habanalabs: revise prints on FD close
authorOded Gabbay <ogabbay@kernel.org>
Tue, 13 Jul 2021 05:11:54 +0000 (08:11 +0300)
committerOded Gabbay <ogabbay@kernel.org>
Sun, 29 Aug 2021 06:47:48 +0000 (09:47 +0300)
The driver quietly handles memory mappings that were not freed so no
need to print a warning about that when user closes the FD.

Accordingly, revise the text that is printed in case the device is
still in use after the user process closed the FD.

Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
drivers/misc/habanalabs/common/device.c
drivers/misc/habanalabs/common/memory.c

index b09bdd1d462f5b3158d0b2f19028860c52facc88..4f76c7a51605890442faeb585b902b4e78c94433 100644 (file)
@@ -129,8 +129,8 @@ static int hl_device_release(struct inode *inode, struct file *filp)
        hl_ctx_mgr_fini(hdev, &hpriv->ctx_mgr);
 
        if (!hl_hpriv_put(hpriv))
-               dev_warn(hdev->dev,
-                       "Device is still in use because there are live CS and/or memory mappings\n");
+               dev_notice(hdev->dev,
+                       "User process closed FD but device still in use\n");
 
        hdev->last_open_session_duration_jif =
                jiffies - hdev->last_successful_open_jif;
index a05d98db485762914ef0b90d84bdb94365ad30a5..efc460e9db5d2f2a858094cbb4496b54d5227636 100644 (file)
@@ -2126,7 +2126,7 @@ void hl_vm_ctx_fini(struct hl_ctx *ctx)
         * another side effect error
         */
        if (!hdev->hard_reset_pending && !hash_empty(ctx->mem_hash))
-               dev_notice(hdev->dev,
+               dev_dbg(hdev->dev,
                        "user released device without removing its memory mappings\n");
 
        hash_for_each_safe(ctx->mem_hash, i, tmp_node, hnode, node) {