The DPSUB DT bindings now specify ports to model the connections with
the programmable logic and the DisplayPort output. Add them to the
device tree.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Michal Simek <michal.simek@amd.com>
Add partial live video support, with a single video input that bypasses
blending. Skip registration of the DRM device in that case, but register
the DRM bridge instead. The DRM device will be created by the driver for
the display controller in the PL.
Full live video mode with concurrent usage of the video and gfx inputs,
and blending in the DPSUB video pipeline, is currently unsupported.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
To prepare for usage of the DPSUB as a DisplayPort bridge without
creating a DRM device, make initialization and usage of the DMA engine
optional. The flag that controls this feature is currently hardcoded to
operating with the DMA engine, this will be made dynamic based on the
device tree configuration in a subsequent change.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Add a mode parameter to the zynqmp_disp_layer_enable() to set the layer
mode, to prepare for live mode support.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
To prepare for live video input support, parse the device tree to find
the connected ports. Warn about unsupported configurations, and error
out when invalid.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The better convey its purpose, rename the zynqmp_dpsub_handle_vblank()
function that belongs to the DRM layer to
zynqmp_dpsub_drm_handle_vblank().
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
To complete the decoupling of the DRM device from the zynqmp_dpsub,
group all DRM-related structures in a zynqmp_dpsub_drm structure and
allocate it separately from the zynqmp_dpsub. The DRM managed allocation
of the drm_device now doesn't cover the zynqmp_dpsub anymore, so we need
to register a cleanup action to release the zynqmp_dpsub when the
drm_device is released.
The will allow usage of the DisplayPort encoder as a standalone bridge,
without registering a DRM device in this driver.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Continue the isolation of DRM/KMS code by moving all DRM init and
cleanup from zynqmp_dpsub.c to zynqmp_kms.c.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The zynqmp_disp and zynqmp_dp structures are allocated with
drmm_kzalloc(). While this simplifies management of memory, it requires
a DRM device, which will not be available at probe time when the DP
bridge will be used standalone, with a DRM device in the PL. To prepare
for this, switch to manual allocation for zynqmp_disp and zynqmp_dp. The
cleanup still uses the DRM managed infrastructure, but one level up, at
the top level. This will be addressed separately.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
There's no need to delay bridge initialization, move it to
zynqmp_dp_probe() and drop the zynqmp_dp_drm_init() function.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
To prepare for operating as a standalone DP bridge with the DRM device
implemented in the PL, move registration of the AUX bus to bridge attach
time, as that's the earliest point when a DRM device is available.
The DRM device pointer stored in zynqmp_dp isn't used anymore, drop it.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Decouple the planes handling from the display controller programming by
moving the corresponding code from zynqmp_disp.c to zynqmp_kms.c. This
prepares for using the DPSUB with a live video input, without creating
DRM planes in the DPSUB driver.
While at it, fix a typo in a comment.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Decouple the CRTC handling from the display controller programming by
moving the corresponding code from zynqmp_disp.c to zynqmp_kms.c. This
prepares for using the DPSUB with a live video input, without creating a
DRM CRTC in the DPSUB driver.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Start preparation for using the DPSUB as a standalone DisplayPort
encoder without a display controller by moving the DRM/KMS
initialization to a new zynqmp_kms.c file. No functional change
intended.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Decouple the zynqmp_disp, which handles the hardware configuration, from
the DRM planes by moving the planes to the zynqmp_dpsub structure. The
planes handling code will be moved to a separate file in a subsequent
step.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Decouple the zynqmp_disp, which handles the hardware configuration, from
the DRM CRTC by moving the CRTC to the zynqmp_dpsub structure. The CRTC
handling code will be moved to a separate file in a subsequent step.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The audio clock is an external resource from the DPSUB point of view,
not a resource internal to the display controller. Move it to the
zynqmp_dpsub structure, to allow accessing it from outside the disp
code.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The video clock is an external resource from the DPSUB point of view,
not a resource internal to the display controller. Move it to the
zynqmp_dpsub structure, to allow accessing it from outside the disp
code.
While at it, rename the fields from pclk and pclk_from_ps to vid_clk and
vid_clk_from_ps, to better reflect their purpose and match the
documentation.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The array of formats passed to drm_universal_plane_init() doesn't need
to outlive the function call, as it's copied internally. Use kcalloc()
instead of drmm_kcalloc() to allocate it, and free it right after usage.
While at it, move the allocation and initialization of the formats array
to a separate function, to prepare for splitting the DRM plane handling
to a separate file.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Use the ARRAY_SIZE() macro to iterate over arrays, instead of hardcoding
their size. This makes the code less error-prone should the array size
change.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The zynqmp_disp_layer_set_format() function only needs format
information, not a full plane state. Get the necessary info from the
plane state in the caller and pass it to zynqmp_disp_layer_set_format().
This prepares for calling the function from non-DRM code. This doesn't
introduce any functional change.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reuse the local info variable instead of going through the layer pointer
in zynqmp_disp_layer_update(). This doesn't introduce any functional
change.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
To prepare for control of the blender outside of the CRTC code, move the
setup of the blender to the zynqmp_disp_enable() function. This doesn't
introduce any functional change.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
To prepare for usage of the clock setup function outside of the CRTC
code, replace the DRM-specific structures passed as parameters with a
pointer to the zynqmp_disp and the requested clock rate. This doesn't
introduce any functional change.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Now that the driver uses the connector bridge helper, HPD can be
reported directly for the connector through the drm_bridge_hpd_notify()
function.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Replace the manual connector implementation and registration in the DP
encoder with the DRM connector bridge helper. This removes boilerplate
code and simplifies the driver.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The next component in the display chain, after the DP encoder, is most
likely a DP connector. The display connector driver registers a bridge
for it. That bridge doesn't need to be controlled, but is needed in
order to use the DRM connector bridge helper. Retrieve it at init time,
and attach to it in the DP bridge attach handler.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
As part of the transitition of the DP encoder to a DRM bridge, turn the
DRM encoder into a dummy encoder and move it out of the DP code, to the
DPSUB core. DP encoder operations are handled by the DP bridge, which is
now attached to the encoder.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Connector creation requires the DRM encoder, and it thus typically
performed in the bridge attach operation. Move it there, to prepare for
registration of the DRM bridge. For now the zynqmp_dp_bridge_attach() is
called manually at initialization time.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
To prepare for the removal of the connector from the DP encoder, pass
the display info pointer to the zynqmp_dp_set_format() function instead
of accessing the connector internally. The display info is NULL when the
function is called at initialization time, as we have no display info at
that point. This doesn't change the existing behaviour, given that the
zynqmp_dp_set_format() was already handling this as a special case (the
display info isn't initialized at init time and is all zeroes).
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The DP encoder is currently modelled as a DRM encoder and DRM connector.
This doesn't support system configurations where the DP encoder is
driven by the FPGA programmable logic, using the live video input to the
DP subsystem. To enable such use cases, we need to model the encoder as
a DRM bridge.
As a first step, create a DRM bridge in the DP encoder driver. Move and
delegate the implementation of the DRM encoder and connector operations
to the bridge to prepare for the transition. The bridge will be
registered with the DRM core as a separate change.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The zynqmp_dp_encoder_mode_set_transfer_unit() function takes a mode
pointer argument that it doesn't need to modify. Make it const.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
To prepare for the transition to the DRM bridge API, switch the encoder
operations to the atomic versions of .enable() and .disable(). This
doesn't cause any functional change by itself.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The DPSUB doesn't live in isolation, but is connected to the
programmable logic for live inputs and outputs, and also has a
DisplayPort output. Model all those using OF graph.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Pull more random number generator updates from Jason Donenfeld:
"This time with some large scale treewide cleanups.
The intent of this pull is to clean up the way callers fetch random
integers. The current rules for doing this right are:
- If you want a secure or an insecure random u64, use get_random_u64()
- If you want a secure or an insecure random u32, use get_random_u32()
The old function prandom_u32() has been deprecated for a while
now and is just a wrapper around get_random_u32(). Same for
get_random_int().
- If you want a secure or an insecure random u16, use get_random_u16()
- If you want a secure or an insecure random u8, use get_random_u8()
- If you want secure or insecure random bytes, use get_random_bytes().
The old function prandom_bytes() has been deprecated for a while
now and has long been a wrapper around get_random_bytes()
- If you want a non-uniform random u32, u16, or u8 bounded by a
certain open interval maximum, use prandom_u32_max()
I say "non-uniform", because it doesn't do any rejection sampling
or divisions. Hence, it stays within the prandom_*() namespace, not
the get_random_*() namespace.
I'm currently investigating a "uniform" function for 6.2. We'll see
what comes of that.
By applying these rules uniformly, we get several benefits:
- By using prandom_u32_max() with an upper-bound that the compiler
can prove at compile-time is ≤65536 or ≤256, internally
get_random_u16() or get_random_u8() is used, which wastes fewer
batched random bytes, and hence has higher throughput.
- By using prandom_u32_max() instead of %, when the upper-bound is
not a constant, division is still avoided, because
prandom_u32_max() uses a faster multiplication-based trick instead.
- By using get_random_u16() or get_random_u8() in cases where the
return value is intended to indeed be a u16 or a u8, we waste fewer
batched random bytes, and hence have higher throughput.
This series was originally done by hand while I was on an airplane
without Internet. Later, Kees and I worked on retroactively figuring
out what could be done with Coccinelle and what had to be done
manually, and then we split things up based on that.
So while this touches a lot of files, the actual amount of code that's
hand fiddled is comfortably small"
* tag 'random-6.1-rc1-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/crng/random:
prandom: remove unused functions
treewide: use get_random_bytes() when possible
treewide: use get_random_u32() when possible
treewide: use get_random_{u8,u16}() when possible, part 2
treewide: use get_random_{u8,u16}() when possible, part 1
treewide: use prandom_u32_max() when possible, part 2
treewide: use prandom_u32_max() when possible, part 1
Pull more perf tools updates from Arnaldo Carvalho de Melo:
- Use BPF CO-RE (Compile Once, Run Everywhere) to support old kernels
when using bperf (perf BPF based counters) with cgroups.
- Support HiSilicon PCIe Performance Monitoring Unit (PMU), that
monitors bandwidth, latency, bus utilization and buffer occupancy.
Documented in Documentation/admin-guide/perf/hisi-pcie-pmu.rst.
- User space tasks can migrate between CPUs, so when tracing selected
CPUs, system-wide sideband is still needed, fix it in the setup of
Intel PT on hybrid systems.
- Fix metricgroups title message in 'perf list', it should state that
the metrics groups are to be used with the '-M' option, not '-e'.
- Sync the msr-index.h copy with the kernel sources, adding support for
using "AMD64_TSC_RATIO" in filter expressions in 'perf trace' as well
as decoding it when printing the MSR tracepoint arguments.
- Fix program header size and alignment when generating a JIT ELF in
'perf inject'.
- Add multiple new Intel PT 'perf test' entries, including a jitdump
one.
- Fix the 'perf test' entries for 'perf stat' CSV and JSON output when
running on PowerPC due to an invalid topology number in that arch.
- Fix the 'perf test' for arm_coresight failures on the ARM Juno
system.
- Fix the 'perf test' attr entry for PERF_FORMAT_LOST, adding this
option to the or expression expected in the intercepted
perf_event_open() syscall.
- Add missing condition flags ('hs', 'lo', 'vc', 'vs') for arm64 in the
'perf annotate' asm parser.
- Fix 'perf mem record -C' option processing, it was being chopped up
when preparing the underlying 'perf record -e mem-events' and thus
being ignored, requiring using '-- -C CPUs' as a workaround.
- Improvements and tidy ups for 'perf test' shell infra.
- Fix Intel PT information printing segfault in uClibc, where a NULL
format was being passed to fprintf.
* tag 'perf-tools-for-v6.1-2-2022-10-16' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux: (23 commits)
tools arch x86: Sync the msr-index.h copy with the kernel sources
perf auxtrace arm64: Add support for parsing HiSilicon PCIe Trace packet
perf auxtrace arm64: Add support for HiSilicon PCIe Tune and Trace device driver
perf auxtrace arm: Refactor event list iteration in auxtrace_record__init()
perf tests stat+json_output: Include sanity check for topology
perf tests stat+csv_output: Include sanity check for topology
perf intel-pt: Fix system_wide dummy event for hybrid
perf intel-pt: Fix segfault in intel_pt_print_info() with uClibc
perf test: Fix attr tests for PERF_FORMAT_LOST
perf test: test_intel_pt.sh: Add 9 tests
perf inject: Fix GEN_ELF_TEXT_OFFSET for jit
perf test: test_intel_pt.sh: Add jitdump test
perf test: test_intel_pt.sh: Tidy some alignment
perf test: test_intel_pt.sh: Print a message when skipping kernel tracing
perf test: test_intel_pt.sh: Tidy some perf record options
perf test: test_intel_pt.sh: Fix return checking again
perf: Skip and warn on unknown format 'configN' attrs
perf list: Fix metricgroups title message
perf mem: Fix -C option behavior for perf mem record
perf annotate: Add missing condition flags for arm64
...
Pull Kbuild fixes from Masahiro Yamada:
- Fix CONFIG_DEBUG_INFO_DWARF_TOOLCHAIN_DEFAULT=y compile error for the
combination of Clang >= 14 and GAS <= 2.35.
- Drop vmlinux.bz2 from the rpm package as it just annoyingly increased
the package size.
- Fix modpost error under build environments using musl.
- Make *.ll files keep value names for easier debugging
- Fix single directory build
- Prevent RISC-V from selecting the broken DWARF5 support when Clang
and GAS are used together.
* tag 'kbuild-fixes-v6.1' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
lib/Kconfig.debug: Add check for non-constant .{s,u}leb128 support to DWARF5
kbuild: fix single directory build
kbuild: add -fno-discard-value-names to cmd_cc_ll_c
scripts/clang-tools: Convert clang-tidy args to list
modpost: put modpost options before argument
kbuild: Stop including vmlinux.bz2 in the rpm's
Kconfig.debug: add toolchain checks for DEBUG_INFO_DWARF_TOOLCHAIN_DEFAULT
Kconfig.debug: simplify the dependency of DEBUG_INFO_DWARF4/5
Pull more clk updates from Stephen Boyd:
"This is the final part of the clk patches for this merge window.
The clk rate range series needed another week to fully bake. Maxime
fixed the bug that broke clk notifiers and prevented this from being
included in the first pull request. He also added a unit test on top
to make sure it doesn't break so easily again. The majority of the
series fixes up how the clk_set_rate_*() APIs work, particularly
around when the rate constraints are dropped and how they move around
when reparenting clks. Overall it's a much needed improvement to the
clk rate range APIs that used to be pretty broken if you looked
sideways.
Beyond the core changes there are a few driver fixes for a compilation
issue or improper data causing clks to fail to register or have the
wrong parents. These are good to get in before the first -rc so that
the system actually boots on the affected devices"
* tag 'clk-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux: (31 commits)
clk: tegra: Fix Tegra PWM parent clock
clk: at91: fix the build with binutils 2.27
clk: qcom: gcc-msm8660: Drop hardcoded fixed board clocks
clk: mediatek: clk-mux: Add .determine_rate() callback
clk: tests: Add tests for notifiers
clk: Update req_rate on __clk_recalc_rates()
clk: tests: Add missing test case for ranges
clk: qcom: clk-rcg2: Take clock boundaries into consideration for gfx3d
clk: Introduce the clk_hw_get_rate_range function
clk: Zero the clk_rate_request structure
clk: Stop forwarding clk_rate_requests to the parent
clk: Constify clk_has_parent()
clk: Introduce clk_core_has_parent()
clk: Switch from __clk_determine_rate to clk_core_round_rate_nolock
clk: Add our request boundaries in clk_core_init_rate_req
clk: Introduce clk_hw_init_rate_request()
clk: Move clk_core_init_rate_req() from clk_core_round_rate_nolock() to its caller
clk: Change clk_core_init_rate_req prototype
clk: Set req_rate on reparenting
clk: Take into account uncached clocks in clk_set_rate_range()
...
Pull more cifs updates from Steve French:
- fix a regression in guest mounts to old servers
- improvements to directory leasing (caching directory entries safely
beyond the root directory)
- symlink improvement (reducing roundtrips needed to process symlinks)
- an lseek fix (to problem where some dir entries could be skipped)
- improved ioctl for returning more detailed information on directory
change notifications
- clarify multichannel interface query warning
- cleanup fix (for better aligning buffers using ALIGN and round_up)
- a compounding fix
- fix some uninitialized variable bugs found by Coverity and the kernel
test robot
* tag '6.1-rc-smb3-client-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6:
smb3: improve SMB3 change notification support
cifs: lease key is uninitialized in two additional functions when smb1
cifs: lease key is uninitialized in smb1 paths
smb3: must initialize two ACL struct fields to zero
cifs: fix double-fault crash during ntlmssp
cifs: fix static checker warning
cifs: use ALIGN() and round_up() macros
cifs: find and use the dentry for cached non-root directories also
cifs: enable caching of directories for which a lease is held
cifs: prevent copying past input buffer boundaries
cifs: fix uninitialised var in smb2_compound_op()
cifs: improve symlink handling for smb2+
smb3: clarify multichannel warning
cifs: fix regression in very old smb1 mounts
cifs: fix skipping to incorrect offset in emit_cached_dirents
When building with a RISC-V kernel with DWARF5 debug info using clang
and the GNU assembler, several instances of the following error appear:
/tmp/vgettimeofday-48aa35.s:2963: Error: non-constant .uleb128 is not supported
Dumping the .s file reveals these .uleb128 directives come from
.debug_loc and .debug_ranges:
.Ldebug_loc0:
.byte 4 # DW_LLE_offset_pair
.uleb128 .Lfunc_begin0-.Lfunc_begin0 # starting offset
.uleb128 .Ltmp1-.Lfunc_begin0 # ending offset
.byte 1 # Loc expr size
.byte 90 # DW_OP_reg10
.byte 0 # DW_LLE_end_of_list
.Ldebug_ranges0:
.byte 4 # DW_RLE_offset_pair
.uleb128 .Ltmp6-.Lfunc_begin0 # starting offset
.uleb128 .Ltmp27-.Lfunc_begin0 # ending offset
.byte 4 # DW_RLE_offset_pair
.uleb128 .Ltmp28-.Lfunc_begin0 # starting offset
.uleb128 .Ltmp30-.Lfunc_begin0 # ending offset
.byte 0 # DW_RLE_end_of_list
There is an outstanding binutils issue to support a non-constant operand
to .sleb128 and .uleb128 in GAS for RISC-V but there does not appear to
be any movement on it, due to concerns over how it would work with
linker relaxation.
To avoid these build errors, prevent DWARF5 from being selected when
using clang and an assembler that does not have support for these symbol
deltas, which can be easily checked in Kconfig with as-instr plus the
small test program from the dwz test suite from the binutils issue.
Link: https://sourceware.org/bugzilla/show_bug.cgi?id=27215
Link: https://github.com/ClangBuiltLinux/linux/issues/1719
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Commit f110e5a250 ("kbuild: refactor single builds of *.ko") was wrong.
KBUILD_MODULES _is_ needed for single builds.
Otherwise, "make foo/bar/baz/" does not build module objects at all.
Fixes: f110e5a250 ("kbuild: refactor single builds of *.ko")
Reported-by: David Sterba <dsterba@suse.cz>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Tested-by: David Sterba <dsterba@suse.com>
Pull slab hotfix from Vlastimil Babka:
"A single fix for the common-kmalloc series, for warnings on mips and
sparc64 reported by Guenter Roeck"
* tag 'slab-for-6.1-rc1-hotfix' of git://git.kernel.org/pub/scm/linux/kernel/git/vbabka/slab:
mm/slab: use kmalloc_node() for off slab freelist_idx_t array allocation
Pull OpenRISC updates from Stafford Horne:
"I have relocated to London so not much work from me while I get
settled.
Still, OpenRISC picked up two patches in this window:
- Fix for kernel page table walking from Jann Horn
- MAINTAINER entry cleanup from Palmer Dabbelt"
* tag 'for-linus' of https://github.com/openrisc/linux:
MAINTAINERS: git://github -> https://github.com for openrisc
openrisc: Fix pagewalk usage in arch_dma_{clear, set}_uncached
Pull pci fix from Bjorn Helgaas:
"Revert the attempt to distribute spare resources to unconfigured
hotplug bridges at boot time.
This fixed some dock hot-add scenarios, but Jonathan Cameron reported
that it broke a topology with a multi-function device where one
function was a Switch Upstream Port and the other was an Endpoint"
* tag 'pci-v6.1-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci:
Revert "PCI: Distribute available resources for root buses, too"
After commit d6a71648db ("mm/slab: kmalloc: pass requests larger than
order-1 page to page allocator"), SLAB passes large ( > PAGE_SIZE * 2)
requests to buddy like SLUB does.
SLAB has been using kmalloc caches to allocate freelist_idx_t array for
off slab caches. But after the commit, freelist_size can be bigger than
KMALLOC_MAX_CACHE_SIZE.
Instead of using pointer to kmalloc cache, use kmalloc_node() and only
check if the kmalloc cache is off slab during calculate_slab_order().
If freelist_size > KMALLOC_MAX_CACHE_SIZE, no looping condition happens
as it allocates freelist_idx_t array directly from buddy.
Link: https://lore.kernel.org/all/20221014205818.GA1428667@roeck-us.net/
Reported-and-tested-by: Guenter Roeck <linux@roeck-us.net>
Fixes: d6a71648db ("mm/slab: kmalloc: pass requests larger than order-1 page to page allocator")
Signed-off-by: Hyeonggon Yoo <42.hyeyoo@gmail.com>
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
