has 24 bits.
+
+ read-opcode:
+ $ref: /schemas/types.yaml#/definitions/uint8
+ description:
+ Value to be used in Read Approval packets. 1 byte.
+
+ write-opcode:
+ $ref: /schemas/types.yaml#/definitions/uint8
+ description:
+ Value to be used in Write Approval packets. 1 byte.
+
+required:
+ - compatible
+ - interrupts
+ - reset-gpios
+ - vdd-supply
+ - input-report-header-address
+ - input-report-body-address
+ - output-report-address
+ - read-opcode
+ - write-opcode
+
+additionalProperties: false
+
+examples:
+ - |
+ #include <dt-bindings/interrupt-controller/irq.h>
+ #include <dt-bindings/gpio/gpio.h>
+
+ spi {
+ #address-cells = <1>;
+ #size-cells = <0>;
+
+ hid@0 {
+ compatible = "microsoft,g6-touch-digitizer", "hid-over-spi";
+ reg = <0x0>;
+ interrupts-extended = <&gpio 42 IRQ_TYPE_EDGE_FALLING>;
+ reset-gpios = <&gpio 27 GPIO_ACTIVE_LOW>;
+ vdd-supply = <&pm8350c_l3>;
+ pinctrl-names = "default";
+ pinctrl-0 = <&ts_d6_int_bias>;
+ input-report-header-address = <0x1000>;
+ input-report-body-address = <0x1004>;
+ output-report-address = <0x2000>;
+ read-opcode = /bits/ 8 <0x0b>;
+ write-opcode = /bits/ 8 <0x02>;
+ };
+ };
--
2.53.0.1185.g05d4b7b318-goog
^ permalink raw reply related
* [PATCH v3 10/11] HID: spi-hid: add power management implementation
From: Jingyuan Liang @ 2026-04-02 1:59 UTC (permalink / raw)
To: Jiri Kosina, Benjamin Tissoires, Jonathan Corbet, Mark Brown,
Steven Rostedt, Masami Hiramatsu, Mathieu Desnoyers,
Dmitry Torokhov, Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: linux-input, linux-doc, linux-kernel, linux-spi,
linux-trace-kernel, devicetree, hbarnor, tfiga, Jingyuan Liang
In-Reply-To: <20260402-send-upstream-v3-0-6091c458d357@chromium.org>
Implement HID over SPI driver power management callbacks.
Signed-off-by: Jingyuan Liang <jingyliang@chromium.org>
---
drivers/hid/spi-hid/spi-hid-acpi.c | 1 +
drivers/hid/spi-hid/spi-hid-core.c | 107 +++++++++++++++++++++++++++++++++++++
drivers/hid/spi-hid/spi-hid-of.c | 1 +
drivers/hid/spi-hid/spi-hid.h | 1 +
4 files changed, 110 insertions(+)
diff --git a/drivers/hid/spi-hid/spi-hid-acpi.c b/drivers/hid/spi-hid/spi-hid-acpi.c
index 298e3ba44d8a..15cfc4e6cc2f 100644
--- a/drivers/hid/spi-hid/spi-hid-acpi.c
+++ b/drivers/hid/spi-hid/spi-hid-acpi.c
@@ -238,6 +238,7 @@ static struct spi_driver spi_hid_acpi_driver = {
.driver = {
.name = "spi_hid_acpi",
.owner = THIS_MODULE,
+ .pm = &spi_hid_core_pm,
.acpi_match_table = spi_hid_acpi_match,
.probe_type = PROBE_PREFER_ASYNCHRONOUS,
.dev_groups = spi_hid_groups,
diff --git a/drivers/hid/spi-hid/spi-hid-core.c b/drivers/hid/spi-hid/spi-hid-core.c
index d48175c764b9..5f7a5bb692d9 100644
--- a/drivers/hid/spi-hid/spi-hid-core.c
+++ b/drivers/hid/spi-hid/spi-hid-core.c
@@ -35,6 +35,8 @@
#include <linux/list.h>
#include <linux/module.h>
#include <linux/mutex.h>
+#include <linux/pm.h>
+#include <linux/pm_wakeirq.h>
#include <linux/slab.h>
#include <linux/spi/spi.h>
#include <linux/string.h>
@@ -244,6 +246,81 @@ static const char *spi_hid_power_mode_string(enum hidspi_power_state power_state
}
}
+static void spi_hid_suspend(struct spi_hid *shid)
+{
+ int error;
+ struct device *dev = &shid->spi->dev;
+
+ guard(mutex)(&shid->power_lock);
+ if (shid->power_state == HIDSPI_OFF)
+ return;
+
+ if (shid->hid) {
+ error = hid_driver_suspend(shid->hid, PMSG_SUSPEND);
+ if (error) {
+ dev_err(dev, "%s failed to suspend hid driver: %d",
+ __func__, error);
+ return;
+ }
+ }
+
+ disable_irq(shid->spi->irq);
+
+ clear_bit(SPI_HID_READY, &shid->flags);
+
+ if (!device_may_wakeup(dev)) {
+ set_bit(SPI_HID_RESET_PENDING, &shid->flags);
+
+ shid->ops->assert_reset(shid->ops);
+
+ error = shid->ops->power_down(shid->ops);
+ if (error) {
+ dev_err(dev, "%s: could not power down.", __func__);
+ shid->regulator_error_count++;
+ shid->regulator_last_error = error;
+ return;
+ }
+
+ shid->power_state = HIDSPI_OFF;
+ }
+}
+
+static void spi_hid_resume(struct spi_hid *shid)
+{
+ int error;
+ struct device *dev = &shid->spi->dev;
+
+ guard(mutex)(&shid->power_lock);
+ if (shid->power_state == HIDSPI_ON)
+ return;
+
+ enable_irq(shid->spi->irq);
+
+ if (!device_may_wakeup(dev)) {
+ shid->ops->assert_reset(shid->ops);
+
+ shid->ops->sleep_minimal_reset_delay(shid->ops);
+
+ error = shid->ops->power_up(shid->ops);
+ if (error) {
+ dev_err(dev, "%s: could not power up.", __func__);
+ shid->regulator_error_count++;
+ shid->regulator_last_error = error;
+ return;
+ }
+ shid->power_state = HIDSPI_ON;
+
+ shid->ops->deassert_reset(shid->ops);
+ }
+
+ if (shid->hid) {
+ error = hid_driver_reset_resume(shid->hid);
+ if (error)
+ dev_err(dev, "%s: failed to reset resume hid driver: %d.",
+ __func__, error);
+ }
+}
+
static void spi_hid_stop_hid(struct spi_hid *shid)
{
struct hid_device *hid = shid->hid;
@@ -1200,6 +1277,13 @@ int spi_hid_core_probe(struct spi_device *spi, struct spihid_ops *ops,
dev_err(dev, "%s: unable to request threaded IRQ.", __func__);
return error;
}
+ if (device_may_wakeup(dev)) {
+ error = dev_pm_set_wake_irq(dev, spi->irq);
+ if (error) {
+ dev_err(dev, "%s: failed to set wake IRQ.", __func__);
+ return error;
+ }
+ }
error = shid->ops->power_up(shid->ops);
if (error) {
@@ -1231,6 +1315,29 @@ void spi_hid_core_remove(struct spi_device *spi)
}
EXPORT_SYMBOL_GPL(spi_hid_core_remove);
+static int spi_hid_core_pm_suspend(struct device *dev)
+{
+ struct spi_hid *shid = dev_get_drvdata(dev);
+
+ spi_hid_suspend(shid);
+
+ return 0;
+}
+
+static int spi_hid_core_pm_resume(struct device *dev)
+{
+ struct spi_hid *shid = dev_get_drvdata(dev);
+
+ spi_hid_resume(shid);
+
+ return 0;
+}
+
+const struct dev_pm_ops spi_hid_core_pm = {
+ SYSTEM_SLEEP_PM_OPS(spi_hid_core_pm_suspend, spi_hid_core_pm_resume)
+};
+EXPORT_SYMBOL_GPL(spi_hid_core_pm);
+
MODULE_DESCRIPTION("HID over SPI transport driver");
MODULE_AUTHOR("Dmitry Antipov <dmanti@microsoft.com>");
MODULE_LICENSE("GPL");
diff --git a/drivers/hid/spi-hid/spi-hid-of.c b/drivers/hid/spi-hid/spi-hid-of.c
index 651456b6906d..80c481b77149 100644
--- a/drivers/hid/spi-hid/spi-hid-of.c
+++ b/drivers/hid/spi-hid/spi-hid-of.c
@@ -227,6 +227,7 @@ static struct spi_driver spi_hid_of_driver = {
.driver = {
.name = "spi_hid_of",
.owner = THIS_MODULE,
+ .pm = &spi_hid_core_pm,
.of_match_table = spi_hid_of_match,
.probe_type = PROBE_PREFER_ASYNCHRONOUS,
.dev_groups = spi_hid_groups,
diff --git a/drivers/hid/spi-hid/spi-hid.h b/drivers/hid/spi-hid/spi-hid.h
index f5a5f4d54beb..17b2fdf192ed 100644
--- a/drivers/hid/spi-hid/spi-hid.h
+++ b/drivers/hid/spi-hid/spi-hid.h
@@ -41,5 +41,6 @@ int spi_hid_core_probe(struct spi_device *spi, struct spihid_ops *ops,
void spi_hid_core_remove(struct spi_device *spi);
extern const struct attribute_group *spi_hid_groups[];
+extern const struct dev_pm_ops spi_hid_core_pm;
#endif /* SPI_HID_H */
--
2.53.0.1185.g05d4b7b318-goog
^ permalink raw reply related
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox