kernel/linux.git/drivers/dax, branch master

Merge tag 'cxl-for-7.1' of git://git.kernel.org/pub/scm/linux/kernel/git/cxl/cxl

2026-04-17T22:52:58+00:00

Pull CXL (Compute Express Link) updates from Dave Jiang: "The significant change of interest is the handling of soft reserved memory conflict between CXL and HMEM. In essence CXL will be the first to claim the soft reserved memory ranges that belongs to CXL and attempt to enumerate them with best effort. If CXL is not able to enumerate the ranges it will punt them to HMEM. There are also MAINTAINERS email changes from Dan Williams and Jonathan Cameron" * tag 'cxl-for-7.1' of git://git.kernel.org/pub/scm/linux/kernel/git/cxl/cxl: (37 commits) MAINTAINERS: Update Jonathan Cameron's email address cxl/hdm: Add support for 32 switch decoders MAINTAINERS: Update address for Dan Williams tools/testing/cxl: Enable replay of user regions as auto regions cxl/region: Add a region sysfs interface for region lock status tools/testing/cxl: Test dax_hmem takeover of CXL regions tools/testing/cxl: Simulate auto-assembly failure dax/hmem: Parent dax_hmem devices dax/hmem: Fix singleton confusion between dax_hmem_work and hmem devices dax/hmem: Reduce visibility of dax_cxl coordination symbols cxl/region: Constify cxl_region_resource_contains() cxl/region: Limit visibility of cxl_region_contains_resource() dax/cxl: Fix HMEM dependencies cxl/region: Fix use-after-free from auto assembly failure cxl/core: Check existence of cxl_memdev_state in poison test cxl/core: use cleanup.h for devm_cxl_add_dax_region cxl/core/region: move dax region device logic into region_dax.c cxl/core/region: move pmem region driver logic into region_pmem.c dax/hmem, cxl: Defer and resolve Soft Reserved ownership cxl/region: Add helper to check Soft Reserved containment by CXL regions ...

mm: rename VMA flag helpers to be more readable

2026-04-05T20:53:18+00:00

Patch series "mm: vma flag tweaks". The ongoing work around introducing non-system word VMA flags has introduced a number of helper functions and macros to make life easier when working with these flags and to make conversions from the legacy use of VM_xxx flags more straightforward. This series improves these to reduce confusion as to what they do and to improve consistency and readability. Firstly the series renames vma_flags_test() to vma_flags_test_any() to make it abundantly clear that this function tests whether any of the flags are set (as opposed to vma_flags_test_all()). It then renames vma_desc_test_flags() to vma_desc_test_any() for the same reason. Note that we drop the 'flags' suffix here, as vma_desc_test_any_flags() would be cumbersome and 'test' implies a flag test. Similarly, we rename vma_test_all_flags() to vma_test_all() for consistency. Next, we have a couple of instances (erofs, zonefs) where we are now testing for vma_desc_test_any(desc, VMA_SHARED_BIT) && vma_desc_test_any(desc, VMA_MAYWRITE_BIT). This is silly, so this series introduces vma_desc_test_all() so these callers can instead invoke vma_desc_test_all(desc, VMA_SHARED_BIT, VMA_MAYWRITE_BIT). We then observe that quite a few instances of vma_flags_test_any() and vma_desc_test_any() are in fact only testing against a single flag. Using the _any() variant here is just confusing - 'any' of single item reads strangely and is liable to cause confusion. So in these instances the series reintroduces vma_flags_test() and vma_desc_test() as helpers which test against a single flag. The fact that vma_flags_t is a struct and that vma_flag_t utilises sparse to avoid confusion with vm_flags_t makes it impossible for a user to misuse these helpers without it getting flagged somewhere. The series also updates __mk_vma_flags() and functions invoked by it to explicitly mark them always inline to match expectation and to be consistent with other VMA flag helpers. It also renames vma_flag_set() to vma_flags_set_flag() (a function only used by __mk_vma_flags()) to be consistent with other VMA flag helpers. Finally it updates the VMA tests for each of these changes, and introduces explicit tests for vma_flags_test() and vma_desc_test() to assert that they behave as expected. This patch (of 6): On reflection, it's confusing to have vma_flags_test() and vma_desc_test_flags() test whether any comma-separated VMA flag bit is set, while also having vma_flags_test_all() and vma_test_all_flags() separately test whether all flags are set. Firstly, rename vma_flags_test() to vma_flags_test_any() to eliminate this confusion. Secondly, since the VMA descriptor flag functions are becoming rather cumbersome, prefer vma_desc_test*() to vma_desc_test_flags*(), and also rename vma_desc_test_flags() to vma_desc_test_any(). Finally, rename vma_test_all_flags() to vma_test_all() to keep the VMA-specific helper consistent with the VMA descriptor naming convention and to help avoid confusion vs. vma_flags_test_all(). While we're here, also update whitespace to be consistent in helper functions. Link: https://lkml.kernel.org/r/cover.1772704455.git.ljs@kernel.org Link: https://lkml.kernel.org/r/0f9cb3c511c478344fac0b3b3b0300bb95be95e9.1772704455.git.ljs@kernel.org Signed-off-by: Lorenzo Stoakes (Oracle) Suggested-by: Pedro Falcato Acked-by: David Hildenbrand (Arm) Reviewed-by: Pedro Falcato Cc: Arnd Bergmann Cc: Babu Moger Cc: Baolin Wang Cc: Chao Yu Cc: Chatre, Reinette Cc: Chunhai Guo Cc: Damien Le Maol Cc: Dan Williams Cc: Dave Jiang Cc: Dave Martin Cc: Gao Xiang Cc: Greg Kroah-Hartman Cc: Hongbo Li Cc: Hugh Dickins Cc: James Morse Cc: Jan Kara Cc: Jann Horn Cc: Jason Gunthorpe Cc: Jeffle Xu Cc: Johannes Thumshirn Cc: Konstantin Komarov Cc: Liam Howlett Cc: "Luck, Tony" Cc: Matthew Wilcox (Oracle) Cc: Michal Hocko Cc: Mike Rapoport Cc: Muchun Song Cc: Naohiro Aota Cc: Oscar Salvador Cc: Sandeep Dhavale Cc: Suren Baghdasaryan Cc: Vishal Verma Cc: Vlastimil Babka Cc: Yue Hu Signed-off-by: Andrew Morton

dax/hmem: Parent dax_hmem devices

2026-04-01T15:12:18+00:00

For test purposes it is useful to be able to determine which "hmem_platform" device is hosting a given sub-device. Register hmem devices underneath "hmem_platform". Signed-off-by: Dan Williams Reviewed-by: Ira Weiny Reviewed-by: Alison Schofield Reviewed-by: Dave Jiang Link: https://patch.msgid.link/20260327052821.440749-8-dan.j.williams@intel.com Signed-off-by: Dave Jiang

dax/hmem: Fix singleton confusion between dax_hmem_work and hmem devices

2026-04-01T15:12:18+00:00

dax_hmem (ab)uses a platform device to allow for a module to autoload in the presence of "Soft Reserved" resources. The dax_hmem driver had no dependencies on the "hmem_platform" device being a singleton until the recent "dax_hmem vs dax_cxl" takeover solution. Replace the layering violation of dax_hmem_work assuming that there will never be more than one "hmem_platform" device associated with a global work item with a dax_hmem local workqueue that can theoretically support any number of hmem_platform devices. Fixup the reference counting to only pin the device while it is live in the queue. Signed-off-by: Dan Williams Reviewed-by: Ira Weiny Reviewed-by: Alison Schofield Reviewed-by: Dave Jiang Link: https://patch.msgid.link/20260327052821.440749-7-dan.j.williams@intel.com Signed-off-by: Dave Jiang

dax/hmem: Reduce visibility of dax_cxl coordination symbols

2026-04-01T15:12:17+00:00

No other module or use case should be using dax_hmem_initial_probe or dax_hmem_flush_work(). Limit their use to dax_hmem, and dax_cxl respectively. Signed-off-by: Dan Williams Reviewed-by: Ira Weiny Reviewed-by: Alison Schofield Reviewed-by: Dave Jiang Link: https://patch.msgid.link/20260327052821.440749-6-dan.j.williams@intel.com Signed-off-by: Dave Jiang

cxl/region: Constify cxl_region_resource_contains()

2026-04-01T15:12:17+00:00

The call to cxl_region_resource_contains() in hmem_register_cxl_device() need not cast away 'const'. The problem is the usage of the bus_for_each_dev() API which does not mark its @data parameter as 'const'. Switch to bus_find_device() which does take 'const' @data, fixup cxl_region_resource_contains() and its caller. Signed-off-by: Dan Williams Reviewed-by: Ira Weiny Reviewed-by: Alison Schofield Reviewed-by: Dave Jiang Link: https://patch.msgid.link/20260327052821.440749-5-dan.j.williams@intel.com Signed-off-by: Dave Jiang

cxl/region: Limit visibility of cxl_region_contains_resource()

2026-04-01T15:12:17+00:00

The dax_hmem dependency on cxl_region_contains_resource() is a one-off special case. It is not suitable for other use cases. Move the definition to the other CONFIG_CXL_REGION guarded definitions in drivers/cxl/cxl.h and include that by a relative path include. This matches what drivers/dax/cxl.c does for its limited private usage of CXL core symbols. Reduce the symbol export visibility from global to just dax_hmem, to further clarify its applicability. Signed-off-by: Dan Williams Reviewed-by: Ira Weiny Reviewed-by: Alison Schofield Reviewed-by: Dave Jiang Link: https://patch.msgid.link/20260327052821.440749-4-dan.j.williams@intel.com Signed-off-by: Dave Jiang

dax/cxl: Fix HMEM dependencies

2026-04-01T15:12:17+00:00

The expectation is that DEV_DAX_HMEM=y should be disallowed if any of CXL_ACPI, or CXL_PCI are set =m. Also DEV_DAX_CXL=y should be disallowed if DEV_DAX_HMEM=m. Use "$config || !$config" syntax for each dependency. Otherwise, the invalid DEV_DAX_HMEM=m && DEV_DAX_CXL=y configuration is allowed. Lastly, dax_hmem depends on the availability of the cxl_region_contains_resource() symbol published by the cxl_core.ko module. So, also prevent DEV_DAX_HMEM from being built-in when the cxl_core module is not built-in. Signed-off-by: Dan Williams Reviewed-by: Ira Weiny Reviewed-by: Alison Schofield Reviewed-by: Dave Jiang Link: https://patch.msgid.link/20260327052821.440749-3-dan.j.williams@intel.com Signed-off-by: Dave Jiang

dax/hmem, cxl: Defer and resolve Soft Reserved ownership

2026-03-27T17:25:47+00:00

The current probe time ownership check for Soft Reserved memory based solely on CXL window intersection is insufficient. dax_hmem probing is not always guaranteed to run after CXL enumeration and region assembly, which can lead to incorrect ownership decisions before the CXL stack has finished publishing windows and assembling committed regions. Introduce deferred ownership handling for Soft Reserved ranges that intersect CXL windows. When such a range is encountered during the initial dax_hmem probe, schedule deferred work to wait for the CXL stack to complete enumeration and region assembly before deciding ownership. Once the deferred work runs, evaluate each Soft Reserved range individually: if a CXL region fully contains the range, skip it and let dax_cxl bind. Otherwise, register it with dax_hmem. This per-range ownership model avoids the need for CXL region teardown and alloc_dax_region() resource exclusion prevents double claiming. Introduce a boolean flag dax_hmem_initial_probe to live inside device.c so it survives module reload. Ensure dax_cxl defers driver registration until dax_hmem has completed ownership resolution. dax_cxl calls dax_hmem_flush_work() before cxl_driver_register(), which both waits for the deferred work to complete and creates a module symbol dependency that forces dax_hmem.ko to load before dax_cxl. Co-developed-by: Dan Williams Signed-off-by: Smita Koralahalli Reviewed-by: Dave Jiang Reviewed-by: Jonathan Cameron Link: https://patch.msgid.link/20260322195343.206900-9-Smita.KoralahalliChannabasappa@amd.com Signed-off-by: Dan Williams Signed-off-by: Dave Jiang

dax: Track all dax_region allocations under a global resource tree

2026-03-27T17:21:43+00:00

Introduce a global "DAX Regions" resource root and register each dax_region->res under it via request_resource(). Release the resource on dax_region teardown. By enforcing a single global namespace for dax_region allocations, this ensures only one of dax_hmem or dax_cxl can successfully register a dax_region for a given range. Suggested-by: Dan Williams Signed-off-by: Smita Koralahalli Reviewed-by: Jonathan Cameron Reviewed-by: Dave Jiang Link: https://patch.msgid.link/20260322195343.206900-7-Smita.KoralahalliChannabasappa@amd.com Signed-off-by: Dan Williams Signed-off-by: Dave Jiang