On 21/06/21 11:11 am, Wolfram Sang wrote:

On 21/06/21 10:54 am, Adrian Hunter wrote:

quoted

On 21/06/21 10:32 am, Ulrich Hecht wrote:

quoted

quoted

On 06/21/2021 9:15 AM Adrian Hunter [off-list ref] wrote:
Can we clarify, is the only problem that the error message is confusing?

AFAICT there are no ill effects of the retune failing apart from the error message.

So maybe the simplest thing to do is just amend the message:
e.g.

diff --git a/drivers/mmc/core/core.c b/drivers/mmc/core/core.c
index 4e52eb14198a..5cbf05e331c4 100644
--- a/drivers/mmc/core/core.c
+++ b/drivers/mmc/core/core.c

@@ -936,13 +936,22 @@ int mmc_execute_tuning(struct mmc_card *card)
 		opcode = MMC_SEND_TUNING_BLOCK;
 
 	err = host->ops->execute_tuning(host, opcode);
-
 	if (err)
-		pr_err("%s: tuning execution failed: %d\n",
-			mmc_hostname(host), err);
-	else
-		mmc_retune_enable(host);
+		goto out_err;
+
+	mmc_retune_enable(host);
 
+	return 0;
+
+out_err:
+	if (mmc_card_is_removable(host)) {
+		if (err != -ENOMEDIUM)
+			pr_err("%s: tuning execution failed: %d (this is normal if card removed)\n",
+			       mmc_hostname(host), err);

Hmm, an error message saying "this is normal" doesn't look like a good
option to me. Can't we surpress the message somehow or even avoid tuning
somehow if the card is removed? Sorry, I can't look this up myself right
now, working on another task today.

With the code above, if the host controller knows the card has been
removed, it can return -ENOMEDIUM from ->execute_tuning() to suppress
the message.

Otherwise, you need to introduce a new card state or flag to indicate
that the card may not be present, and use that to suppress the message.

quoted

quoted

+	} else {

+		pr_err("%s: tuning execution failed: %d\n",
+		       mmc_hostname(host), err);
+	}
 	return err;
 }
 

quoted