[PATCH v3] f2fs: allow to change discard policy based on cached discard cmds

[PATCH v3] f2fs: allow to change discard policy based on cached discard cmds

[Sahitya Tummala]

With the default DPOLICY_BG discard thread is ioaware, which prevents
the discard thread from issuing the discard commands. On low RAM setups,
it is observed that these discard commands in the cache are consuming
high memory. This patch aims to relax the memory pressure on the system
due to f2fs pending discard cmds by changing the policy to DPOLICY_FORCE
based on the nm_i->ram_thresh configured.

Signed-off-by: Sahitya Tummala <stummala@codeaurora.org>
---
v3:
Fix sporadic null pointer dereference issue when accessing NM_I(sbi)
in f2fs_available_free_memory(). It is because this function can be
called by discard thread before f2fs_build_node_manager() in
f2fs_fill_super().

 fs/f2fs/node.c    | 8 ++++++++
 fs/f2fs/node.h    | 1 +
 fs/f2fs/segment.c | 3 ++-
 3 files changed, 11 insertions(+), 1 deletion(-)

diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
index 3a24423..256c39b 100644
--- a/fs/f2fs/node.c
+++ b/fs/f2fs/node.c
@@ -44,11 +44,15 @@ int f2fs_check_nid_range(struct f2fs_sb_info *sbi, nid_t nid)
 bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type)
 {
 	struct f2fs_nm_info *nm_i = NM_I(sbi);
+	struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
 	struct sysinfo val;
 	unsigned long avail_ram;
 	unsigned long mem_size = 0;
 	bool res = false;
 
+	if (!nm_i)
+		return true;
+
 	si_meminfo(&val);
 
 	/* only uses low memory */
@@ -90,6 +94,10 @@ bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type)
 		/* it allows 20% / total_ram for inmemory pages */
 		mem_size = get_pages(sbi, F2FS_INMEM_PAGES);
 		res = mem_size < (val.totalram / 5);
+	} else if (type == DISCARD_CACHE) {
+		mem_size = (atomic_read(&dcc->discard_cmd_cnt) *
+				sizeof(struct discard_cmd)) >> PAGE_SHIFT;
+		res = mem_size < (avail_ram * nm_i->ram_thresh / 100);
 	} else {
 		if (!sbi->sb->s_bdi->wb.dirty_exceeded)
 			return true;
diff --git a/fs/f2fs/node.h b/fs/f2fs/node.h
index f84541b..7a45c0f 100644
--- a/fs/f2fs/node.h
+++ b/fs/f2fs/node.h
@@ -147,6 +147,7 @@ enum mem_type {
 	INO_ENTRIES,	/* indicates inode entries */
 	EXTENT_CACHE,	/* indicates extent cache */
 	INMEM_PAGES,	/* indicates inmemory pages */
+	DISCARD_CACHE,	/* indicates memory of cached discard cmds */
 	BASE_CHECK,	/* check kernel status */
 };
 
diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index f214b69..d7076796 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -1775,7 +1775,8 @@ static int issue_discard_thread(void *data)
 		if (!atomic_read(&dcc->discard_cmd_cnt))
 			continue;
 
-		if (sbi->gc_mode == GC_URGENT_HIGH)
+		if (sbi->gc_mode == GC_URGENT_HIGH ||
+			!f2fs_available_free_memory(sbi, DISCARD_CACHE))
 			__init_discard_policy(sbi, &dpolicy, DPOLICY_FORCE, 1);
 
 		sb_start_intwrite(sbi->sb);
-- 
Qualcomm India Private Limited, on behalf of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project.

Re: [PATCH v3] f2fs: allow to change discard policy based on cached discard cmds

[Chao Yu]

On 2021/3/16 17:29, Sahitya Tummala wrote:
> With the default DPOLICY_BG discard thread is ioaware, which prevents
> the discard thread from issuing the discard commands. On low RAM setups,
> it is observed that these discard commands in the cache are consuming
> high memory. This patch aims to relax the memory pressure on the system
> due to f2fs pending discard cmds by changing the policy to DPOLICY_FORCE
> based on the nm_i->ram_thresh configured.
> 
> Signed-off-by: Sahitya Tummala <stummala@codeaurora.org>

Reviewed-by: Chao Yu <yuchao0@huawei.com>

Thanks,

Re: [PATCH v3] f2fs: allow to change discard policy based on cached discard cmds

[Sahitya Tummala]

Hi Jaegeuk,

This latest v3 patch needs to be updated in f2fs tree.
The f2fs tree currently points to older version of patch.

Please make a note of it.

Thanks,
Sahitya.

On Tue, Mar 16, 2021 at 07:08:58PM +0800, Chao Yu wrote:
> On 2021/3/16 17:29, Sahitya Tummala wrote:
> >With the default DPOLICY_BG discard thread is ioaware, which prevents
> >the discard thread from issuing the discard commands. On low RAM setups,
> >it is observed that these discard commands in the cache are consuming
> >high memory. This patch aims to relax the memory pressure on the system
> >due to f2fs pending discard cmds by changing the policy to DPOLICY_FORCE
> >based on the nm_i->ram_thresh configured.
> >
> >Signed-off-by: Sahitya Tummala <stummala@codeaurora.org>
> 
> Reviewed-by: Chao Yu <yuchao0@huawei.com>
> 
> Thanks,

-- 
--
Sent by a consultant of the Qualcomm Innovation Center, Inc.
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum.

Re: [PATCH v3] f2fs: allow to change discard policy based on cached discard cmds

[Jaegeuk Kim]

On 03/26, Sahitya Tummala wrote:
> Hi Jaegeuk,
> 
> This latest v3 patch needs to be updated in f2fs tree.
> The f2fs tree currently points to older version of patch.
> 
> Please make a note of it.

Ha, need more coffee. Thanks for pointing it out. :)

> 
> Thanks,
> Sahitya.
> 
> On Tue, Mar 16, 2021 at 07:08:58PM +0800, Chao Yu wrote:
> > On 2021/3/16 17:29, Sahitya Tummala wrote:
> > >With the default DPOLICY_BG discard thread is ioaware, which prevents
> > >the discard thread from issuing the discard commands. On low RAM setups,
> > >it is observed that these discard commands in the cache are consuming
> > >high memory. This patch aims to relax the memory pressure on the system
> > >due to f2fs pending discard cmds by changing the policy to DPOLICY_FORCE
> > >based on the nm_i->ram_thresh configured.
> > >
> > >Signed-off-by: Sahitya Tummala <stummala@codeaurora.org>
> > 
> > Reviewed-by: Chao Yu <yuchao0@huawei.com>
> > 
> > Thanks,
> 
> -- 
> --
> Sent by a consultant of the Qualcomm Innovation Center, Inc.
> The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum.