Thread (23 messages) 23 messages, 2 authors, 2026-03-06

Re: [PATCH 2/7] soc: st: add RISAB dump debug driver

From: Gatien CHEVALLIER <gatien.chevallier@foss.st.com>
Date: 2026-02-10 09:57:46
Also in: linux-arm-kernel, lkml


On 2/10/26 08:50, Krzysztof Kozlowski wrote:
On 09/02/2026 15:59, Gatien Chevallier wrote:
quoted
+
+static int stm32_risab_register_debugfs(struct risab_pdata *pdata)
+{
+	struct dentry *root = NULL;
+
+	root = debugfs_lookup("stm32_firewall", NULL);
+	if (!root)
+		root = debugfs_create_dir("stm32_firewall", NULL);
+
+	if (IS_ERR(root))
+		return PTR_ERR(root);
+
+	pdata->dbg_entry = debugfs_create_file(dev_name(pdata->dev), 0444,
+					       root, pdata, &stm32_risab_conf_dump_fops);
soc drivers should not have any user-space interfaces. This was
requested by Arnd many times in the past, nothing new, including about
debugfs. Otherwise it opens cans of worms, because soc platform
maintainer can dump here whatever the subsystem rejected.

Find suitable subsystem and get its approval for such debugfs entry.

Your commit msg is so vague it is another reason this is not mergeable.
Explains nothing about the feature, nothing about the risk or impact on
the system when user does (while (1); read stm32_risab_conf_dump_fops).

Not mentioning that it does not explain WHY do we want this driver in
the first place...

Best regards,
Krzysztof
My bad, I did not know that and only checked in the directory to see
that there were other debugfs entries in it.

You're right, it's too vague, I will develop the commit message in V2.

Best regards,
Gatien
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help