From: Waiman Long Subject: Re: [PATCH v5 3/4] mm/page_owner: Print memcg information Date: Tue, 8 Feb 2022 13:40:57 -0500 Message-ID: References: <20220208000532.1054311-1-longman@redhat.com> <20220208000532.1054311-4-longman@redhat.com> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Return-path: DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1644345674; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=xML1xNkIw3e0hTYk74bi0fSg0bcnbMZm5u6L4MPv8BA=; b=bck0yfKicJmKETmL+MCTzBS/qgOqMz1tJSizTsus6ltfq9Cp0UGNPIQn/SEu3tSNynlTHH onTPxwk711oQoqxDMFJJzbycFeZldhZVO8NqtHgGeXsREglZo8IUT2ldqiNSjW1ZZzW1mf WlZHGyi6MUlWPl3tfygVWYmD8iZ8wLk= Content-Language: en-US In-Reply-To: List-ID: Content-Type: text/plain; charset="us-ascii"; format="flowed" To: Michal Hocko Cc: Johannes Weiner , Vladimir Davydov , Andrew Morton , Petr Mladek , Steven Rostedt , Sergey Senozhatsky , Andy Shevchenko , Rasmus Villemoes , linux-kernel@vger.kernel.org, cgroups@vger.kernel.org, linux-mm@kvack.org, Ira Weiny , Mike Rapoport , David Rientjes , Roman Gushchin , Rafael Aquini , Mike Rapoport On 2/8/22 07:13, Michal Hocko wrote: > On Mon 07-02-22 19:05:31, Waiman Long wrote: >> It was found that a number of dying memcgs were not freed because >> they were pinned by some charged pages that were present. Even "echo 1 > >> /proc/sys/vm/drop_caches" wasn't able to free those pages. These dying >> but not freed memcgs tend to increase in number over time with the side >> effect that percpu memory consumption as shown in /proc/meminfo also >> increases over time. > I still believe that this is very suboptimal way to debug offline memcgs > but memcg information can be useful in other contexts and it doesn't > cost us anything except for an additional output so I am fine with this. I am planning to have a follow-up patch to add a new debugfs file for just printing page information associated with dying memcgs only. It will be based on the existing page_owner code, though. So I need to get this patch in first. > >> In order to find out more information about those pages that pin >> dying memcgs, the page_owner feature is extended to print memory >> cgroup information especially whether the cgroup is dying or not. >> RCU read lock is taken when memcg is being accessed to make sure >> that it won't be freed. >> >> Signed-off-by: Waiman Long >> Acked-by: David Rientjes >> Acked-by: Roman Gushchin >> Acked-by: Mike Rapoport > With few comments/questions below. > >> --- >> mm/page_owner.c | 44 ++++++++++++++++++++++++++++++++++++++++++++ >> 1 file changed, 44 insertions(+) >> >> diff --git a/mm/page_owner.c b/mm/page_owner.c >> index 28dac73e0542..d4c311455753 100644 >> --- a/mm/page_owner.c >> +++ b/mm/page_owner.c >> @@ -10,6 +10,7 @@ >> #include >> #include >> #include >> +#include >> #include >> >> #include "internal.h" >> @@ -325,6 +326,47 @@ void pagetypeinfo_showmixedcount_print(struct seq_file *m, >> seq_putc(m, '\n'); >> } >> >> +/* >> + * Looking for memcg information and print it out >> + */ > I am not sure this is particularly useful comment. Right, I can remove that. > >> +static inline int print_page_owner_memcg(char *kbuf, size_t count, int ret, >> + struct page *page) >> +{ >> +#ifdef CONFIG_MEMCG >> + unsigned long memcg_data; >> + struct mem_cgroup *memcg; >> + bool dying; >> + >> + rcu_read_lock(); >> + memcg_data = READ_ONCE(page->memcg_data); >> + if (!memcg_data) >> + goto out_unlock; >> + >> + if (memcg_data & MEMCG_DATA_OBJCGS) >> + ret += scnprintf(kbuf + ret, count - ret, >> + "Slab cache page\n"); >> + >> + memcg = page_memcg_check(page); >> + if (!memcg) >> + goto out_unlock; >> + >> + dying = (memcg->css.flags & CSS_DYING); > Is there any specific reason why you haven't used mem_cgroup_online? Not really. However, I think checking for CSS_DYING makes more sense now that I using the term "dying". > >> + ret += scnprintf(kbuf + ret, count - ret, >> + "Charged %sto %smemcg ", >> + PageMemcgKmem(page) ? "(via objcg) " : "", >> + dying ? "dying " : ""); >> + >> + /* Write cgroup name directly into kbuf */ >> + cgroup_name(memcg->css.cgroup, kbuf + ret, count - ret); >> + ret += strlen(kbuf + ret); > cgroup_name should return the length of the path added to the buffer. I realized that after I sent out the patch. I will remove te redundant strlen() in a future update. > >> + ret += scnprintf(kbuf + ret, count - ret, "\n"); > I do not see any overflow prevention here. I believe you really need to > check ret >= count after each scnprintf/cgroup_name. As you have realized, the beauty of using scnprintf() is to not needing an overflow check after each invocation. Cheers, Longman