From: Minchan Kim <minchan@kernel.org>
To: Vitaly Wool <vitaly.wool@konsulko.com>
Cc: Tian Tao <tiantao6@hisilicon.com>,
Seth Jennings <sjenning@redhat.com>,
Dan Streetman <ddstreet@ieee.org>,
Andrew Morton <akpm@linux-foundation.org>,
Barry Song <song.bao.hua@hisilicon.com>,
Linux-MM <linux-mm@kvack.org>, Shakeel Butt <shakeelb@google.com>,
Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com>
Subject: Re: [RFC mm/zswap 1/2] mm/zswap: add the flag can_sleep_mapped
Date: Thu, 14 Jan 2021 10:56:41 -0800 [thread overview]
Message-ID: <YACT6Ss+6Z+rLxtY@google.com> (raw)
In-Reply-To: <CAM4kBBJHL7jUYgfLPbMTsY4gZLLrps-tnmUr-UutGi0cE6gh1w@mail.gmail.com>
On Thu, Jan 14, 2021 at 07:40:50PM +0100, Vitaly Wool wrote:
> On Thu, Jan 14, 2021 at 7:29 PM Minchan Kim <minchan@kernel.org> wrote:
> >
> > On Fri, Dec 25, 2020 at 07:02:50PM +0800, Tian Tao wrote:
> > > add a flag to zpool, named is "can_sleep_mapped", and have it set true
> > > for zbud/z3fold, set false for zsmalloc. Then zswap could go the current
> > > path if the flag is true; and if it's false, copy data from src to a
> > > temporary buffer, then unmap the handle, take the mutex, process the
> > > buffer instead of src to avoid sleeping function called from atomic
> > > context.
> > >
> > > Signed-off-by: Tian Tao <tiantao6@hisilicon.com>
> > > ---
> > > include/linux/zpool.h | 3 +++
> > > mm/zpool.c | 13 +++++++++++++
> > > mm/zswap.c | 50 +++++++++++++++++++++++++++++++++++++++++++++-----
> > > 3 files changed, 61 insertions(+), 5 deletions(-)
> > >
> > > diff --git a/include/linux/zpool.h b/include/linux/zpool.h
> > > index 51bf430..e899701 100644
> > > --- a/include/linux/zpool.h
> > > +++ b/include/linux/zpool.h
> > > @@ -73,6 +73,7 @@ u64 zpool_get_total_size(struct zpool *pool);
> > > * @malloc: allocate mem from a pool.
> > > * @free: free mem from a pool.
> > > * @shrink: shrink the pool.
> > > + * @sleep_mapped: whether zpool driver can sleep during map.
> >
> > I don't think it's a good idea. It just breaks zpool abstraction
> > in that it exposes internal implementation to user to avoid issue
> > zswap recently introduced. It also conflicts zpool_map_handle's
> > semantic.
> >
> > Rather than introducing another break in zpool due to the new
> > zswap feature recenlty introduced, zswap could introduce
> > CONFIG_ZSWAP_HW_COMPRESSOR. Once it's configured, zsmalloc could
> > be disabled. And with disabling CONFIG_ZSWAP_HW_COMPRESSOR, zswap
> > doesn't need to make any bounce buffer copy so that no existing
> > zsmalloc user will see performance regression.
>
> I believe it won't help that much -- the new compressor API presumes
> that the caller may sleep during compression and that will be an
> accident waiting to happen as long as we use it and keep the handle
> mapped in zsmalloc case.
>
> Or maybe I interpreted you wrong and you are suggesting re-introducing
> calls to the old API under this #ifdef, is that the case?
Yub. zswap could abstract that part under #ifdef to keep old behavior.
>
> Best regards,
> Vitaly
>
> >
> > > * @map: map a handle.
> > > * @unmap: unmap a handle.
> > > * @total_size: get total size of a pool.
> > > @@ -100,6 +101,7 @@ struct zpool_driver {
> > > int (*shrink)(void *pool, unsigned int pages,
> > > unsigned int *reclaimed);
> > >
> > > + bool sleep_mapped;
> > > void *(*map)(void *pool, unsigned long handle,
> > > enum zpool_mapmode mm);
> > > void (*unmap)(void *pool, unsigned long handle);
> > > @@ -112,5 +114,6 @@ void zpool_register_driver(struct zpool_driver *driver);
> > > int zpool_unregister_driver(struct zpool_driver *driver);
> > >
> > > bool zpool_evictable(struct zpool *pool);
> > > +bool zpool_can_sleep_mapped(struct zpool *pool);
> > >
> > > #endif
> > > diff --git a/mm/zpool.c b/mm/zpool.c
> > > index 3744a2d..5ed7120 100644
> > > --- a/mm/zpool.c
> > > +++ b/mm/zpool.c
> > > @@ -23,6 +23,7 @@ struct zpool {
> > > void *pool;
> > > const struct zpool_ops *ops;
> > > bool evictable;
> > > + bool can_sleep_mapped;
> > >
> > > struct list_head list;
> > > };
> > > @@ -183,6 +184,7 @@ struct zpool *zpool_create_pool(const char *type, const char *name, gfp_t gfp,
> > > zpool->pool = driver->create(name, gfp, ops, zpool);
> > > zpool->ops = ops;
> > > zpool->evictable = driver->shrink && ops && ops->evict;
> > > + zpool->can_sleep_mapped = driver->sleep_mapped;
> > >
> > > if (!zpool->pool) {
> > > pr_err("couldn't create %s pool\n", type);
> > > @@ -393,6 +395,17 @@ bool zpool_evictable(struct zpool *zpool)
> > > return zpool->evictable;
> > > }
> > >
> > > +/**
> > > + * zpool_can_sleep_mapped - Test if zpool can sleep when do mapped.
> > > + * @zpool: The zpool to test
> > > + *
> > > + * Returns: true if zpool can sleep; false otherwise.
> > > + */
> > > +bool zpool_can_sleep_mapped(struct zpool *zpool)
> > > +{
> > > + return zpool->can_sleep_mapped;
> > > +}
> > > +
> > > MODULE_LICENSE("GPL");
> > > MODULE_AUTHOR("Dan Streetman <ddstreet@ieee.org>");
> > > MODULE_DESCRIPTION("Common API for compressed memory storage");
> > > diff --git a/mm/zswap.c b/mm/zswap.c
> > > index 182f6ad..67d4555 100644
> > > --- a/mm/zswap.c
> > > +++ b/mm/zswap.c
> > > @@ -935,13 +935,20 @@ static int zswap_writeback_entry(struct zpool *pool, unsigned long handle)
> > > struct scatterlist input, output;
> > > struct crypto_acomp_ctx *acomp_ctx;
> > >
> > > - u8 *src;
> > > + u8 *src, *tmp;
> > > unsigned int dlen;
> > > int ret;
> > > struct writeback_control wbc = {
> > > .sync_mode = WB_SYNC_NONE,
> > > };
> > >
> > > + if (!zpool_can_sleep_mapped(pool)) {
> > > +
> > > + tmp = kmalloc(entry->length, GFP_ATOMIC);
> > > + if (!tmp)
> > > + return -ENOMEM;
> > > + }
> > > +
> > > /* extract swpentry from data */
> > > zhdr = zpool_map_handle(pool, handle, ZPOOL_MM_RO);
> > > swpentry = zhdr->swpentry; /* here */
> > > @@ -979,6 +986,14 @@ static int zswap_writeback_entry(struct zpool *pool, unsigned long handle)
> > > dlen = PAGE_SIZE;
> > > src = (u8 *)zhdr + sizeof(struct zswap_header);
> > >
> > > + if (!zpool_can_sleep_mapped(pool)) {
> > > +
> > > + memcpy(tmp, src, entry->length);
> > > + src = tmp;
> > > +
> > > + zpool_unmap_handle(pool, handle);
> > > + }
> > > +
> > > mutex_lock(acomp_ctx->mutex);
> > > sg_init_one(&input, src, entry->length);
> > > sg_init_table(&output, 1);
> > > @@ -1033,7 +1048,11 @@ static int zswap_writeback_entry(struct zpool *pool, unsigned long handle)
> > > spin_unlock(&tree->lock);
> > >
> > > end:
> > > - zpool_unmap_handle(pool, handle);
> > > + if (zpool_can_sleep_mapped(pool))
> > > + zpool_unmap_handle(pool, handle);
> > > + else
> > > + kfree(tmp);
> > > +
> > > return ret;
> > > }
> > >
> > > @@ -1235,7 +1254,7 @@ static int zswap_frontswap_load(unsigned type, pgoff_t offset,
> > > struct zswap_entry *entry;
> > > struct scatterlist input, output;
> > > struct crypto_acomp_ctx *acomp_ctx;
> > > - u8 *src, *dst;
> > > + u8 *src, *dst, *tmp;
> > > unsigned int dlen;
> > > int ret;
> > >
> > > @@ -1256,12 +1275,29 @@ static int zswap_frontswap_load(unsigned type, pgoff_t offset,
> > > goto freeentry;
> > > }
> > >
> > > + if (!zpool_can_sleep_mapped(entry->pool->zpool)) {
> > > +
> > > + tmp = kmalloc(entry->length, GFP_ATOMIC);
> > > + if (!tmp) {
> > > + ret = -ENOMEM;
> > > + goto freeentry;
> > > + }
> > > + }
> > > +
> > > /* decompress */
> > > dlen = PAGE_SIZE;
> > > src = zpool_map_handle(entry->pool->zpool, entry->handle, ZPOOL_MM_RO);
> > > if (zpool_evictable(entry->pool->zpool))
> > > src += sizeof(struct zswap_header);
> > >
> > > + if (!zpool_can_sleep_mapped(entry->pool->zpool)) {
> > > +
> > > + memcpy(tmp, src, entry->length);
> > > + src = tmp;
> > > +
> > > + zpool_unmap_handle(entry->pool->zpool, entry->handle);
> > > + }
> > > +
> > > acomp_ctx = raw_cpu_ptr(entry->pool->acomp_ctx);
> > > mutex_lock(acomp_ctx->mutex);
> > > sg_init_one(&input, src, entry->length);
> > > @@ -1271,7 +1307,11 @@ static int zswap_frontswap_load(unsigned type, pgoff_t offset,
> > > ret = crypto_wait_req(crypto_acomp_decompress(acomp_ctx->req), &acomp_ctx->wait);
> > > mutex_unlock(acomp_ctx->mutex);
> > >
> > > - zpool_unmap_handle(entry->pool->zpool, entry->handle);
> > > + if (zpool_can_sleep_mapped(entry->pool->zpool))
> > > + zpool_unmap_handle(entry->pool->zpool, entry->handle);
> > > + else
> > > + kfree(tmp);
> > > +
> > > BUG_ON(ret);
> > >
> > > freeentry:
> > > @@ -1279,7 +1319,7 @@ static int zswap_frontswap_load(unsigned type, pgoff_t offset,
> > > zswap_entry_put(tree, entry);
> > > spin_unlock(&tree->lock);
> > >
> > > - return 0;
> > > + return ret;
> > > }
> > >
> > > /* frees an entry in zswap */
> > > --
> > > 2.7.4
> > >
> > >
next prev parent reply other threads:[~2021-01-14 18:56 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-12-25 11:02 [RFC mm/zswap 0/2] Fix the compatibility of zsmalloc and zswap Tian Tao
2020-12-25 11:02 ` [RFC mm/zswap 1/2] mm/zswap: add the flag can_sleep_mapped Tian Tao
2021-01-14 18:28 ` Minchan Kim
2021-01-14 18:40 ` Vitaly Wool
2021-01-14 18:56 ` Minchan Kim [this message]
2021-01-14 19:05 ` Vitaly Wool
2021-01-14 19:21 ` Shakeel Butt
2021-01-14 19:23 ` Minchan Kim
2021-01-14 19:53 ` Vitaly Wool
2021-01-14 21:09 ` Shakeel Butt
2021-01-14 22:41 ` Vitaly Wool
2021-01-19 1:28 ` tiantao (H)
2021-01-14 18:43 ` Shakeel Butt
2021-01-14 18:53 ` Vitaly Wool
2021-01-21 9:17 ` Vitaly Wool
2021-01-21 23:15 ` Song Bao Hua (Barry Song)
2020-12-25 11:02 ` [RFC mm/zswap 2/2] mm: set the sleep_mapped to true for zbud and z3fold Tian Tao
2021-01-14 18:46 ` [RFC mm/zswap 0/2] Fix the compatibility of zsmalloc and zswap Vitaly Wool
2021-01-15 1:17 ` tiantao (H)
2021-01-19 1:31 ` tiantao (H)
2021-01-19 2:39 ` Mike Galbraith
2021-01-18 13:44 ` Sebastian Andrzej Siewior
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=YACT6Ss+6Z+rLxtY@google.com \
--to=minchan@kernel.org \
--cc=akpm@linux-foundation.org \
--cc=ddstreet@ieee.org \
--cc=linux-mm@kvack.org \
--cc=sergey.senozhatsky.work@gmail.com \
--cc=shakeelb@google.com \
--cc=sjenning@redhat.com \
--cc=song.bao.hua@hisilicon.com \
--cc=tiantao6@hisilicon.com \
--cc=vitaly.wool@konsulko.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).