On Fri 30-08-19 08:49:21, Tejun Heo wrote:

Hello, Jan.

On Fri, Aug 30, 2019 at 05:40:23PM +0200, Jan Kara wrote:

quoted

quoted

+	TP_fast_assign(
+		strncpy(__entry->name,	dev_name(wb->bdi->dev), 32);
+		__entry->bdi_id		= wb->bdi->id;
+		__entry->ino		= page->mapping->host->i_ino;
+		__entry->memcg_id	= wb->memcg_css->id;
+		__entry->cgroup_ino	= __trace_wb_assign_cgroup(wb);
+		__entry->page_cgroup_ino = page->mem_cgroup->css.cgroup->kn->id.ino;
+	),

Are the page dereferences above safe? I suppose lock_page_memcg() protects
the page->mem_cgroup->css.cgroup->kn->id dereference? But page->mapping
does not seem to be protected by page lock?

Hah, I assumed it would work because there are preceding if
(page_mapping()) tests in the dirty paths -
e.g. __set_page_dirty_nobuffers().  Oh, regardless of that assumption,
I should have used page_mapping().

Well, but if you look at __set_page_dirty_nobuffers() it is careful. It
does:

struct address_space *mapping = page_mapping(page);

if (!mapping) {
	bail
}
... use mapping

Exactly because page->mapping can become NULL under your hands if you don't
hold page lock. So I think you either need something similar in your
tracepoint or handle this in the caller.

								Honza
-- 
Jan Kara [off-list ref]
SUSE Labs, CR