| Age | Commit message (Collapse) | Author |
|---|
|
The Multi-queue Priority (MQPRIO) and Earliest TxTime First (ETF) offloads
utilize the Time Sensitive Networking (TSN) Tx mode. This mode is always
coupled to IEEE 802.1Qbv time aware shaper (Qbv). Therefore, the driver
sets a default Qbv schedule of all gates opened and a cycle time of
1s. This schedule is set during probe.
However, the following sequence of events lead to Tx issues:
- Boot a dual core system
igc_probe():
igc_tsn_clear_schedule():
-> Default Schedule is set
Note: At this point the driver has allocated two Tx/Rx queues, because
there are only two CPUs.
- ethtool -L enp3s0 combined 4
igc_ethtool_set_channels():
igc_reinit_queues()
-> Default schedule is gone, per Tx ring start and end time are zero
- tc qdisc replace dev enp3s0 handle 100 parent root mqprio \
num_tc 4 map 3 3 2 2 0 1 1 1 3 3 3 3 3 3 3 3 \
queues 1@0 1@1 1@2 1@3 hw 1
igc_tsn_offload_apply():
igc_tsn_enable_offload():
-> Writes zeros to IGC_STQT(i) and IGC_ENDQT(i), causing Tx to stall/fail
Therefore, restore the default Qbv schedule after changing the number of
channels.
Furthermore, add a restriction to not allow queue reconfiguration when
TSN/Qbv is enabled, because it may lead to inconsistent states.
Fixes: c814a2d2d48f ("igc: Use default cycle 'start' and 'end' values for queues")
Signed-off-by: Kurt Kanzenbach <kurt@linutronix.de>
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Tested-by: Avigail Dahan <avigailx.dahan@intel.com>
Acked-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
|
|
The commit 5f6df173f92e ("ice: implement and use rd32_poll_timeout for
ice_sq_done timeout") converted ICE_CTL_Q_SQ_CMD_TIMEOUT from jiffies
to microseconds.
But the ice_release_res() function was missed, and its logic still
treats ICE_CTL_Q_SQ_CMD_TIMEOUT as a jiffies value.
So correct the issue by usecs_to_jiffies().
Found by inspection of the DDP downloading process.
Compile and modprobe tested only.
Fixes: 5f6df173f92e ("ice: implement and use rd32_poll_timeout for ice_sq_done timeout")
Signed-off-by: Ding Hui <dinghui@sangfor.com.cn>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Tested-by: Rinitha S <sx.rinitha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
|
|
When the user issues an administrative down to an interface that is the
primary for an aggregate bond, the prune lists are being purged. This
breaks communication to the secondary interface, which shares a prune
list on the main switch block while bonded together.
For the primary interface of an aggregate, avoid deleting these prune
lists during stop, and since they are hardcoded to specific values for
the default vlan and QinQ vlans, the attempt to re-add them during the
up phase will quietly fail without any additional problem.
Fixes: 1e0f9881ef79 ("ice: Flesh out implementation of support for SRIOV on bonded interface")
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Marcin Szycik <marcin.szycik@linux.intel.com>
Signed-off-by: Dave Ertman <david.m.ertman@intel.com>
Tested-by: Rinitha S <sx.rinitha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
|
|
The u64_stats_sync structure is empty on 64-bit systems. However, on 32-bit
systems it contains a seqcount_t which needs to be initialized. While the
memory is zero-initialized, a lack of u64_stats_init means that lockdep
won't get initialized properly. Fix this by adding u64_stats_init() calls
to the rings just after allocation.
Fixes: 2b245cb29421 ("ice: Implement transmit and NAPI support")
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Tested-by: Rinitha S <sx.rinitha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
|
|
Most Qualcomm platforms feature Gunyah hypervisor, which typically
handles IOMMU configuration. This includes mapping memory regions and
device memory resources for remote processors by intercepting
qcom_scm_pas_auth_and_reset() calls. These mappings are later removed
during teardown. Additionally, SHM bridge setup is required to enable
memory protection for both remoteproc metadata and its memory regions.
When the aforementioned hypervisor is absent, the operating system must
perform these configurations instead.
When Linux runs as the hypervisor (@ EL2) on a SoC, it will have its
own device tree overlay file that specifies the firmware stream ID now
managed by Linux for a particular remote processor. If the iommus
property is specified in the remoteproc device tree node, it indicates
that IOMMU configuration must be handled by Linux. In this case, the
has_iommu flag is set for the remote processor, which ensures that the
resource table, carveouts, and SHM bridge are properly configured before
memory is passed to TrustZone for authentication. Otherwise, the
has_iommu flag remains unset, which indicates default behavior.
Enables Secure PAS support for remote processors when IOMMU configuration
is managed by Linux.
Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260105-kvmrprocv10-v10-13-022e96815380@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
|
|
Qualcomm remote processor may rely on static and dynamic resources for
it to be functional. For most of the Qualcomm SoCs, when run with Gunyah
or older QHEE hypervisor, all the resources whether it is static or
dynamic, is managed by the hypervisor. Dynamic resources if it is
present for a remote processor will always be coming from secure world
via SMC call while static resources may be present in remote processor
firmware binary or it may be coming from SMC call along with dynamic
resources.
Remoteproc already has method like rproc_elf_load_rsc_table() to check
firmware binary has resources or not and if it is not having then we
pass NULL and zero as input resource table and its size argument
respectively to qcom_scm_pas_get_rsc_table() and while it has resource
present then it should pass the present resources to Trustzone(TZ) so that
it could authenticate the present resources and append dynamic resource
to return in output_rt argument along with authenticated resources.
Extend parse_fw callback to include SMC call to get resources from
Trustzone and to leverage resource table parsing and mapping and
unmapping code from the remoteproc framework.
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260105-kvmrprocv10-v10-12-022e96815380@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
|
|
Qualcomm remote processor may rely on Static and Dynamic resources for
it to be functional. Static resources are fixed like for example,
memory-mapped addresses required by the subsystem and dynamic
resources, such as shared memory in DDR etc., are determined at
runtime during the boot process.
For most of the Qualcomm SoCs, when run with Gunyah or older QHEE
hypervisor, all the resources whether it is static or dynamic, is
managed by the hypervisor. Dynamic resources if it is present for a
remote processor will always be coming from secure world via SMC call
while static resources may be present in remote processor firmware
binary or it may be coming qcom_scm_pas_get_rsc_table() SMC call along
with dynamic resources.
Some of the remote processor drivers, such as video, GPU, IPA, etc., do
not check whether resources are present in their remote processor
firmware binary. In such cases, the caller of this function should set
input_rt and input_rt_size as NULL and zero respectively. Remoteproc
framework has method to check whether firmware binary contain resources
or not and they should be pass resource table pointer to input_rt and
resource table size to input_rt_size and this will be forwarded to
TrustZone for authentication. TrustZone will then append the dynamic
resources and return the complete resource table in the passed output
buffer.
More about documentation on resource table format can be found in
include/linux/remoteproc.h
Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260105-kvmrprocv10-v10-11-022e96815380@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
|
|
On SoCs running with a non-Gunyah-based hypervisor, Linux must take
responsibility for creating the SHM bridge both for metadata (before
calling qcom_scm_pas_init_image()) and for remoteproc memory (before
calling qcom_scm_pas_auth_and_reset()). We have taken care the things
required for qcom_scm_pas_auth_and_reset(). Lets put these awareness
of above conditions into qcom_scm_pas_init_image() and
qcom_scm_pas_metadata_release().
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260105-kvmrprocv10-v10-10-022e96815380@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
|
|
Refactor qcom_scm_pas_init_image() by moving the memory allocation,
copy, and free operations to a higher-level function, and isolate the
actual SMC call in a separate function. The main intention is to allow
flexibility for different allocators and to respect any constraints that
the allocator API may impose before invoking the actual SCM function.
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260105-kvmrprocv10-v10-9-022e96815380@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
|
|
For memory passed to TrustZone (TZ), it must either be part of a pool
registered with TZ or explicitly registered via SHMbridge SMC calls.
When Gunyah hypervisor is present, PAS SMC calls from Linux running at
EL1 are trapped by Gunyah running @ EL2, which handles SHMbridge
creation for both metadata and remoteproc carveout memory before
invoking the calls to TZ.
On SoCs running with a non-Gunyah-based hypervisor, Linux must take
responsibility for creating the SHM bridge before invoking PAS SMC
calls. For the auth_and_reset() call, the remoteproc carveout memory
must first be registered with TZ via a SHMbridge SMC call and once
authentication and reset are complete, the SHMbridge memory can be
deregistered.
Introduce qcom_scm_pas_prepare_and_auth_reset(), which sets up the SHM
bridge over the remoteproc carveout memory when Linux operates at EL2.
This behavior is indicated by a new field added to the PAS context data
structure. The function then invokes the auth_and_reset SMC call.
Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260105-kvmrprocv10-v10-8-022e96815380@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
|
|
qcom_mdt_pas_init() was previously used only by the remoteproc driver
(drivers/remoteproc/qcom_q6v5_pas.c). Since that driver has now
transitioned to using PAS context-based qcom_mdt_pas_load() function,
making qcom_mdt_pas_init() obsolete for external use.
Removes qcom_mdt_pas_init() from the list of exported symbols and make
it static to limit its scope to internal use within mdtloader.
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260105-kvmrprocv10-v10-7-022e96815380@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
|
|
Introduce a new PAS context-aware function, qcom_mdt_pas_load(), for
remote processor drivers. This function utilizes the PAS context
pointer returned from qcom_scm_pas_ctx_init() to perform firmware
metadata verification and memory setup via SMC calls.
The qcom_mdt_pas_load() and qcom_mdt_load() functions are largely
similar, but the former is designed for clients using the PAS
context-based data structure. Over time, all users of qcom_mdt_load()
can be migrated to use qcom_mdt_pas_load() for consistency and
improved abstraction.
As the remoteproc PAS driver (qcom_q6v5_pas) has already adopted the
PAS context-based approach, update it to use qcom_mdt_pas_load().
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260105-kvmrprocv10-v10-6-022e96815380@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
|
|
As a superset of the existing metadata context, the PAS context
structure enables both remoteproc and non-remoteproc subsystems to
better support scenarios where the SoC runs with or without the Gunyah
hypervisor. To reflect this, relevant SCM and metadata functions are
updated to incorporate PAS context awareness and remove metadata context
data structure completely.
Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260105-kvmrprocv10-v10-5-022e96815380@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
|
|
When the Peripheral Authentication Service (PAS) method runs on a SoC
where Linux operates at EL2 (i.e., without the Gunyah hypervisor), the
reset sequences are handled by TrustZone. In such cases, Linux must
perform additional steps before invoking PAS SMC calls, such as creating
a SHM bridge. Therefore, PAS SMC calls require awareness and handling of
these additional steps when Linux runs at EL2.
To support this, there is a need for a data structure that can be
initialized prior to invoking any SMC or MDT functions. This structure
allows those functions to determine whether they are operating in the
presence or absence of the Gunyah hypervisor and behave accordingly.
Currently, remoteproc and non-remoteproc subsystems use different
variants of the MDT loader helper API, primarily due to differences in
metadata context handling. Remoteproc subsystems retain the metadata
context until authentication and reset are completed, while
non-remoteproc subsystems (e.g., video, graphics, IPA, etc.) do not
retain the metadata context and can free it within the
qcom_scm_pas_init() call by passing a NULL context parameter and due to
these differences, it is not possible to extend metadata context
handling to support remoteproc and non remoteproc subsystem use PAS
operations, when Linux operates at EL2.
Add PAS context data structure allocator helper function.
Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260105-kvmrprocv10-v10-4-022e96815380@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
|
|
Peripheral and pas_id refers to unique id for a subsystem and used only
when peripheral authentication service from secure world is utilized.
Lets rename peripheral to pas_id to reflect closer to its meaning.
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260105-kvmrprocv10-v10-3-022e96815380@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
|
|
The qcom_scm_pas_metadata_release() function already frees the allocated
memory and sets ctx->ptr to NULL. Resetting ctx->phys and ctx->size to
zero is unnecessary because the context is expected to be discarded
after release.
Lets removes redundant assignments.
Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260105-kvmrprocv10-v10-2-022e96815380@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
|
|
On most Qualcomm platforms, the Qualcomm Hypervisor Execution
Environment (QHEE) is either used as a standalone hypervisor or
co-exists as a module with the Gunyah hypervisor. When QHEE is
present, it configures firmware streams for remote processors.
If QHEE is not available, the operating system must perform
these configurations instead and for that remote processor
firmware stream should be presented to the operating system.
To support this, add the iommus property as an optional property
for PAS-supported devices.
Acked-by: Rob Herring (Arm) <robh@kernel.org>
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260105-kvmrprocv10-v10-1-022e96815380@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
|
|
Unlike the CPU, the GPU does not throttle its speed automatically when it
reaches high temperatures.
Set up GPU cooling by throttling the GPU speed
when reaching 115°C.
Signed-off-by: Gaurav Kohli <quic_gkohli@quicinc.com>
Signed-off-by: Akhil P Oommen <akhilpo@oss.qualcomm.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20250903-a623-gpu-support-v5-4-5398585e2981@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2
Pull gfs2 revert from Andreas Gruenbacher:
"Revert bad commit "gfs2: Fix use of bio_chain"
I was originally assuming that there must be a bug in gfs2
because gfs2 chains bios in the opposite direction of what
bio_chain_and_submit() expects.
It turns out that the bio chains are set up in "reverse direction"
intentionally so that the first bio's bi_end_io callback is invoked
rather than the last bio's callback.
We want the first bio's callback invoked for the following reason: The
initial bio starts page aligned and covers one or more pages. When it
terminates at a non-page-aligned offset, subsequent bios are added to
handle the remaining portion of the final page.
Upon completion of the bio chain, all affected pages need to be be
marked as read, and only the first bio references all of these pages"
* tag 'gfs2-for-6.19-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2:
Revert "gfs2: Fix use of bio_chain"
|
|
The func_id parameter is not needed in check_func_proto.
This patch removes it.
Signed-off-by: Song Chen <chensong_2000@189.cn>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://patch.msgid.link/20260105155009.4581-1-chensong_2000@189.cn
|
|
Pull x86 kvm fixes from Paolo Bonzini:
- Avoid freeing stack-allocated node in kvm_async_pf_queue_task
- Clear XSTATE_BV[i] in guest XSAVE state whenever XFD[i]=1
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
selftests: kvm: Verify TILELOADD actually #NM faults when XFD[18]=1
selftests: kvm: try getting XFD and XSAVE state out of sync
selftests: kvm: replace numbered sync points with actions
x86/fpu: Clear XSTATE_BV[i] in guest XSAVE state whenever XFD[i]=1
x86/kvm: Avoid freeing stack-allocated node in kvm_async_pf_queue_task
|
|
ASPEED BMC controllers have VGA and USB functions behind a PCIe-to-PCI
bridge that causes them to share the same StreamID:
[e0]---00.0-[e1-e2]----00.0-[e2]--+-00.0 ASPEED Graphics Family
\-02.0 ASPEED USB Controller
Both devices get StreamID 0x5e200 due to bridge aliasing, causing the USB
controller to be rejected with 'Aliasing StreamID unsupported'.
Per ASPEED, the AST1150 doesn't use a real PCI bus and always forwards
the original Requester ID from downstream devices rather than replacing
it with any alias.
Add a new PCI_DEV_FLAGS_PCI_BRIDGE_NO_ALIAS flag and apply it to the
AST1150.
Suggested-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Nirmoy Das <nirmoyd@nvidia.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Link: https://patch.msgid.link/20251217154529.377586-2-nirmoyd@nvidia.com
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/hyperv/linux
Pull hyperv fixes from Wei Liu:
- Minor fixes and cleanups for the MSHV driver
* tag 'hyperv-fixes-signed-20260112' of git://git.kernel.org/pub/scm/linux/kernel/git/hyperv/linux:
mshv: release mutex on region invalidation failure
hyperv: Avoid -Wflex-array-member-not-at-end warning
mshv: hide x86-specific functions on arm64
mshv: Initialize local variables early upon region invalidation
mshv: Use PMD_ORDER instead of HPAGE_PMD_ORDER when processing regions
|
|
Puranjay Mohan says:
====================
bpf: Recognize special arithmetic shift in the verifier
v3: https://lore.kernel.org/all/20260103022310.935686-1-puranjay@kernel.org/
Changes in v3->v4:
- Fork verifier state while processing BPF_OR when src_reg has [-1,0]
range and 2nd operand is a constant. This is to detect the following pattern:
i32 X > -1 ? C1 : -1 --> (X >>s 31) | C1
- Add selftests for above.
- Remove __description("s>>=63") (Eduard in another patchset)
v2: https://lore.kernel.org/bpf/20251115022611.64898-1-alexei.starovoitov@gmail.com/
Changes in v2->v3:
- fork verifier state while processing BPF_AND when src_reg has [-1,0]
range and 2nd operand is a constant.
v1->v2:
Use __mark_reg32_known() or __mark_reg_known() for zero too.
Add comment to selftest.
v1:
https://lore.kernel.org/bpf/20251114031039.63852-1-alexei.starovoitov@gmail.com/
====================
Link: https://patch.msgid.link/20260112201424.816836-1-puranjay@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
Add tests for special arithmetic shift right.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Co-developed-by: Puranjay Mohan <puranjay@kernel.org>
Signed-off-by: Puranjay Mohan <puranjay@kernel.org>
Link: https://lore.kernel.org/r/20260112201424.816836-3-puranjay@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
cilium bpf_wiregard.bpf.c when compiled with -O1 fails to load
with the following verifier log:
192: (79) r2 = *(u64 *)(r10 -304) ; R2=pkt(r=40) R10=fp0 fp-304=pkt(r=40)
...
227: (85) call bpf_skb_store_bytes#9 ; R0=scalar()
228: (bc) w2 = w0 ; R0=scalar() R2=scalar(smin=0,smax=umax=0xffffffff,var_off=(0x0; 0xffffffff))
229: (c4) w2 s>>= 31 ; R2=scalar(smin=0,smax=umax=0xffffffff,smin32=-1,smax32=0,var_off=(0x0; 0xffffffff))
230: (54) w2 &= -134 ; R2=scalar(smin=0,smax=umax=umax32=0xffffff7a,smax32=0x7fffff7a,var_off=(0x0; 0xffffff7a))
...
232: (66) if w2 s> 0xffffffff goto pc+125 ; R2=scalar(smin=umin=umin32=0x80000000,smax=umax=umax32=0xffffff7a,smax32=-134,var_off=(0x80000000; 0x7fffff7a))
...
238: (79) r4 = *(u64 *)(r10 -304) ; R4=scalar() R10=fp0 fp-304=scalar()
239: (56) if w2 != 0xffffff78 goto pc+210 ; R2=0xffffff78 // -136
...
258: (71) r1 = *(u8 *)(r4 +0)
R4 invalid mem access 'scalar'
The error might confuse most bpf authors, since fp-304 slot had 'pkt'
pointer at insn 192 and became 'scalar' at 238. That happened because
bpf_skb_store_bytes() clears all packet pointers including those in
the stack. On the first glance it might look like a bug in the source
code, since ctx->data pointer should have been reloaded after the call
to bpf_skb_store_bytes().
The relevant part of cilium source code looks like this:
// bpf/lib/nodeport.h
int dsr_set_ipip6()
{
if (ctx_adjust_hroom(...))
return DROP_INVALID; // -134
if (ctx_store_bytes(...))
return DROP_WRITE_ERROR; // -141
return 0;
}
bool dsr_fail_needs_reply(int code)
{
if (code == DROP_FRAG_NEEDED) // -136
return true;
return false;
}
tail_nodeport_ipv6_dsr()
{
ret = dsr_set_ipip6(...);
if (!IS_ERR(ret)) {
...
} else {
if (dsr_fail_needs_reply(ret))
return dsr_reply_icmp6(...);
}
}
The code doesn't have arithmetic shift by 31 and it reloads ctx->data
every time it needs to access it. So it's not a bug in the source code.
The reason is DAGCombiner::foldSelectCCToShiftAnd() LLVM transformation:
// If this is a select where the false operand is zero and the compare is a
// check of the sign bit, see if we can perform the "gzip trick":
// select_cc setlt X, 0, A, 0 -> and (sra X, size(X)-1), A
// select_cc setgt X, 0, A, 0 -> and (not (sra X, size(X)-1)), A
The conditional branch in dsr_set_ipip6() and its return values
are optimized into BPF_ARSH plus BPF_AND:
227: (85) call bpf_skb_store_bytes#9
228: (bc) w2 = w0
229: (c4) w2 s>>= 31 ; R2=scalar(smin=0,smax=umax=0xffffffff,smin32=-1,smax32=0,var_off=(0x0; 0xffffffff))
230: (54) w2 &= -134 ; R2=scalar(smin=0,smax=umax=umax32=0xffffff7a,smax32=0x7fffff7a,var_off=(0x0; 0xffffff7a))
after insn 230 the register w2 can only be 0 or -134,
but the verifier approximates it, since there is no way to
represent two scalars in bpf_reg_state.
After fallthough at insn 232 the w2 can only be -134,
hence the branch at insn
239: (56) if w2 != -136 goto pc+210
should be always taken, and trapping insn 258 should never execute.
LLVM generated correct code, but the verifier follows impossible
path and rejects valid program. To fix this issue recognize this
special LLVM optimization and fork the verifier state.
So after insn 229: (c4) w2 s>>= 31
the verifier has two states to explore:
one with w2 = 0 and another with w2 = 0xffffffff
which makes the verifier accept bpf_wiregard.c
A similar pattern exists were OR operation is used in place of the AND
operation, the verifier detects that pattern as well by forking the
state before the OR operation with a scalar in range [-1,0].
Note there are 20+ such patterns in bpf_wiregard.o compiled
with -O1 and -O2, but they're rarely seen in other production
bpf programs, so push_stack() approach is not a concern.
Reported-by: Hao Sun <sunhao.th@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Co-developed-by: Puranjay Mohan <puranjay@kernel.org>
Signed-off-by: Puranjay Mohan <puranjay@kernel.org>
Link: https://lore.kernel.org/r/20260112201424.816836-2-puranjay@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
Yonghong Song says:
====================
Fix a few selftest failure due to 64K page
Fix a few arm64 selftest failures due to 64K page. Please see each
indvidual patch for why the test failed and how the test gets fixed.
====================
Link: https://patch.msgid.link/20260113061018.3797051-1-yonghong.song@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
With 64K page on arm64, verifier_arena_globals1 failed like below:
...
libbpf: map 'arena': failed to create: -E2BIG
...
#509/1 verifier_arena_globals1/check_reserve1:FAIL
...
For 64K page, if the number of arena pages is (1UL << 20), the total
memory will exceed 4G and this will cause map creation failure.
Adjusting ARENA_PAGES based on the actual page size fixed the problem.
Cc: Emil Tsalapatis <emil@etsalapatis.com>
Signed-off-by: Yonghong Song <yonghong.song@linux.dev>
Reviewed-by: Emil Tsalapatis <emil@etsalapatis.com>
Link: https://lore.kernel.org/r/20260113061033.3798549-1-yonghong.song@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
The current selftest sk_bypass_prot_mem only supports 4K page.
When running with 64K page on arm64, the following failure happens:
...
check_bypass:FAIL:no bypass unexpected no bypass: actual 3 <= expected 32
...
#385/1 sk_bypass_prot_mem/TCP :FAIL
...
check_bypass:FAIL:no bypass unexpected no bypass: actual 4 <= expected 32
...
#385/2 sk_bypass_prot_mem/UDP :FAIL
...
Adding support to 64K page as well fixed the failure.
Cc: Kuniyuki Iwashima <kuniyu@google.com>
Signed-off-by: Yonghong Song <yonghong.song@linux.dev>
Link: https://lore.kernel.org/r/20260113061028.3798326-1-yonghong.song@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
On arm64 with 64K page , I observed the following test failure:
...
subtest_dmabuf_iter_check_lots_of_buffers:FAIL:total_bytes_read unexpected total_bytes_read:
actual 4696 <= expected 65536
#97/3 dmabuf_iter/lots_of_buffers:FAIL
With 4K page on x86, the total_bytes_read is 4593.
With 64K page on arm64, the total_byte_read is 4696.
In progs/dmabuf_iter.c, for each iteration, the output is
BPF_SEQ_PRINTF(seq, "%lu\n%llu\n%s\n%s\n", inode, size, name, exporter);
The only difference between 4K and 64K page is 'size' in
the above BPF_SEQ_PRINTF. The 4K page will output '4096' and
the 64K page will output '65536'. So the total_bytes_read with 64K page
is slighter greater than 4K page.
Adjusting the total_bytes_read from 65536 to 4096 fixed the issue.
Cc: T.J. Mercier <tjmercier@google.com>
Signed-off-by: Yonghong Song <yonghong.song@linux.dev>
Link: https://lore.kernel.org/r/20260113061023.3798085-1-yonghong.song@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
It's quite likely that only register opcode restrictions exists, in
which case we'd never need to check the normal opcodes. Split
ctx->restricted into two separate fields, one for I/O opcodes, and one
for register opcodes.
Reviewed-by: Gabriel Krisman Bertazi <krisman@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
Just a cleanup, makes the code easier to read without too many dependent
nested checks.
Reviewed-by: Gabriel Krisman Bertazi <krisman@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
Rather than defer this until the rings are enabled, just set it
upfront when the restrictions are parsed and enabled anyway. There's
no reason to defer this setting until the rings are enabled.
Reviewed-by: Gabriel Krisman Bertazi <krisman@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
Replace the pattern of declaring a local regs array from cur_regs()
and then indexing into it with the more concise reg_state() helper.
This simplifies the code by eliminating intermediate variables and
makes register access more consistent throughout the verifier.
Signed-off-by: Mykyta Yatsenko <yatsenko@meta.com>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/r/20260113134826.2214860-1-mykyta.yatsenko5@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
Rather than leave this to the caller, have io_parse_restrictions() set
->registered = true if restrictions have been enabled. This is in
preparation for having finer grained restrictions.
Reviewed-by: Gabriel Krisman Bertazi <krisman@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
Rather than return 0 on success, return >= 0 for success, where the
return value is that number of parsed entries. As before, any < 0
return is an error.
Reviewed-by: Gabriel Krisman Bertazi <krisman@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
Update dw_i3c_master_restore_addrs() to preserve existing bits in each
Device Address Table (DAT) entry when restoring addresses. This prevents
overwriting configuration bits during PM runtime resumes.
Signed-off-by: Adrian Ng Ho Yin <adrianhoyin.ng@altera.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Link: https://patch.msgid.link/46112c0da44110f46709cb0e7a4595e312b95c10.1765529948.git.adrianhoyin.ng@altera.com
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
|
|
Add DEV_ADDR_TABLE_DYNAMIC_MASK / DEV_ADDR_TABLE_DYNAMIC_ADDR(x) for
dynamic device addresses and DEV_ADDR_TABLE_STATIC_MASK /
DEV_ADDR_TABLE_STATIC_ADDR(x) for static device addresses in the I3C
address table. Replace manual shift-and-mask with FIELD_PREP() for both
dynamic and static addresses for clarity and maintainability.
Signed-off-by: Adrian Ng Ho Yin <adrianhoyin.ng@altera.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Link: https://patch.msgid.link/d72896e510db1870d26a794f131f600c7e42cf00.1765529948.git.adrianhoyin.ng@altera.com
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
|
|
The DesignWare I3C controller supports automatically retrying transactions
when a device NACKs. This is useful for slave devices that may be
temporarily busy and not ready to respond immediately.
Add new ops to configure all active DAT entry with dev_nack_retry during
runtime. Returns error when value exceeds hw specified limit.
Signed-off-by: Adrian Ng Ho Yin <adrianhoyin.ng@altera.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Link: https://patch.msgid.link/f09ee67e61d31f0a12a0bf48f01e9057ca9e2fb7.1765529948.git.adrianhoyin.ng@altera.com
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
|
|
Document sysfs attribute dev_nack_retry_cnt that controls the number of
automatic retries performed by the I3C controller when a target device
returns a NACK
Add a `dev_nack_retry_count` sysfs attribute to allow reading and updating
the device NACK retry count. A new `dev_nack_retry_count` field and an
optional `set_dev_nack_retry()` callback are added to
i3c_master_controller. The attribute is created only when the callback is
implemented.
Updates are applied under the I3C bus maintenance lock to ensure safe
hardware reconfiguration.
Signed-off-by: Adrian Ng Ho Yin <adrianhoyin.ng@altera.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Link: https://patch.msgid.link/3c4b5082bde64024fc383c44bebeef89ad3c7ed3.1765529948.git.adrianhoyin.ng@altera.com
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
|
|
Merge series from Jon Hunter <jonathanh@nvidia.com>:
This series includes fixes for the realtek,rt5640 dt-binding to address
a few warnings that are observed when running the CHECK_DTBS=y for some
DTBs that use this codec.
|
|
This dereference of sched_clock_timer::function was missed when the
hrtimer callback function pointer was marked private.
Fixes: 04257da0c99c ("hrtimers: Make callback function pointer private")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Thomas Gleixner <tglx@kernel.org>
Link: https://patch.msgid.link/875x95jw7q.ffs@tglx
Closes: https://lore.kernel.org/oe-kbuild-all/202601131713.KsxhXQ0M-lkp@intel.com/
|
|
- Reorder some #define blocks
- Document mixer/volume levels
- Document some Ctl request fields (tag, len)
- replace some magic numbers with macros
No functional change intended. The information is based on reverse
engineering.
Signed-off-by: fenugrec <fenugrec@mail.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Link: https://patch.msgid.link/20260111-preso_clean1-v2-2-44b4e5129a75@mail.com
|
|
Use __le32 types for USB control transfers
Signed-off-by: fenugrec <fenugrec@mail.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Link: https://patch.msgid.link/20260111-preso_clean1-v2-1-44b4e5129a75@mail.com
|
|
Move the selinux_blob_sizes struct so it adjacent to the rest of the
SELinux initialization code and not in the middle of the LSM hook
callbacks.
Signed-off-by: Paul Moore <paul@paul-moore.com>
|
|
The minimum/maximum DSC input (i.e. pipe) and compressed (i.e. link) BPP
limits are computed already in intel_dp_compute_config_limits(), so
there is no need to do this again in
mst_stream_dsc_compute_link_config() called later. Remove the
corresponding alignments from the latter function and use the
precomputed (aligned and within bounds) maximum pipe BPP and the min/max
compressed BPP values instead as-is.
Reviewed-by: Jouni Högander <jouni.hogander@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Link: https://patch.msgid.link/20251222153547.713360-21-imre.deak@intel.com
|
|
intel_edp_dsc_compute_pipe_bpp() matches now
intel_dp_dsc_compute_pipe_bpp(), remove the former function.
Reviewed-by: Vinod Govindapillai <vinod.govindapillai@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Link: https://patch.msgid.link/20251222153547.713360-20-imre.deak@intel.com
|
|
Move computing the eDP compressed BPP value to the function computing
this for DP, allowing further simplifications later.
Reviewed-by: Vinod Govindapillai <vinod.govindapillai@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Link: https://patch.msgid.link/20251222153547.713360-19-imre.deak@intel.com
|
|
If dsc_compute_compressed_bpp() failed with a forced pipe BPP value
(where the forced pipe BPP value itself is valid within the min/max pipe
BPP limits), the function will also fail when called with the maximum
pipe BPP value: dsc_compute_compressed_bpp() will try all compressed
BPPs below the passed in pipe BPP value and if the function failed with
a given (low) compressed BPP value it will also fail with a compressed
BPP value higher than the one which failed already.
Based on the above remove the logic to retry computing a compressed BPP
value with the maximum pipe BPP value if computing the compressed BPP
failed already with the (lower) forced pipe BPP value.
Reviewed-by: Vinod Govindapillai <vinod.govindapillai@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Link: https://patch.msgid.link/20251222153547.713360-18-imre.deak@intel.com
|
|
The maximum pipe BPP value (used as the DSC input BPP) has been aligned
already to the corresponding source/sink input BPP capabilities in
intel_dp_compute_config_limits(). So it isn't needed to perform the same
alignment again in intel_dp_dsc_compute_pipe_bpp() called later, this
function can simply use the already aligned maximum pipe BPP value, do
that.
Also, there is no point in trying pipe BPP values lower than the
maximum: this would only make dsc_compute_compressed_bpp() start with a
lower _compressed_ BPP value, but this lower compressed BPP value has
been tried already when dsc_compute_compressed_bpp() was called with the
higher pipe BPP value (i.e. the first dsc_compute_compressed_bpp() call
tries already all the possible compressed BPP values which are all below
the pipe BPP value passed to it). Simplify the function accordingly
trying only the maximum pipe BPP value.
Reviewed-by: Vinod Govindapillai <vinod.govindapillai@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Link: https://patch.msgid.link/20251222153547.713360-17-imre.deak@intel.com
|
