On Thu, Jan 14, 2021 at 11:05 AM Vitaly Wool [off-list ref] wrote:

On Thu, Jan 14, 2021 at 7:56 PM Minchan Kim [off-list ref] wrote:

quoted

On Thu, Jan 14, 2021 at 07:40:50PM +0100, Vitaly Wool wrote:

quoted

On Thu, Jan 14, 2021 at 7:29 PM Minchan Kim [off-list ref] wrote:

quoted

On Fri, Dec 25, 2020 at 07:02:50PM +0800, Tian Tao wrote:

quoted

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.

We can reconsider this option when zsmalloc implements reclaim
callback. So far it's obviously too much a mess for a reason so weak.

Sorry I don't understand the link between zsmalloc implementing shrink
callback and this patch. This patch is adding an overhead for all
zswap+zsmalloc users irrespective of availability of hardware. If we
want to add support for new hardware, please add without impacting the
current users.

Shakeel