| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <3b792413-184b-20b1-9d90-9e69f0df8cc4@suse.cz>
Date: Tue, 7 Aug 2018 15:18:31 +0200
From: Vlastimil Babka <vbabka@...e.cz>
To: Dennis Zhou <dennisszhou@...il.com>,
Andrew Morton <akpm@...ux-foundation.org>,
Tejun Heo <tj@...nel.org>,
Johannes Weiner <hannes@...xchg.org>,
Christoph Lameter <cl@...ux.com>, Roman Gushchin <guro@...com>
Cc: kernel-team@...com, linux-mm@...ck.org,
linux-kernel@...r.kernel.org, Linux API <linux-api@...r.kernel.org>
Subject: Re: [PATCH] proc: add percpu populated pages count to meminfo
+CC linux-api
On 08/07/2018 02:56 AM, Dennis Zhou wrote:
> From: "Dennis Zhou (Facebook)" <dennisszhou@...il.com>
>
> Currently, percpu memory only exposes allocation and utilization
> information via debugfs. This more or less is only really useful for
> understanding the fragmentation and allocation information at a
> per-chunk level with a few global counters. This is also gated behind a
> config. BPF and cgroup, for example, have seen an increase use causing
> increased use of percpu memory. Let's make it easier for someone to
> identify how much memory is being used.
>
> This patch adds the PercpuPopulated stat to meminfo to more easily
> look up how much percpu memory is in use. This new number includes the
> cost for all backing pages and not just insight at the a unit, per
> chunk level. This stat includes only pages used to back the chunks
> themselves excluding metadata. I think excluding metadata is fair
> because the backing memory scales with the number of cpus and can
> quickly outweigh the metadata. It also makes this calculation light.
>
> Signed-off-by: Dennis Zhou <dennisszhou@...il.com>
Sounds useful and cheap, so why not, I guess.
> ---
> fs/proc/meminfo.c | 2 ++
> include/linux/percpu.h | 2 ++
> mm/percpu.c | 29 +++++++++++++++++++++++++++++
Documentation/filesystems/proc.txt should be updated as well
> 3 files changed, 33 insertions(+)
>
> diff --git a/fs/proc/meminfo.c b/fs/proc/meminfo.c
> index 2fb04846ed11..ddd5249692e9 100644
> --- a/fs/proc/meminfo.c
> +++ b/fs/proc/meminfo.c
> @@ -7,6 +7,7 @@
> #include <linux/mman.h>
> #include <linux/mmzone.h>
> #include <linux/proc_fs.h>
> +#include <linux/percpu.h>
> #include <linux/quicklist.h>
> #include <linux/seq_file.h>
> #include <linux/swap.h>
> @@ -121,6 +122,7 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
> (unsigned long)VMALLOC_TOTAL >> 10);
> show_val_kb(m, "VmallocUsed: ", 0ul);
> show_val_kb(m, "VmallocChunk: ", 0ul);
> + show_val_kb(m, "PercpuPopulated:", pcpu_nr_populated_pages());
>
> #ifdef CONFIG_MEMORY_FAILURE
> seq_printf(m, "HardwareCorrupted: %5lu kB\n",
> diff --git a/include/linux/percpu.h b/include/linux/percpu.h
> index 296bbe49d5d1..1c80be42822c 100644
> --- a/include/linux/percpu.h
> +++ b/include/linux/percpu.h
> @@ -149,4 +149,6 @@ extern phys_addr_t per_cpu_ptr_to_phys(void *addr);
> (typeof(type) __percpu *)__alloc_percpu(sizeof(type), \
> __alignof__(type))
>
> +extern int pcpu_nr_populated_pages(void);
> +
> #endif /* __LINUX_PERCPU_H */
> diff --git a/mm/percpu.c b/mm/percpu.c
> index 0b6480979ac7..08a4341f30c5 100644
> --- a/mm/percpu.c
> +++ b/mm/percpu.c
> @@ -169,6 +169,14 @@ static LIST_HEAD(pcpu_map_extend_chunks);
> */
> int pcpu_nr_empty_pop_pages;
>
> +/*
> + * The number of populated pages in use by the allocator, protected by
> + * pcpu_lock. This number is kept per a unit per chunk (i.e. when a page gets
> + * allocated/deallocated, it is allocated/deallocated in all units of a chunk
> + * and increments/decrements this count by 1).
> + */
> +static int pcpu_nr_populated;
It better be unsigned long, to match others.
> +
> /*
> * Balance work is used to populate or destroy chunks asynchronously. We
> * try to keep the number of populated free pages between
> @@ -1232,6 +1240,7 @@ static void pcpu_chunk_populated(struct pcpu_chunk *chunk, int page_start,
>
> bitmap_set(chunk->populated, page_start, nr);
> chunk->nr_populated += nr;
> + pcpu_nr_populated += nr;
>
> if (!for_alloc) {
> chunk->nr_empty_pop_pages += nr;
> @@ -1260,6 +1269,7 @@ static void pcpu_chunk_depopulated(struct pcpu_chunk *chunk,
> chunk->nr_populated -= nr;
> chunk->nr_empty_pop_pages -= nr;
> pcpu_nr_empty_pop_pages -= nr;
> + pcpu_nr_populated -= nr;
> }
>
> /*
> @@ -2176,6 +2186,9 @@ int __init pcpu_setup_first_chunk(const struct pcpu_alloc_info *ai,
> pcpu_nr_empty_pop_pages = pcpu_first_chunk->nr_empty_pop_pages;
> pcpu_chunk_relocate(pcpu_first_chunk, -1);
>
> + /* include all regions of the first chunk */
> + pcpu_nr_populated += PFN_DOWN(size_sum);
> +
> pcpu_stats_chunk_alloc();
> trace_percpu_create_chunk(base_addr);
>
> @@ -2745,6 +2758,22 @@ void __init setup_per_cpu_areas(void)
>
> #endif /* CONFIG_SMP */
>
> +/*
> + * pcpu_nr_populated_pages - calculate total number of populated backing pages
> + *
> + * This reflects the number of pages populated to back the chunks.
> + * Metadata is excluded in the number exposed in meminfo as the number of
> + * backing pages scales with the number of cpus and can quickly outweigh the
> + * memory used for metadata. It also keeps this calculation nice and simple.
> + *
> + * RETURNS:
> + * Total number of populated backing pages in use by the allocator.
> + */
> +int pcpu_nr_populated_pages(void)
Also unsigned long please.
Thanks,
Vlastimil
> +{
> + return pcpu_nr_populated * pcpu_nr_units;
> +}
> +
> /*
> * Percpu allocator is initialized early during boot when neither slab or
> * workqueue is available. Plug async management until everything is up
>
Powered by blists - more mailing lists