Hello Daniel,

On 5/5/22 15:02, Daniel Vetter wrote:

[snip]

quoted

 static void vesafb_destroy(struct fb_info *info)
 {
 	struct vesafb_par *par = info->par;

@@ -187,7 +191,13 @@ static void vesafb_destroy(struct fb_info *info)
 	arch_phys_wc_del(par->wc_cookie);
 	if (info->screen_base)
 		iounmap(info->screen_base);
+
+	if (((struct vesafb_par *)(info->par))->region)
+		release_region(0x3c0, 32);

This move seems rather iffy, so maybe justify it with "makes the code
exactly as busted before 27599aacbaef ("fbdev: Hot-unplug firmware fb
devices on forced removal")"

I think that will just drop this change. While being here I wanted the release
order to be the inverse of the order in which the driver acquires them. But I
will only move the framebuffer_release() that is the problematic bit.

Someone if care enough could fix the rest of the driver.
 

Also same comments as on v1 about adding more details about what/how this
fixes, with that: Reviewed-by: Daniel Vetter [off-list ref]

Yes, I'll do that too. Thanks again for your comments and feedback.

-- 
Best regards,

Javier Martinez Canillas
Linux Engineering
Red Hat