IB/hfi1: Fix wrong mmu_node used for user SDMA packet after invalidate

The hfi1 user SDMA pinned-page cache will leave a stale cache entry when
the cache-entry's virtual address range is invalidated but that cache
entry is in-use by an outstanding SDMA request.

Subsequent user SDMA requests with buffers in or spanning the virtual
address range of the stale cache entry will result in packets constructed
from the wrong memory, the physical pages pointed to by the stale cache
entry.

To fix this, remove mmu_rb_node cache entries from the mmu_rb_handler
cache independent of the cache entry's refcount. Add 'struct kref
refcount' to struct mmu_rb_node and manage mmu_rb_node lifetime with
kref_get() and kref_put().

mmu_rb_node.refcount makes sdma_mmu_node.refcount redundant. Remove
'atomic_t refcount' from struct sdma_mmu_node and change sdma_mmu_node
code to use mmu_rb_node.refcount.

Move the mmu_rb_handler destructor call after a
wait-for-SDMA-request-completion call so mmu_rb_nodes that need
mmu_rb_handler's workqueue to queue themselves up for destruction from an
interrupt context may do so.

Fixes: f48ad614c100 ("IB/hfi1: Move driver out of staging")
Fixes: 00cbce5cbf88 ("IB/hfi1: Fix bugs with non-PAGE_SIZE-end multi-iovec user SDMA requests")
Link: https://lore.kernel.org/r/168451393605.3700681.13493776139032178861.stgit@awfm-02.cornelisnetworks.com
Reviewed-by: Dean Luick <dean.luick@cornelisnetworks.com>
Signed-off-by: Brendan Cunningham <bcunningham@cornelisnetworks.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@cornelisnetworks.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

1 files changed, 32 insertions, 15 deletions

diff --git a/drivers/infiniband/hw/hfi1/sdma.h b/drivers/infiniband/hw/hfi1/sdma.h
index 95aaec14c6c2..7fdebab202c4 100644
--- a/drivers/infiniband/hw/hfi1/sdma.h
+++ b/drivers/infiniband/hw/hfi1/sdma.h
@@ -594,9 +594,11 @@ static inline dma_addr_t sdma_mapping_addr(struct sdma_desc *d)
 static inline void make_tx_sdma_desc(
 	struct sdma_txreq *tx,
 	int type,
-	void *pinning_ctx,
 	dma_addr_t addr,
-	size_t len)
+	size_t len,
+	void *pinning_ctx,
+	void (*ctx_get)(void *),
+	void (*ctx_put)(void *))
 {
 	struct sdma_desc *desc = &tx->descp[tx->num_desc];
 
@@ -613,7 +615,11 @@ static inline void make_tx_sdma_desc(
 				<< SDMA_DESC0_PHY_ADDR_SHIFT) |
 			(((u64)len & SDMA_DESC0_BYTE_COUNT_MASK)
 				<< SDMA_DESC0_BYTE_COUNT_SHIFT);
+
 	desc->pinning_ctx = pinning_ctx;
+	desc->ctx_put = ctx_put;
+	if (pinning_ctx && ctx_get)
+		ctx_get(pinning_ctx);
 }
 
 /* helper to extend txreq */
@@ -645,18 +651,20 @@ static inline void _sdma_close_tx(struct hfi1_devdata *dd,
 static inline int _sdma_txadd_daddr(
 	struct hfi1_devdata *dd,
 	int type,
-	void *pinning_ctx,
 	struct sdma_txreq *tx,
 	dma_addr_t addr,
-	u16 len)
+	u16 len,
+	void *pinning_ctx,
+	void (*ctx_get)(void *),
+	void (*ctx_put)(void *))
 {
 	int rval = 0;
 
 	make_tx_sdma_desc(
 		tx,
 		type,
-		pinning_ctx,
-		addr, len);
+		addr, len,
+		pinning_ctx, ctx_get, ctx_put);
 	WARN_ON(len > tx->tlen);
 	tx->num_desc++;
 	tx->tlen -= len;
@@ -676,11 +684,18 @@ static inline int _sdma_txadd_daddr(
 /**
  * sdma_txadd_page() - add a page to the sdma_txreq
  * @dd: the device to use for mapping
- * @pinning_ctx: context to be released at descriptor retirement
  * @tx: tx request to which the page is added
  * @page: page to map
  * @offset: offset within the page
  * @len: length in bytes
+ * @pinning_ctx: context to be stored on struct sdma_desc .pinning_ctx. Not
+ *               added if coalesce buffer is used. E.g. pointer to pinned-page
+ *               cache entry for the sdma_desc.
+ * @ctx_get: optional function to take reference to @pinning_ctx. Not called if
+ *           @pinning_ctx is NULL.
+ * @ctx_put: optional function to release reference to @pinning_ctx after
+ *           sdma_desc completes. May be called in interrupt context so must
+ *           not sleep. Not called if @pinning_ctx is NULL.
  *
  * This is used to add a page/offset/length descriptor.
  *
@@ -692,11 +707,13 @@ static inline int _sdma_txadd_daddr(
  */
 static inline int sdma_txadd_page(
 	struct hfi1_devdata *dd,
-	void *pinning_ctx,
 	struct sdma_txreq *tx,
 	struct page *page,
 	unsigned long offset,
-	u16 len)
+	u16 len,
+	void *pinning_ctx,
+	void (*ctx_get)(void *),
+	void (*ctx_put)(void *))
 {
 	dma_addr_t addr;
 	int rval;
@@ -720,7 +737,8 @@ static inline int sdma_txadd_page(
 		return -ENOSPC;
 	}
 
-	return _sdma_txadd_daddr(dd, SDMA_MAP_PAGE, pinning_ctx, tx, addr, len);
+	return _sdma_txadd_daddr(dd, SDMA_MAP_PAGE, tx, addr, len,
+				 pinning_ctx, ctx_get, ctx_put);
 }
 
 /**
@@ -754,8 +772,8 @@ static inline int sdma_txadd_daddr(
 			return rval;
 	}
 
-	return _sdma_txadd_daddr(dd, SDMA_MAP_NONE, NULL, tx,
-				 addr, len);
+	return _sdma_txadd_daddr(dd, SDMA_MAP_NONE, tx, addr, len,
+				 NULL, NULL, NULL);
 }
 
 /**
@@ -801,7 +819,8 @@ static inline int sdma_txadd_kvaddr(
 		return -ENOSPC;
 	}
 
-	return _sdma_txadd_daddr(dd, SDMA_MAP_SINGLE, NULL, tx, addr, len);
+	return _sdma_txadd_daddr(dd, SDMA_MAP_SINGLE, tx, addr, len,
+				 NULL, NULL, NULL);
 }
 
 struct iowait_work;
@@ -1034,6 +1053,4 @@ u16 sdma_get_descq_cnt(void);
 extern uint mod_num_sdma;
 
 void sdma_update_lmc(struct hfi1_devdata *dd, u64 mask, u32 lid);
-
-void system_descriptor_complete(struct hfi1_devdata *dd, struct sdma_desc *descp);
 #endif