[v3,6/9] gpudev: add memory barrier

Message ID 20211009015349.9694-7-eagostini@nvidia.com (mailing list archive)
State Superseded, archived
Delegated to: Thomas Monjalon
Headers
Series GPU library |

Checks

Context Check Description
ci/checkpatch success coding style OK

Commit Message

Elena Agostini Oct. 9, 2021, 1:53 a.m. UTC
  From: Elena Agostini <eagostini@nvidia.com>

Add a function for the application to ensure the coherency
of the writes executed by another device into the GPU memory.

Signed-off-by: Elena Agostini <eagostini@nvidia.com>
---
 doc/guides/prog_guide/gpudev.rst |  8 ++++++++
 lib/gpudev/gpudev.c              | 19 +++++++++++++++++++
 lib/gpudev/gpudev_driver.h       |  3 +++
 lib/gpudev/rte_gpudev.h          | 18 ++++++++++++++++++
 lib/gpudev/version.map           |  1 +
 5 files changed, 49 insertions(+)
  

Comments

Thomas Monjalon Oct. 8, 2021, 8:16 p.m. UTC | #1
09/10/2021 03:53, eagostini@nvidia.com:
> From: Elena Agostini <eagostini@nvidia.com>
> 
> Add a function for the application to ensure the coherency
> of the writes executed by another device into the GPU memory.
> 
> Signed-off-by: Elena Agostini <eagostini@nvidia.com>
> ---
> +/**
> + * @warning
> + * @b EXPERIMENTAL: this API may change without prior notice.
> + *
> + * Enforce a GPU memory write barrier.
> + *
> + * @param dev_id
> + *   Reference device ID.
> + *
> + * @return
> + *   0 on success, -rte_errno otherwise:
> + *   - ENODEV if invalid dev_id
> + *   - ENOTSUP if operation not supported by the driver
> + *   - EPERM if driver error
> + */
> +__rte_experimental
> +int rte_gpu_mbw(int16_t dev_id);

I would replace mbw with wmb.

Also it may be worth adding few more words about the goal:
ensure that previous writes in GPU memory are complete?
Does it work for writes done from CPU? from GPU?
  

Patch

diff --git a/doc/guides/prog_guide/gpudev.rst b/doc/guides/prog_guide/gpudev.rst
index 9aca69038c..eb5f0af817 100644
--- a/doc/guides/prog_guide/gpudev.rst
+++ b/doc/guides/prog_guide/gpudev.rst
@@ -65,3 +65,11 @@  gpudev can register a CPU memory area to make it visible from a GPU device.
 Later, it's also possible to unregister that memory with gpudev.
 CPU memory registered outside of the gpudev library
 (e.g. with GPU specific library) cannot be unregistered by the gpudev library.
+
+Memory Barrier
+~~~~~~~~~~~~~~
+
+Some GPU drivers may need, under certain conditions,
+to enforce the coherency of external devices writes (e.g. NIC receiving packets)
+into the GPU memory.
+gpudev abstracts and exposes this capability.
diff --git a/lib/gpudev/gpudev.c b/lib/gpudev/gpudev.c
index 1d8318f769..cefefd737a 100644
--- a/lib/gpudev/gpudev.c
+++ b/lib/gpudev/gpudev.c
@@ -624,3 +624,22 @@  rte_gpu_free(int16_t dev_id, void *ptr)
 	}
 	return GPU_DRV_RET(dev->ops.mem_free(dev, ptr));
 }
+
+int
+rte_gpu_mbw(int16_t dev_id)
+{
+	struct rte_gpu *dev;
+
+	dev = gpu_get_by_id(dev_id);
+	if (dev == NULL) {
+		GPU_LOG(ERR, "memory barrier for invalid device ID %d", dev_id);
+		rte_errno = ENODEV;
+		return -rte_errno;
+	}
+
+	if (dev->ops.mbw == NULL) {
+		rte_errno = ENOTSUP;
+		return -rte_errno;
+	}
+	return GPU_DRV_RET(dev->ops.mbw(dev));
+}
diff --git a/lib/gpudev/gpudev_driver.h b/lib/gpudev/gpudev_driver.h
index 11015944a6..ab24de9e28 100644
--- a/lib/gpudev/gpudev_driver.h
+++ b/lib/gpudev/gpudev_driver.h
@@ -31,6 +31,7 @@  typedef int (rte_gpu_mem_alloc_t)(struct rte_gpu *dev, size_t size, void **ptr);
 typedef int (rte_gpu_free_t)(struct rte_gpu *dev, void *ptr);
 typedef int (rte_gpu_mem_register_t)(struct rte_gpu *dev, size_t size, void *ptr);
 typedef int (rte_gpu_mem_unregister_t)(struct rte_gpu *dev, void *ptr);
+typedef int (rte_gpu_mbw_t)(struct rte_gpu *dev);
 
 struct rte_gpu_ops {
 	/* Get device info. If NULL, info is just copied. */
@@ -45,6 +46,8 @@  struct rte_gpu_ops {
 	rte_gpu_free_t *mem_free;
 	/* Unregister CPU memory in device. */
 	rte_gpu_mem_unregister_t *mem_unregister;
+	/* Enforce GPU memory write barrier. */
+	rte_gpu_mbw_t *mbw;
 };
 
 struct rte_gpu_mpshared {
diff --git a/lib/gpudev/rte_gpudev.h b/lib/gpudev/rte_gpudev.h
index 3c276581c0..e790b3e2b7 100644
--- a/lib/gpudev/rte_gpudev.h
+++ b/lib/gpudev/rte_gpudev.h
@@ -387,6 +387,24 @@  int rte_gpu_register(int16_t dev_id, size_t size, void * ptr);
 __rte_experimental
 int rte_gpu_unregister(int16_t dev_id, void *ptr);
 
+/**
+ * @warning
+ * @b EXPERIMENTAL: this API may change without prior notice.
+ *
+ * Enforce a GPU memory write barrier.
+ *
+ * @param dev_id
+ *   Reference device ID.
+ *
+ * @return
+ *   0 on success, -rte_errno otherwise:
+ *   - ENODEV if invalid dev_id
+ *   - ENOTSUP if operation not supported by the driver
+ *   - EPERM if driver error
+ */
+__rte_experimental
+int rte_gpu_mbw(int16_t dev_id);
+
 #ifdef __cplusplus
 }
 #endif
diff --git a/lib/gpudev/version.map b/lib/gpudev/version.map
index d4a65ebd52..d72d470d8e 100644
--- a/lib/gpudev/version.map
+++ b/lib/gpudev/version.map
@@ -13,6 +13,7 @@  EXPERIMENTAL {
 	rte_gpu_init;
 	rte_gpu_is_valid;
 	rte_gpu_malloc;
+	rte_gpu_mbw;
 	rte_gpu_register;
 	rte_gpu_unregister;
 };