Merge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma

Pull rdma updates from Jason Gunthorpe:
 "Usual collection of driver fixes:

   - Small bug fixes and cleansup in hfi, hns, rxe, mlx5, mana siw

   - Further ODP functionality in rxe

   - Remote access MRs in mana, along with more page sizes

   - Improve CM scalability with a rwlock around the agent

   - More trace points for hns

   - ODP hmm conversion to the new two step dma API

   - Support the ethernet HW device in mana as well as the RNIC

   - Cleanups:
       - Use secs_to_jiffies() when appropriate
       - Use ERR_CAST() instead of naked casts
       - Don't use %pK in printk
       - Unusued functions removed
       - Allocation type matching"

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma: (57 commits)
  RDMA/cma: Fix hang when cma_netevent_callback fails to queue_work
  RDMA/bnxt_re: Support extended stats for Thor2 VF
  RDMA/hns: Fix endian issue in trace events
  RDMA/mlx5: Avoid flexible array warning
  IB/cm: Remove dead code and adjust naming
  RDMA/core: Avoid hmm_dma_map_alloc() for virtual DMA devices
  RDMA/rxe: Break endless pagefault loop for RO pages
  RDMA/bnxt_re: Fix return code of bnxt_re_configure_cc
  RDMA/bnxt_re: Fix missing error handling for tx_queue
  RDMA/bnxt_re: Fix incorrect display of inactivity_cp in debugfs output
  RDMA/mlx5: Add support for 200Gbps per lane speeds
  RDMA/mlx5: Remove the redundant MLX5_IB_STAGE_UAR stage
  RDMA/iwcm: Fix use-after-free of work objects after cm_id destruction
  net: mana: Add support for auxiliary device servicing events
  RDMA/mana_ib: unify mana_ib functions to support any gdma device
  RDMA/mana_ib: Add support of mana_ib for RNIC and ETH nic
  net: mana: Probe rdma device in mana driver
  RDMA/siw: replace redundant ternary operator with just rv
  RDMA/umem: Separate implicit ODP initialization from explicit ODP
  RDMA/core: Convert UMEM ODP DMA mapping to caching IOVA and page linkage
  ...

3 files changed, 56 insertions, 2 deletions

diff --git a/include/linux/hmm-dma.h b/include/linux/hmm-dma.h
new file mode 100644
index 000000000000..f58b9fc71999
--- /dev/null
+++ b/include/linux/hmm-dma.h
@@ -0,0 +1,33 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/* Copyright (c) 2024 NVIDIA Corporation & Affiliates */
+#ifndef LINUX_HMM_DMA_H
+#define LINUX_HMM_DMA_H
+
+#include <linux/dma-mapping.h>
+
+struct dma_iova_state;
+struct pci_p2pdma_map_state;
+
+/*
+ * struct hmm_dma_map - array of PFNs and DMA addresses
+ *
+ * @state: DMA IOVA state
+ * @pfns: array of PFNs
+ * @dma_list: array of DMA addresses
+ * @dma_entry_size: size of each DMA entry in the array
+ */
+struct hmm_dma_map {
+	struct dma_iova_state state;
+	unsigned long *pfn_list;
+	dma_addr_t *dma_list;
+	size_t dma_entry_size;
+};
+
+int hmm_dma_map_alloc(struct device *dev, struct hmm_dma_map *map,
+		      size_t nr_entries, size_t dma_entry_size);
+void hmm_dma_map_free(struct device *dev, struct hmm_dma_map *map);
+dma_addr_t hmm_dma_map_pfn(struct device *dev, struct hmm_dma_map *map,
+			   size_t idx,
+			   struct pci_p2pdma_map_state *p2pdma_state);
+bool hmm_dma_unmap_pfn(struct device *dev, struct hmm_dma_map *map, size_t idx);
+#endif /* LINUX_HMM_DMA_H */
diff --git a/include/linux/hmm.h b/include/linux/hmm.h
index 126a36571667..db75ffc949a7 100644
--- a/include/linux/hmm.h
+++ b/include/linux/hmm.h
@@ -23,6 +23,10 @@ struct mmu_interval_notifier;
  * HMM_PFN_WRITE - if the page memory can be written to (requires HMM_PFN_VALID)
  * HMM_PFN_ERROR - accessing the pfn is impossible and the device should
  *                 fail. ie poisoned memory, special pages, no vma, etc
+ * HMM_PFN_P2PDMA - P2P page
+ * HMM_PFN_P2PDMA_BUS - Bus mapped P2P transfer
+ * HMM_PFN_DMA_MAPPED - Flag preserved on input-to-output transformation
+ *                      to mark that page is already DMA mapped
  *
  * On input:
  * 0                 - Return the current state of the page, do not fault it.
@@ -36,13 +40,21 @@ enum hmm_pfn_flags {
 	HMM_PFN_VALID = 1UL << (BITS_PER_LONG - 1),
 	HMM_PFN_WRITE = 1UL << (BITS_PER_LONG - 2),
 	HMM_PFN_ERROR = 1UL << (BITS_PER_LONG - 3),
-	HMM_PFN_ORDER_SHIFT = (BITS_PER_LONG - 8),
+	/*
+	 * Sticky flags, carried from input to output,
+	 * don't forget to update HMM_PFN_INOUT_FLAGS
+	 */
+	HMM_PFN_DMA_MAPPED = 1UL << (BITS_PER_LONG - 4),
+	HMM_PFN_P2PDMA     = 1UL << (BITS_PER_LONG - 5),
+	HMM_PFN_P2PDMA_BUS = 1UL << (BITS_PER_LONG - 6),
+
+	HMM_PFN_ORDER_SHIFT = (BITS_PER_LONG - 11),
 
 	/* Input flags */
 	HMM_PFN_REQ_FAULT = HMM_PFN_VALID,
 	HMM_PFN_REQ_WRITE = HMM_PFN_WRITE,
 
-	HMM_PFN_FLAGS = 0xFFUL << HMM_PFN_ORDER_SHIFT,
+	HMM_PFN_FLAGS = ~((1UL << HMM_PFN_ORDER_SHIFT) - 1),
 };
 
 /*
@@ -58,6 +70,14 @@ static inline struct page *hmm_pfn_to_page(unsigned long hmm_pfn)
 }
 
 /*
+ * hmm_pfn_to_phys() - return physical address pointed to by a device entry
+ */
+static inline phys_addr_t hmm_pfn_to_phys(unsigned long hmm_pfn)
+{
+	return __pfn_to_phys(hmm_pfn & ~HMM_PFN_FLAGS);
+}
+
+/*
  * hmm_pfn_to_map_order() - return the CPU mapping size order
  *
  * This is optionally useful to optimize processing of the pfn result
diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index d1dfbad9a447..e6ba8f4f4bd1 100644
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -398,6 +398,7 @@ struct mlx5_core_rsc_common {
 	enum mlx5_res_type	res;
 	refcount_t		refcount;
 	struct completion	free;
+	bool			invalid;
 };
 
 struct mlx5_uars_page {