In order to prevent mptcpify prog from affecting the running results
of other BPF tests, a pid limit was added to restrict it from only
modifying its own program.
Suggested-by: Martin KaFai Lau <martin.lau@kernel.org>
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/8987e2938e15e8ec390b85b5dcbee704751359dc.1712054986.git.tanggeliang@kylinos.cn
If user request --no-msr or is not able to access the MSRs,
turbostat should clear all the counters added with --add.
Because MSR access permission checks are done after the cmdline is
parsed, the decision has to be defered up until the transition into
no-msr mode happen.
Signed-off-by: Len Brown <len.brown@intel.com>
Checking early if the permissions are even needed gets rid of the
warnings about some of them missing. Earlier we issued a warning in case
of missing MSR and/or perf permissions, even when user never asked for
counters that require those.
Signed-off-by: Patryk Wlazlyn <patryk.wlazlyn@linux.intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
To allow unprivileged user to run turbostat seamlessly.
Signed-off-by: Patryk Wlazlyn <patryk.wlazlyn@linux.intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
By using the perf API we spend less time in between the reads of the
counters, resulting in more accurate calculations of the dependent
metrics.
Using perf API is also usually faster overall, although cache miss, if
we get one, is more costly when using perf vs MSR driver.
We would fallback to the msr reads if the sysfs isn't there or when in
--no-perf mode.
Signed-off-by: Patryk Wlazlyn <patryk.wlazlyn@linux.intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
Add the --no-perf option to allow users to run turbostat without
accessing perf.
Signed-off-by: Patryk Wlazlyn <patryk.wlazlyn@linux.intel.com>
Reviewed-by: Len Brown <len.brown@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
Add --no-msr option to allow users to run turbostat without
accessing MSRs via the MSR driver.
Signed-off-by: Patryk Wlazlyn <patryk.wlazlyn@linux.intel.com>
Reviewed-by: Len Brown <len.brown@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
Eliminate redundant debug output for core and package scope counters.
Include name and path for all "ADDED" counters.
Signed-off-by: Len Brown <len.brown@intel.com>
Previously a failed read of /dev/cpu_dma_latency erroneously complained
turbostat: capget(CAP_SYS_ADMIN) failed, try "# setcap cap_sys_admin=ep ./turbostat
This went unnoticed because this file is typically visible to root,
and turbostat was typically run as root.
Going forward, when a non-root user can run turbostat...
Complain about failed read access to this file only if --debug is used.
Signed-off-by: Len Brown <len.brown@intel.com>
If MSRs cannot be read, values can be obtained from cpuid.
Signed-off-by: Patryk Wlazlyn <patryk.wlazlyn@linux.intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
There are lots of maintainers "pings" during the merge window, even
for trivial patches.
Clarify that contributors should not expect progress on *any*
non-urgent patches during the merge window. This applies to all
contributions, not just large ones.
Clarify the language around -rc1. Trees really are closed during the
merge window.
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Acked-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/all/20240322183403.67BAEEFE%40davehans-spike.ostc.intel.com
Commit 20d59ee551 ("libbpf: add bpf_core_cast() macro") added a
bpf_helpers include in bpf_core_read.h as a system include. Usually, the
includes are local, though, like in bpf_tracing.h. This commit adjusts
the include to be local as well.
Signed-off-by: Tobias Böhm <tobias@aibor.de>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/q5d5bgc6vty2fmaazd5e73efd6f5bhiru2le6fxn43vkw45bls@fhlw2s5ootdb
- Ensure perf events programmed to count during guest execution
are actually enabled before entering the guest in the nVHE
configuration.
- Restore out-of-range handler for stage-2 translation faults.
- Several fixes to stage-2 TLB invalidations to avoid stale
translations, possibly including partial walk caches.
- Fix early handling of architectural VHE-only systems to ensure E2H is
appropriately set.
- Correct a format specifier warning in the arch_timer selftest.
- Make the KVM banner message correctly handle all of the possible
configurations.
-----BEGIN PGP SIGNATURE-----
iI0EABYIADUWIQSNXHjWXuzMZutrKNKivnWIJHzdFgUCZgtpWBccb2xpdmVyLnVw
dG9uQGxpbnV4LmRldgAKCRCivnWIJHzdFoilAQCQk6kLIeuih5QOe50fK4XkNsyg
PGcxw0a0BP8cfjtJsgEArwLlfHQOTE4tRWtXyEHvapJfe/bE1hjLmzUJx7BwLQ4=
=6hNq
-----END PGP SIGNATURE-----
Merge tag 'kvmarm-fixes-6.9-1' of git://git.kernel.org/pub/scm/linux/kernel/git/kvmarm/kvmarm into HEAD
KVM/arm64 fixes for 6.9, part #1
- Ensure perf events programmed to count during guest execution
are actually enabled before entering the guest in the nVHE
configuration.
- Restore out-of-range handler for stage-2 translation faults.
- Several fixes to stage-2 TLB invalidations to avoid stale
translations, possibly including partial walk caches.
- Fix early handling of architectural VHE-only systems to ensure E2H is
appropriately set.
- Correct a format specifier warning in the arch_timer selftest.
- Make the KVM banner message correctly handle all of the possible
configurations.
The node_id for multi-gateway ALH DAI's get overwritten with the
group_id during the DAI copier's ipc_prepare op. So, save the ALH
dai_index during the BE DAI hw_params in the dai_index field of struct
ipc4_copier and use that to set the device ID in the configuration blob.
This will avoid errors during copier init after an xrun.
Note that the dai_index is typically set in topology for DMIC/SSP, but
it's not used for ALH. Reclaiming this dai_index field to store the
node_id does not generate a conflict with topology-defined values.
Reviewed-by: Rander Wang <rander.wang@intel.com>
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://msgid.link/r/20240402151828.175002-18-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
With the delayed stops, the xrun handling is problematic: the
applications expects everything to be reset but the firmware and DMA
are still in a PAUSED state.
This patch makes sure the prepare while pending_stop is set is
special-cased.
Reviewed-by: Rander Wang <rander.wang@intel.com>
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Co-developed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://msgid.link/r/20240402151828.175002-17-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Add a state variable to keep track of delayed stops, in case
pcm_ops->platform_stop_during_hw_free is set.
This patch should be iso-functionality, possibly removing no-op
cases. The main purpose of this new state variable is to prepare a
follow-up patch to reset all PCM and DMAs in case of stop/prepare xrun
sequences.
Reviewed-by: Rander Wang <rander.wang@intel.com>
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://msgid.link/r/20240402151828.175002-16-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The same sequence is used twice, use common helper.
Reviewed-by: Rander Wang <rander.wang@intel.com>
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://msgid.link/r/20240402151828.175002-15-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
For some reason the existing code stops on the first error, which
potentially leaves the DMA and widgets in a weird state.
Change to free-up all resources even in case of errors.
Also add a more consistent error handling and logs, with the first
error code returned to the caller.
Reviewed-by: Rander Wang <rander.wang@intel.com>
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://msgid.link/r/20240402151828.175002-14-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
When an xrun happens, the BE DAI hw_params doesn't get invoked before
the stream restarts with a prepare. In this case, clearing the node ID
when the DAI widget is freed and unprepared will result in an error when
it is re-initialized. In order to avoid this, move the code to clear the
node ID to the BE DAI hw_free op to keep it balanced with the BE DAI
hw_params.
Reviewed-by: Rander Wang <rander.wang@intel.com>
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://msgid.link/r/20240402151828.175002-13-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
To allow using widget_to_sdev() in other files, move it as static inline
in shared header file.
Reviewed-by: Rander Wang <rander.wang@intel.com>
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://msgid.link/r/20240402151828.175002-12-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The node_id value needs to be handled specifically for ALH.
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://msgid.link/r/20240402151828.175002-11-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Add intel_alh_id to set the expected gateway node_id in a follow-up
patch.
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://msgid.link/r/20240402151828.175002-10-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
We've already defined the value for dai_index, let's use it instead of
open-coding the same thing. No functionality change.
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://msgid.link/r/20240402151828.175002-9-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
We need to be able to set the dai config differently for SoundWire.
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://msgid.link/r/20240402151828.175002-8-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
sof_ipc4_dma_config_tlv{} is required for ACE2.x. The patch follow the
convention to set the dma_stream_channel_map.mapping device as
"link_id << 8 | pdi_id".
And the mapping in sof_ipc4_alh_configuration_blob{} should be the same
as dma_stream_channel_map.mapping in sof_ipc4_dma_config{}.
The purposes of device id is to map DMA tlv.
Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://msgid.link/r/20240402151828.175002-7-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
We always use the lowest N channels of stream. So, set ch_mask to
GENMASK(params_channels(params) - 1, 0).
Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://msgid.link/r/20240402151828.175002-6-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Each stream needs a dma_config_tlv. We will handle multi dma_config_tlv
in the follow up commits.
Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://msgid.link/r/20240402151828.175002-5-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
This reverts commit f8ba62ac86.
The SoundWire aggregated solution was to use one DMA on multiple links.
But, the solution changed to use one DMA for each link. It means that
we should assign HDaudio stream_tag for each cpu_dai.
Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://msgid.link/r/20240402151828.175002-4-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
This reverts commit 699e146d9e.
Don't reset device_count as we will use the multi-gateway firmware
configuration.
Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://msgid.link/r/20240402151828.175002-3-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The existing code derives the channel map used to program the HDaudio
link DMA from the hw_params, but that is not quite right in the case
of aggregation. The code in soc-pcm.c splits the hw_params depending
on the codec_ch_map, and we need to reconstruct the channel-map to
insert the data in the right places.
This issue is seen only on amplifier feedback capture where the data
from the second amplifier was replaced by that of the first amplifier.
Note that the loop iterator of the macro for_each_rtd_cpu_dais() is
reused in a following loop. This is different to all existing usages
of that macro, hence the use of a boolean flag to avoid an access to
an uninitialized variable.
Fixes: 2960ee5c48 ("ASoC: SOF: Intel: hda-dai: add helpers for SoundWire callbacks")
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Rander Wang <rander.wang@intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://msgid.link/r/20240402151828.175002-2-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Somehow instance_id and id were mixed. Align on instance_id for
consistency.
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Link: https://msgid.link/r/20240402145959.172619-4-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
cppcheck reports this issue:
sound/soc/sof/amd/acp-loader.c:299:6: style: Variable 'ret' is
reassigned a value before the old one has been
used. [redundantAssignment]
ret = request_firmware(&adata->fw_dbin, fw_filename, sdev->dev);
^
sound/soc/sof/amd/acp-loader.c:289:6: note: ret is assigned
ret = snd_sof_dsp_block_write(sdev, SOF_FW_BLK_TYPE_IRAM, 0,
^
sound/soc/sof/amd/acp-loader.c:299:6: note: ret is overwritten
ret = request_firmware(&adata->fw_dbin, fw_filename, sdev->dev);
^
This behavior is probably unintentional, there's no reason to return
an error for the DRAM but not the IRAM.
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Link: https://msgid.link/r/20240402145959.172619-3-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
For better usability group the display drivers together in Kconfig.
With this we will have the following sections:
- Character LCD
- Samsung KS0108 LCD controller
- Single character line display
- Character LCD with non-conforming interface
Keep Makefile entries sorted alphabetically.
While at it, drop redundant 'default n' entries.
Tested-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Translate dev-tools/ubsan.rst into Chinese, add it into
zh_CN/dev-tools/index.rst.
Signed-off-by: Dongliang Mu <dzm91@hust.edu.cn>
Reviewed-by: Yanteng Si <siyanteng@loongson.cn>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Link: https://lore.kernel.org/r/20240302140058.1527765-1-dzm91@hust.edu.cn
Translate Documentation/process/2.Process.rst into Spanish
Signed-off-by: Avadhut Naik <avadhut.naik@amd.com>
Reviewed-by: Carlos Bilbao <carlos.bilbao@amd.com>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Link: https://lore.kernel.org/r/20240305221839.2764380-5-avadhut.naik@amd.com
Translate Documentation/process/1.Intro.rst into Spanish
In order to avoid broken links in the translated document, empty files
have been created for documents which have not yet been translated.
Signed-off-by: Avadhut Naik <avadhut.naik@amd.com>
Reviewed-by: Carlos Bilbao <carlos.bilbao@amd.com>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Link: https://lore.kernel.org/r/20240305221839.2764380-4-avadhut.naik@amd.com
Translate Documentation/process/development-process.rst into Spanish
Signed-off-by: Avadhut Naik <avadhut.naik@amd.com>
Reviewed-by: Carlos Bilbao <carlos.bilbao@amd.com>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Link: https://lore.kernel.org/r/20240305221839.2764380-3-avadhut.naik@amd.com
Commit 329ac9af90 (docs: submitting-patches: Discuss interleaved replies)
updates the original Documentation/process/submitting-patches.rst file.
Translate and add the updates to its corresponding version in Spanish.
Signed-off-by: Avadhut Naik <avadhut.naik@amd.com>
Reviewed-by: Carlos Bilbao <carlos.bilbao@amd.com>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Link: https://lore.kernel.org/r/20240305221839.2764380-2-avadhut.naik@amd.com
Extend commit 84b4cc8189 ("docs: scripts: sphinx-pre-install: Fix
building docs with pyyaml package") and add pyyaml as an optional
package to Mageia, ArchLinux, and Gentoo.
The Python module pyyaml is required to build the docs, but it is only
listed in Documentation/sphinx/requirements.txt and is therefore missing
when Sphinx is installed as a package and not via pip/pypi.
Signed-off-by: Thorsten Blum <thorsten.blum@toblux.com>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Link: https://lore.kernel.org/r/20240323125837.2022-2-thorsten.blum@toblux.com
The updates from commit ae67ee6c5e ("docs: fix Co-Developed-by
docs") in v5.0 were never applied to the Chinese translations.
In addition:
- "Cc" used wrong case,
- "Co-developed-by" lacked a dash,
- "Signed-off-by" was misspelled.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Hu Haowen <2023002089@link.tyut.edu.cn>
Reviewed-by: Alex Shi <alexs@kernel.org>
Reviewed-by: Yanteng Si <siyanteng@loongson.cn>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Link: https://lore.kernel.org/r/22892a8ab5c17d7121ef5b85f7d18d8b1f41e434.1711715655.git.geert+renesas@glider.be
There are bunch of codes in driver like
if (dma_set_mask_and_coherent(dev, DMA_BIT_MASK(64)))
dma_set_mask_and_coherent(dev, DMA_BIT_MASK(32))
Actually it is wrong because if dma_set_mask_and_coherent(64) fails,
dma_set_mask_and_coherent(32) will fail for the same reason.
And dma_set_mask_and_coherent(64) never returns failure.
According to the definition of dma_set_mask(), it indicates the width of
address that device DMA can access. If it can access 64-bit address, it
must access 32-bit address inherently. So only need set biggest address
width.
See below code fragment:
dma_set_mask(mask)
{
mask = (dma_addr_t)mask;
if (!dev->dma_mask || !dma_supported(dev, mask))
return -EIO;
arch_dma_set_mask(dev, mask);
*dev->dma_mask = mask;
return 0;
}
dma_supported() will call dma_direct_supported or iommux's dma_supported
call back function.
int dma_direct_supported(struct device *dev, u64 mask)
{
u64 min_mask = (max_pfn - 1) << PAGE_SHIFT;
/*
* Because 32-bit DMA masks are so common we expect every architecture
* to be able to satisfy them - either by not supporting more physical
* memory, or by providing a ZONE_DMA32. If neither is the case, the
* architecture needs to use an IOMMU instead of the direct mapping.
*/
if (mask >= DMA_BIT_MASK(32))
return 1;
...
}
The iommux's dma_supported() actually means iommu requires devices's
minimized dma capability.
An example:
static int sba_dma_supported( struct device *dev, u64 mask)()
{
...
* check if mask is >= than the current max IO Virt Address
* The max IO Virt address will *always* < 30 bits.
*/
return((int)(mask >= (ioc->ibase - 1 +
(ioc->pdir_size / sizeof(u64) * IOVP_SIZE) )));
...
}
1 means supported. 0 means unsupported.
Correct document to make it more clear and provide correct sample code.
Signed-off-by: Frank Li <Frank.Li@nxp.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
[jc: fixed then/than typo]
Link: https://lore.kernel.org/r/20240401174159.642998-1-Frank.Li@nxp.com
When kmalloc_node() is called without __GFP_THISNODE and the target node
lacks sufficient memory, SLUB allocates a folio from a different node
other than the requested node, instead of taking a partial slab from it.
However, since the allocated folio does not belong to the requested
node, on the following allocation it is deactivated and added to the
partial slab list of the node it belongs to.
This behavior can result in excessive memory usage when the requested
node has insufficient memory, as SLUB will repeatedly allocate folios
from other nodes without reusing the previously allocated ones.
To prevent memory wastage, when a preferred node is indicated (not
NUMA_NO_NODE) but without a prior __GFP_THISNODE constraint:
1) try to get a partial slab from target node only by having
__GFP_THISNODE in pc.flags for get_partial()
2) if 1) failed, try to allocate a new slab from target node with
GFP_NOWAIT | __GFP_THISNODE opportunistically.
3) if 2) failed, retry with original gfpflags which will allow
get_partial() try partial lists of other nodes before potentially
allocating new page from other nodes
Without a preferred node, or with __GFP_THISNODE constraint, the
behavior remains unchanged.
On qemu with 4 numa nodes and each numa has 1G memory. Write a test ko
to call kmalloc_node(196, GFP_KERNEL, 3) for (4 * 1024 + 4) * 1024 times.
cat /proc/slabinfo shows:
kmalloc-256 4200530 13519712 256 32 2 : tunables..
after this patch,
cat /proc/slabinfo shows:
kmalloc-256 4200558 4200768 256 32 2 : tunables..
Signed-off-by: Chen Jun <chenjun102@huawei.com>
Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Add functions to convert a CString to upper- / lowercase, either
in-place or by creating a copy of the original CString.
Naming follows the one from the Rust stdlib, where functions starting
with 'to' create a copy and functions starting with 'make' perform an
in-place conversion.
This is required by the Nova project (GSP only Rust successor of
Nouveau) to convert stringified enum values (representing different GPU
chipsets) to strings in order to generate the corresponding firmware
paths. See also [1].
Link: https://rust-for-linux.zulipchat.com/#narrow/stream/288089-General/topic/String.20manipulation.20in.20kernel.20Rust [1]
Reviewed-by: Alice Ryhl <aliceryhl@google.com>
Signed-off-by: Danilo Krummrich <dakr@redhat.com>
Reviewed-by: Benno Lossin <benno.lossin@proton.me>
Link: https://lore.kernel.org/r/20240223163726.12397-1-dakr@redhat.com
[ Reworded to fix typo and to make the link use the `Link:` tag. ]
Signed-off-by: Miguel Ojeda <ojeda@kernel.org>