From: Phillip Lougher <phillip@squashfs.org.uk>
To: Hui Wang <hui.wang@canonical.com>, Michal Hocko <mhocko@suse.com>
Cc: Gao Xiang <hsiangkao@linux.alibaba.com>,
linux-mm@kvack.org, akpm@linux-foundation.org, surenb@google.com,
colin.i.king@gmail.com, shy828301@gmail.com, hannes@cmpxchg.org,
vbabka@suse.cz, hch@infradead.org, mgorman@suse.de
Subject: Re: [PATCH 1/1] mm/oom_kill: trigger the oom killer if oom occurs without __GFP_FS
Date: Sun, 7 May 2023 22:07:07 +0100 [thread overview]
Message-ID: <b8f57b19-89f5-6a17-d5de-8ab488b96ac6@squashfs.org.uk> (raw)
In-Reply-To: <70000460-ace2-3965-084d-34be65a6bd6a@squashfs.org.uk>
On 03/05/2023 20:10, Phillip Lougher wrote:
>
> On 03/05/2023 12:49, Hui Wang wrote:
>>
>> On 4/29/23 03:53, Michal Hocko wrote:
>>> On Thu 27-04-23 11:47:10, Hui Wang wrote:
>>> [...]
>>>> So Michal,
>>>>
>>>> Don't know if you read the "[PATCH 0/1] mm/oom_kill: system enters
>>>> a state
>>>> something like hang when running stress-ng", do you know why
>>>> out_of_memory()
>>>> will return immediately if there is no __GFP_FS, could we drop
>>>> these lines
>>>> directly:
>>>>
>>>> /*
>>>> * The OOM killer does not compensate for IO-less reclaim.
>>>> * pagefault_out_of_memory lost its gfp context so we have to
>>>> * make sure exclude 0 mask - all other users should have at
>>>> least
>>>> * ___GFP_DIRECT_RECLAIM to get here. But mem_cgroup_oom() has to
>>>> * invoke the OOM killer even if it is a GFP_NOFS allocation.
>>>> */
>>>> if (oc->gfp_mask && !(oc->gfp_mask & __GFP_FS) &&
>>>> !is_memcg_oom(oc))
>>>> return true;
>>> The comment is rather hard to grasp without an intimate knowledge of
>>> the
>>> memory reclaim. The primary reason is that the allocation context
>>> without __GFP_FS (and also __GFP_IO) cannot perform a full memory
>>> reclaim because fs or the storage subsystem might be holding locks
>>> required for the memory reclaim. This means that a large amount of
>>> reclaimable memory is out of sight of the specific direct reclaim
>>> context. If we allowed oom killer to trigger we could invoke the oom
>>> killer while there is a lot of otherwise reclaimable memory. As you can
>>> imagine not something many users would appreciate as the oom kill is a
>>> very disruptive operation. In this case we rely on kswapd or other
>>> GFP_KERNEL like allocation context to make forward instead. If there is
>>> really nothing reclaimable then the oom killer would eventually hit
>>> from
>>> elsewhere.
>>>
>>> HTH
>> Hi Michal,
>>
>> Understand. Thanks for explanation. So we can't remove those 2 lines
>> of code.
>>
>> Here in my patch, letting a kthread allocate a page with GFP_KERNEL,
>> It could possibly trigger the reclaim and if nothing reclaimable,
>> trigger the oom killer. Do you think it is a safe workaround for the
>> issue we are facing currently?
>>
>>
>> And Hi Phillip,
>>
>> What is your opinion on it, do you have a direction to solve this
>> issue from filesystem?
>>
>
> The following patch creates the concept of "squashfs contexts", which
> moves all memory dynamically allocated (in a readahead/read_page path)
> into a single structure which can be allocated and deleted once. It
> then creates a pool of these at filesystem mount time. Threads
> entering readahead/read_page will take a context from the pool, and
> will then perform no dynamic memory allocation.
>
> The final patch-series will make this a non-default build option for
> systems that need this.
>
> Phillip
>
>
An updated version of the patch.
From 24f22dbd7fed8d0a90d11a4ce80545d30a9acfdc Mon Sep 17 00:00:00 2001
From: Phillip Lougher <phillip@squashfs.org.uk>
Date: Sun, 7 May 2023 20:16:55 +0100
Subject: [PATCH] read_context v2
Signed-off-by: Phillip Lougher <phillip@squashfs.org.uk>
---
fs/squashfs/Makefile | 1 +
fs/squashfs/block.c | 67 ++++-----
fs/squashfs/cache.c | 66 ++++++---
fs/squashfs/decompressor.c | 14 +-
fs/squashfs/decompressor_multi.c | 2 +
fs/squashfs/decompressor_multi_percpu.c | 2 +
fs/squashfs/decompressor_single.c | 2 +
fs/squashfs/file.c | 61 ++++----
fs/squashfs/file_direct.c | 59 ++++----
fs/squashfs/page_actor.c | 37 ++---
fs/squashfs/page_actor.h | 8 +-
fs/squashfs/read_context.c | 181 ++++++++++++++++++++++++
fs/squashfs/read_context.h | 31 ++++
fs/squashfs/squashfs.h | 12 +-
fs/squashfs/squashfs_fs_sb.h | 3 +-
fs/squashfs/super.c | 13 +-
fs/squashfs/symlink.c | 3 +-
17 files changed, 390 insertions(+), 172 deletions(-)
create mode 100644 fs/squashfs/read_context.c
create mode 100644 fs/squashfs/read_context.h
diff --git a/fs/squashfs/Makefile b/fs/squashfs/Makefile
index 477c89a519ee..93511720e94f 100644
--- a/fs/squashfs/Makefile
+++ b/fs/squashfs/Makefile
@@ -6,6 +6,7 @@
obj-$(CONFIG_SQUASHFS) += squashfs.o
squashfs-y += block.o cache.o dir.o export.o file.o fragment.o id.o inode.o
squashfs-y += namei.o super.o symlink.o decompressor.o page_actor.o
+squashfs-y += read_context.o
squashfs-$(CONFIG_SQUASHFS_FILE_CACHE) += file_cache.o
squashfs-$(CONFIG_SQUASHFS_FILE_DIRECT) += file_direct.o
squashfs-$(CONFIG_SQUASHFS_DECOMP_SINGLE) += decompressor_single.o
diff --git a/fs/squashfs/block.c b/fs/squashfs/block.c
index bed3bb8b27fa..110fcfd2a0eb 100644
--- a/fs/squashfs/block.c
+++ b/fs/squashfs/block.c
@@ -23,9 +23,10 @@
#include "squashfs_fs.h"
#include "squashfs_fs_sb.h"
+#include "read_context.h"
+#include "page_actor.h"
#include "squashfs.h"
#include "decompressor.h"
-#include "page_actor.h"
/*
* Returns the amount of bytes copied to the page actor.
@@ -77,7 +78,7 @@ static int copy_bio_to_actor(struct bio *bio,
}
static int squashfs_bio_read(struct super_block *sb, u64 index, int length,
- struct bio **biop, int *block_offset)
+ struct squashfs_context *context, int *block_offset)
{
struct squashfs_sb_info *msblk = sb->s_fs_info;
const u64 read_start = round_down(index, msblk->devblksize);
@@ -87,44 +88,33 @@ static int squashfs_bio_read(struct super_block *sb, u64 index, int length,
int offset = read_start - round_down(index, PAGE_SIZE);
int total_len = (block_end - block) << msblk->devblksize_log2;
const int page_count = DIV_ROUND_UP(total_len + offset, PAGE_SIZE);
- int error, i;
- struct bio *bio;
+ struct bio *bio = context->bio;
+ int error = -EIO, i;
- bio = bio_kmalloc(page_count, GFP_NOIO);
- if (!bio)
- return -ENOMEM;
bio_init(bio, sb->s_bdev, bio->bi_inline_vecs, page_count, REQ_OP_READ);
bio->bi_iter.bi_sector = block * (msblk->devblksize >> SECTOR_SHIFT);
for (i = 0; i < page_count; ++i) {
unsigned int len =
min_t(unsigned int, PAGE_SIZE - offset, total_len);
- struct page *page = alloc_page(GFP_NOIO);
+ int res = bio_add_page(bio, context->bio_page[i], len, offset);
+
+ if (!res)
+ goto out_uninit_bio;
- if (!page) {
- error = -ENOMEM;
- goto out_free_bio;
- }
- if (!bio_add_page(bio, page, len, offset)) {
- error = -EIO;
- goto out_free_bio;
- }
offset = 0;
total_len -= len;
}
error = submit_bio_wait(bio);
if (error)
- goto out_free_bio;
+ goto out_uninit_bio;
- *biop = bio;
*block_offset = index & ((1 << msblk->devblksize_log2) - 1);
return 0;
-out_free_bio:
- bio_free_pages(bio);
- bio_uninit(bio);
- kfree(bio);
+out_uninit_bio:
+ bio_uninit(context->bio);
return error;
}
@@ -138,10 +128,10 @@ static int squashfs_bio_read(struct super_block *sb, u64 index, int length,
* algorithms).
*/
int squashfs_read_data(struct super_block *sb, u64 index, int length,
- u64 *next_index, struct squashfs_page_actor *output)
+ u64 *next_index, struct squashfs_context *context)
{
struct squashfs_sb_info *msblk = sb->s_fs_info;
- struct bio *bio = NULL;
+ struct squashfs_page_actor *output = context->actor;
int compressed;
int res;
int offset;
@@ -166,13 +156,13 @@ int squashfs_read_data(struct super_block *sb, u64 index, int length,
res = -EIO;
goto out;
}
- res = squashfs_bio_read(sb, index, 2, &bio, &offset);
+ res = squashfs_bio_read(sb, index, 2, context, &offset);
if (res)
goto out;
- if (WARN_ON_ONCE(!bio_next_segment(bio, &iter_all))) {
+ if (WARN_ON_ONCE(!bio_next_segment(context->bio, &iter_all))) {
res = -EIO;
- goto out_free_bio;
+ goto out_uninit_bio;
}
/* Extract the length of the metadata block */
data = bvec_virt(bvec);
@@ -180,16 +170,14 @@ int squashfs_read_data(struct super_block *sb, u64 index, int length,
if (offset < bvec->bv_len - 1) {
length |= data[offset + 1] << 8;
} else {
- if (WARN_ON_ONCE(!bio_next_segment(bio, &iter_all))) {
+ if (WARN_ON_ONCE(!bio_next_segment(context->bio, &iter_all))) {
res = -EIO;
- goto out_free_bio;
+ goto out_uninit_bio;
}
data = bvec_virt(bvec);
length |= data[0] << 8;
}
- bio_free_pages(bio);
- bio_uninit(bio);
- kfree(bio);
+ bio_uninit(context->bio);
compressed = SQUASHFS_COMPRESSED(length);
length = SQUASHFS_COMPRESSED_SIZE(length);
@@ -207,24 +195,23 @@ int squashfs_read_data(struct super_block *sb, u64 index, int length,
if (next_index)
*next_index = index + length;
- res = squashfs_bio_read(sb, index, length, &bio, &offset);
+ res = squashfs_bio_read(sb, index, length, context, &offset);
if (res)
goto out;
if (compressed) {
if (!msblk->stream) {
res = -EIO;
- goto out_free_bio;
+ goto out_uninit_bio;
}
- res = msblk->thread_ops->decompress(msblk, bio, offset, length, output);
+ res = msblk->thread_ops->decompress(msblk, context->bio,
+ offset, length, output);
} else {
- res = copy_bio_to_actor(bio, output, offset, length);
+ res = copy_bio_to_actor(context->bio, output, offset, length);
}
-out_free_bio:
- bio_free_pages(bio);
- bio_uninit(bio);
- kfree(bio);
+out_uninit_bio:
+ bio_uninit(context->bio);
out:
if (res < 0) {
ERROR("Failed to read block 0x%llx: %d\n", index, res);
diff --git a/fs/squashfs/cache.c b/fs/squashfs/cache.c
index 5062326d0efb..98af4f696edd 100644
--- a/fs/squashfs/cache.c
+++ b/fs/squashfs/cache.c
@@ -42,19 +42,24 @@
#include "squashfs_fs.h"
#include "squashfs_fs_sb.h"
-#include "squashfs.h"
+#include "read_context.h"
#include "page_actor.h"
+#include "squashfs.h"
/*
* Look-up block in cache, and increment usage count. If not in cache, read
* and decompress it from disk.
*/
struct squashfs_cache_entry *squashfs_cache_get(struct super_block *sb,
- struct squashfs_cache *cache, u64 block, int length)
+ struct squashfs_cache *cache, struct squashfs_contexts *contexts,
+ struct squashfs_context *context, u64 block, int length)
{
- int i, n;
+ int i, n, need_context = context == NULL;
struct squashfs_cache_entry *entry;
+ if (cache->contexts)
+ contexts = cache->contexts;
+
spin_lock(&cache->lock);
while (1) {
@@ -107,8 +112,17 @@ struct squashfs_cache_entry *squashfs_cache_get(struct super_block *sb,
entry->error = 0;
spin_unlock(&cache->lock);
+ if (need_context)
+ context = squashfs_get_context(contexts);
+
+ squashfs_page_actor_init(context->actor, entry->data,
+ cache->pages, 0);
+
entry->length = squashfs_read_data(sb, block, length,
- &entry->next_index, entry->actor);
+ &entry->next_index, context);
+
+ if (need_context)
+ squashfs_put_context(contexts, context);
spin_lock(&cache->lock);
@@ -207,9 +221,9 @@ void squashfs_cache_delete(struct squashfs_cache *cache)
kfree(cache->entry[i].data[j]);
kfree(cache->entry[i].data);
}
- kfree(cache->entry[i].actor);
}
+ squashfs_delete_contexts(cache->contexts);
kfree(cache->entry);
kfree(cache);
}
@@ -221,7 +235,7 @@ void squashfs_cache_delete(struct squashfs_cache *cache)
* is allocated as a sequence of kmalloced PAGE_SIZE buffers.
*/
struct squashfs_cache *squashfs_cache_init(char *name, int entries,
- int block_size)
+ int block_size, int alloc_contexts)
{
int i, j;
struct squashfs_cache *cache = kzalloc(sizeof(*cache), GFP_KERNEL);
@@ -237,6 +251,15 @@ struct squashfs_cache *squashfs_cache_init(char *name, int entries,
goto cleanup;
}
+ if (alloc_contexts) {
+ cache->contexts = squashfs_create_contexts(block_size, entries,
+ 0, 0);
+ if (cache->contexts == NULL) {
+ ERROR("Failed to allocate %s cache\n", name);
+ goto cleanup;
+ }
+ }
+
cache->curr_blk = 0;
cache->next_blk = 0;
cache->unused = entries;
@@ -268,13 +291,6 @@ struct squashfs_cache *squashfs_cache_init(char *name, int entries,
goto cleanup;
}
}
-
- entry->actor = squashfs_page_actor_init(entry->data,
- cache->pages, 0);
- if (entry->actor == NULL) {
- ERROR("Failed to allocate %s cache entry\n", name);
- goto cleanup;
- }
}
return cache;
@@ -341,7 +357,8 @@ int squashfs_read_metadata(struct super_block *sb, void *buffer,
return -EIO;
while (length) {
- entry = squashfs_cache_get(sb, msblk->block_cache, *block, 0);
+ entry = squashfs_cache_get(sb, msblk->block_cache, NULL, NULL,
+ *block, 0);
if (entry->error) {
res = entry->error;
goto error;
@@ -377,12 +394,12 @@ int squashfs_read_metadata(struct super_block *sb, void *buffer,
* filesystem. If necessary read and decompress it from disk.
*/
struct squashfs_cache_entry *squashfs_get_fragment(struct super_block *sb,
- u64 start_block, int length)
+ u64 start_block, int length, struct squashfs_context *context)
{
struct squashfs_sb_info *msblk = sb->s_fs_info;
- return squashfs_cache_get(sb, msblk->fragment_cache, start_block,
- length);
+ return squashfs_cache_get(sb, msblk->fragment_cache, msblk->contexts,
+ context, start_block, length);
}
@@ -396,7 +413,8 @@ struct squashfs_cache_entry *squashfs_get_datablock(struct super_block *sb,
{
struct squashfs_sb_info *msblk = sb->s_fs_info;
- return squashfs_cache_get(sb, msblk->read_page, start_block, length);
+ return squashfs_cache_get(sb, msblk->read_page, msblk->contexts, NULL,
+ start_block, length);
}
@@ -408,7 +426,7 @@ void *squashfs_read_table(struct super_block *sb, u64 block, int length)
int pages = (length + PAGE_SIZE - 1) >> PAGE_SHIFT;
int i, res;
void *table, *buffer, **data;
- struct squashfs_page_actor *actor;
+ struct squashfs_context *context;
table = buffer = kmalloc(length, GFP_KERNEL);
if (table == NULL)
@@ -420,20 +438,22 @@ void *squashfs_read_table(struct super_block *sb, u64 block, int length)
goto failed;
}
- actor = squashfs_page_actor_init(data, pages, length);
- if (actor == NULL) {
+ context = squashfs_create_context(length, 0, 0);
+ if (context == NULL) {
res = -ENOMEM;
goto failed2;
}
+ squashfs_page_actor_init(context->actor, data, pages, length);
+
for (i = 0; i < pages; i++, buffer += PAGE_SIZE)
data[i] = buffer;
res = squashfs_read_data(sb, block, length |
- SQUASHFS_COMPRESSED_BIT_BLOCK, NULL, actor);
+ SQUASHFS_COMPRESSED_BIT_BLOCK, NULL, context);
+ squashfs_delete_context(context);
kfree(data);
- kfree(actor);
if (res < 0)
goto failed;
diff --git a/fs/squashfs/decompressor.c b/fs/squashfs/decompressor.c
index 8893cb9b4198..d12255edb129 100644
--- a/fs/squashfs/decompressor.c
+++ b/fs/squashfs/decompressor.c
@@ -15,9 +15,10 @@
#include "squashfs_fs.h"
#include "squashfs_fs_sb.h"
+#include "read_context.h"
+#include "page_actor.h"
#include "decompressor.h"
#include "squashfs.h"
-#include "page_actor.h"
/*
* This file (and decompressor.h) implements a decompressor framework for
@@ -89,7 +90,7 @@ static void *get_comp_opts(struct super_block *sb, unsigned short flags)
{
struct squashfs_sb_info *msblk = sb->s_fs_info;
void *buffer = NULL, *comp_opts;
- struct squashfs_page_actor *actor = NULL;
+ struct squashfs_context *context;
int length = 0;
/*
@@ -102,14 +103,16 @@ static void *get_comp_opts(struct super_block *sb, unsigned short flags)
goto out;
}
- actor = squashfs_page_actor_init(&buffer, 1, 0);
- if (actor == NULL) {
+ context = squashfs_create_context(PAGE_SIZE, 0, 0);
+ if (context == NULL) {
comp_opts = ERR_PTR(-ENOMEM);
goto out;
}
+ squashfs_page_actor_init(context->actor, &buffer, 1, 0);
length = squashfs_read_data(sb,
- sizeof(struct squashfs_super_block), 0, NULL, actor);
+ sizeof(struct squashfs_super_block), 0, NULL, context);
+ squashfs_delete_context(context);
if (length < 0) {
comp_opts = ERR_PTR(length);
@@ -120,7 +123,6 @@ static void *get_comp_opts(struct super_block *sb, unsigned short flags)
comp_opts = squashfs_comp_opts(msblk, buffer, length);
out:
- kfree(actor);
kfree(buffer);
return comp_opts;
}
diff --git a/fs/squashfs/decompressor_multi.c b/fs/squashfs/decompressor_multi.c
index 416c53eedbd1..0a31012abec7 100644
--- a/fs/squashfs/decompressor_multi.c
+++ b/fs/squashfs/decompressor_multi.c
@@ -13,6 +13,8 @@
#include "squashfs_fs.h"
#include "squashfs_fs_sb.h"
+#include "read_context.h"
+#include "page_actor.h"
#include "decompressor.h"
#include "squashfs.h"
diff --git a/fs/squashfs/decompressor_multi_percpu.c b/fs/squashfs/decompressor_multi_percpu.c
index 1dfadf76ed9a..12094aaba19f 100644
--- a/fs/squashfs/decompressor_multi_percpu.c
+++ b/fs/squashfs/decompressor_multi_percpu.c
@@ -12,6 +12,8 @@
#include "squashfs_fs.h"
#include "squashfs_fs_sb.h"
+#include "read_context.h"
+#include "page_actor.h"
#include "decompressor.h"
#include "squashfs.h"
diff --git a/fs/squashfs/decompressor_single.c b/fs/squashfs/decompressor_single.c
index 6f161887710b..8d9b73103165 100644
--- a/fs/squashfs/decompressor_single.c
+++ b/fs/squashfs/decompressor_single.c
@@ -11,6 +11,8 @@
#include "squashfs_fs.h"
#include "squashfs_fs_sb.h"
+#include "read_context.h"
+#include "page_actor.h"
#include "decompressor.h"
#include "squashfs.h"
diff --git a/fs/squashfs/file.c b/fs/squashfs/file.c
index 8ba8c4c50770..d8ca63a3eb0e 100644
--- a/fs/squashfs/file.c
+++ b/fs/squashfs/file.c
@@ -38,8 +38,9 @@
#include "squashfs_fs.h"
#include "squashfs_fs_sb.h"
#include "squashfs_fs_i.h"
-#include "squashfs.h"
+#include "read_context.h"
#include "page_actor.h"
+#include "squashfs.h"
/*
* Locate cache slot in range [offset, index] for specified inode. If
@@ -424,7 +425,7 @@ static int squashfs_readpage_fragment(struct page *page, int expected)
struct inode *inode = page->mapping->host;
struct squashfs_cache_entry *buffer = squashfs_get_fragment(inode->i_sb,
squashfs_i(inode)->fragment_block,
- squashfs_i(inode)->fragment_size);
+ squashfs_i(inode)->fragment_size, NULL);
int res = buffer->error;
if (res)
@@ -497,13 +498,13 @@ static int squashfs_read_folio(struct file *file, struct folio *folio)
return res;
}
-static int squashfs_readahead_fragment(struct page **page,
+static int squashfs_readahead_fragment(struct squashfs_context *context,
unsigned int pages, unsigned int expected)
{
- struct inode *inode = page[0]->mapping->host;
+ struct inode *inode = context->page[0]->mapping->host;
struct squashfs_cache_entry *buffer = squashfs_get_fragment(inode->i_sb,
squashfs_i(inode)->fragment_block,
- squashfs_i(inode)->fragment_size);
+ squashfs_i(inode)->fragment_size, context);
struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info;
unsigned int n, mask = (1 << (msblk->block_log - PAGE_SHIFT)) - 1;
int error = buffer->error;
@@ -514,18 +515,18 @@ static int squashfs_readahead_fragment(struct page **page,
expected += squashfs_i(inode)->fragment_offset;
for (n = 0; n < pages; n++) {
- unsigned int base = (page[n]->index & mask) << PAGE_SHIFT;
+ unsigned int base = (context->page[n]->index & mask) << PAGE_SHIFT;
unsigned int offset = base + squashfs_i(inode)->fragment_offset;
if (expected > offset) {
unsigned int avail = min_t(unsigned int, expected -
offset, PAGE_SIZE);
- squashfs_fill_page(page[n], buffer, offset, avail);
+ squashfs_fill_page(context->page[n], buffer, offset, avail);
}
- unlock_page(page[n]);
- put_page(page[n]);
+ unlock_page(context->page[n]);
+ put_page(context->page[n]);
}
out:
@@ -541,16 +542,15 @@ static void squashfs_readahead(struct readahead_control *ractl)
unsigned short shift = msblk->block_log - PAGE_SHIFT;
loff_t start = readahead_pos(ractl) & ~mask;
size_t len = readahead_length(ractl) + readahead_pos(ractl) - start;
- struct squashfs_page_actor *actor;
+ struct squashfs_context *context;
unsigned int nr_pages = 0;
- struct page **pages;
int i, file_end = i_size_read(inode) >> msblk->block_log;
unsigned int max_pages = 1UL << shift;
readahead_expand(ractl, start, (len | mask) + 1);
- pages = kmalloc_array(max_pages, sizeof(void *), GFP_KERNEL);
- if (!pages)
+ context = squashfs_get_context(msblk->contexts);
+ if (context == NULL)
return;
for (;;) {
@@ -566,21 +566,21 @@ static void squashfs_readahead(struct readahead_control *ractl)
max_pages = (expected + PAGE_SIZE - 1) >> PAGE_SHIFT;
- nr_pages = __readahead_batch(ractl, pages, max_pages);
+ nr_pages = __readahead_batch(ractl, context->page, max_pages);
if (!nr_pages)
break;
if (readahead_pos(ractl) >= i_size_read(inode))
goto skip_pages;
- index = pages[0]->index >> shift;
+ index = context->page[0]->index >> shift;
- if ((pages[nr_pages - 1]->index >> shift) != index)
+ if ((context->page[nr_pages - 1]->index >> shift) != index)
goto skip_pages;
if (index == file_end && squashfs_i(inode)->fragment_block !=
SQUASHFS_INVALID_BLK) {
- res = squashfs_readahead_fragment(pages, nr_pages,
+ res = squashfs_readahead_fragment(context, nr_pages,
expected);
if (res)
goto skip_pages;
@@ -591,14 +591,13 @@ static void squashfs_readahead(struct readahead_control *ractl)
if (bsize == 0)
goto skip_pages;
- actor = squashfs_page_actor_init_special(msblk, pages, nr_pages,
- expected);
- if (!actor)
- goto skip_pages;
+ squashfs_page_actor_init_special(context, msblk, nr_pages,
+ expected);
- res = squashfs_read_data(inode->i_sb, block, bsize, NULL, actor);
+ res = squashfs_read_data(inode->i_sb, block, bsize, NULL,
+ context);
- last_page = squashfs_page_actor_free(actor);
+ last_page = squashfs_page_actor_free(context->actor);
if (res == expected) {
int bytes;
@@ -610,26 +609,26 @@ static void squashfs_readahead(struct readahead_control *ractl)
PAGE_SIZE - bytes);
for (i = 0; i < nr_pages; i++) {
- flush_dcache_page(pages[i]);
- SetPageUptodate(pages[i]);
+ flush_dcache_page(context->page[i]);
+ SetPageUptodate(context->page[i]);
}
}
for (i = 0; i < nr_pages; i++) {
- unlock_page(pages[i]);
- put_page(pages[i]);
+ unlock_page(context->page[i]);
+ put_page(context->page[i]);
}
}
- kfree(pages);
+ squashfs_put_context(msblk->contexts, context);
return;
skip_pages:
for (i = 0; i < nr_pages; i++) {
- unlock_page(pages[i]);
- put_page(pages[i]);
+ unlock_page(context->page[i]);
+ put_page(context->page[i]);
}
- kfree(pages);
+ squashfs_put_context(msblk->contexts, context);
}
const struct address_space_operations squashfs_aops = {
diff --git a/fs/squashfs/file_direct.c b/fs/squashfs/file_direct.c
index f1ccad519e28..ab7b990cf71d 100644
--- a/fs/squashfs/file_direct.c
+++ b/fs/squashfs/file_direct.c
@@ -15,8 +15,9 @@
#include "squashfs_fs.h"
#include "squashfs_fs_sb.h"
#include "squashfs_fs_i.h"
-#include "squashfs.h"
+#include "read_context.h"
#include "page_actor.h"
+#include "squashfs.h"
/* Read separately compressed datablock directly into page cache */
int squashfs_readpage_block(struct page *target_page, u64 block, int bsize,
@@ -31,8 +32,7 @@ int squashfs_readpage_block(struct page *target_page, u64 block, int bsize,
int start_index = target_page->index & ~mask;
int end_index = start_index | mask;
int i, n, pages, bytes, res = -ENOMEM;
- struct page **page;
- struct squashfs_page_actor *actor;
+ struct squashfs_context *context;
void *pageaddr;
if (end_index > file_end)
@@ -40,21 +40,21 @@ int squashfs_readpage_block(struct page *target_page, u64 block, int bsize,
pages = end_index - start_index + 1;
- page = kmalloc_array(pages, sizeof(void *), GFP_KERNEL);
- if (page == NULL)
+ context = squashfs_get_context(msblk->contexts);
+ if (context == NULL)
return res;
/* Try to grab all the pages covered by the Squashfs block */
for (i = 0, n = start_index; n <= end_index; n++) {
- page[i] = (n == target_page->index) ? target_page :
+ context->page[i] = (n == target_page->index) ? target_page :
grab_cache_page_nowait(target_page->mapping, n);
- if (page[i] == NULL)
+ if (context->page[i] == NULL)
continue;
- if (PageUptodate(page[i])) {
- unlock_page(page[i]);
- put_page(page[i]);
+ if (PageUptodate(context->page[i])) {
+ unlock_page(context->page[i]);
+ put_page(context->page[i]);
continue;
}
@@ -64,17 +64,15 @@ int squashfs_readpage_block(struct page *target_page, u64 block, int bsize,
pages = i;
/*
- * Create a "page actor" which will kmap and kunmap the
+ * Initialise "page actor" which will kmap and kunmap the
* page cache pages appropriately within the decompressor
*/
- actor = squashfs_page_actor_init_special(msblk, page, pages, expected);
- if (actor == NULL)
- goto out;
+ squashfs_page_actor_init_special(context, msblk, pages, expected);
/* Decompress directly into the page cache buffers */
- res = squashfs_read_data(inode->i_sb, block, bsize, NULL, actor);
+ res = squashfs_read_data(inode->i_sb, block, bsize, NULL, context);
- squashfs_page_actor_free(actor);
+ squashfs_page_actor_free(context->actor);
if (res < 0)
goto mark_errored;
@@ -86,22 +84,22 @@ int squashfs_readpage_block(struct page *target_page, u64 block, int bsize,
/* Last page (if present) may have trailing bytes not filled */
bytes = res % PAGE_SIZE;
- if (page[pages - 1]->index == end_index && bytes) {
- pageaddr = kmap_local_page(page[pages - 1]);
+ if (context->page[pages - 1]->index == end_index && bytes) {
+ pageaddr = kmap_local_page(context->page[pages - 1]);
memset(pageaddr + bytes, 0, PAGE_SIZE - bytes);
kunmap_local(pageaddr);
}
/* Mark pages as uptodate, unlock and release */
for (i = 0; i < pages; i++) {
- flush_dcache_page(page[i]);
- SetPageUptodate(page[i]);
- unlock_page(page[i]);
- if (page[i] != target_page)
- put_page(page[i]);
+ flush_dcache_page(context->page[i]);
+ SetPageUptodate(context->page[i]);
+ unlock_page(context->page[i]);
+ if (context->page[i] != target_page)
+ put_page(context->page[i]);
}
- kfree(page);
+ squashfs_put_context(msblk->contexts, context);
return 0;
@@ -110,15 +108,14 @@ int squashfs_readpage_block(struct page *target_page, u64 block, int bsize,
* dealt with by the caller
*/
for (i = 0; i < pages; i++) {
- if (page[i] == NULL || page[i] == target_page)
+ if (context->page[i] == NULL || context->page[i] == target_page)
continue;
- flush_dcache_page(page[i]);
- SetPageError(page[i]);
- unlock_page(page[i]);
- put_page(page[i]);
+ flush_dcache_page(context->page[i]);
+ SetPageError(context->page[i]);
+ unlock_page(context->page[i]);
+ put_page(context->page[i]);
}
-out:
- kfree(page);
+ squashfs_put_context(msblk->contexts, context);
return res;
}
diff --git a/fs/squashfs/page_actor.c b/fs/squashfs/page_actor.c
index 81af6c4ca115..8ff50034afee 100644
--- a/fs/squashfs/page_actor.c
+++ b/fs/squashfs/page_actor.c
@@ -8,8 +8,9 @@
#include <linux/slab.h>
#include <linux/pagemap.h>
#include "squashfs_fs_sb.h"
-#include "decompressor.h"
+#include "read_context.h"
#include "page_actor.h"
+#include "decompressor.h"
/*
* This file contains implementations of page_actor for decompressing into
@@ -40,14 +41,9 @@ static void cache_finish_page(struct squashfs_page_actor *actor)
/* empty */
}
-struct squashfs_page_actor *squashfs_page_actor_init(void **buffer,
+void squashfs_page_actor_init(struct squashfs_page_actor *actor, void **buffer,
int pages, int length)
{
- struct squashfs_page_actor *actor = kmalloc(sizeof(*actor), GFP_KERNEL);
-
- if (actor == NULL)
- return NULL;
-
actor->length = length ? : pages * PAGE_SIZE;
actor->buffer = buffer;
actor->pages = pages;
@@ -56,7 +52,6 @@ struct squashfs_page_actor *squashfs_page_actor_init(void **buffer,
actor->squashfs_first_page = cache_first_page;
actor->squashfs_next_page = cache_next_page;
actor->squashfs_finish_page = cache_finish_page;
- return actor;
}
/* Implementation of page_actor for decompressing directly into page cache. */
@@ -102,35 +97,23 @@ static void direct_finish_page(struct squashfs_page_actor *actor)
kunmap_local(actor->pageaddr);
}
-struct squashfs_page_actor *squashfs_page_actor_init_special(struct squashfs_sb_info *msblk,
- struct page **page, int pages, int length)
+void squashfs_page_actor_init_special(struct squashfs_context *context,
+ struct squashfs_sb_info *msblk, int pages, int length)
{
- struct squashfs_page_actor *actor = kmalloc(sizeof(*actor), GFP_KERNEL);
-
- if (actor == NULL)
- return NULL;
-
- if (msblk->decompressor->alloc_buffer) {
- actor->tmp_buffer = kmalloc(PAGE_SIZE, GFP_KERNEL);
-
- if (actor->tmp_buffer == NULL) {
- kfree(actor);
- return NULL;
- }
- } else
- actor->tmp_buffer = NULL;
+ struct squashfs_page_actor *actor = context->actor;
+ actor->tmp_buffer = context->hole;
actor->length = length ? : pages * PAGE_SIZE;
- actor->page = page;
+ actor->page = context->page;
actor->pages = pages;
actor->next_page = 0;
actor->returned_pages = 0;
- actor->next_index = page[0]->index & ~((1 << (msblk->block_log - PAGE_SHIFT)) - 1);
+ actor->next_index = context->page[0]->index & ~((1 <<
+ (msblk->block_log - PAGE_SHIFT)) - 1);
actor->pageaddr = NULL;
actor->last_page = NULL;
actor->alloc_buffer = msblk->decompressor->alloc_buffer;
actor->squashfs_first_page = direct_first_page;
actor->squashfs_next_page = direct_next_page;
actor->squashfs_finish_page = direct_finish_page;
- return actor;
}
diff --git a/fs/squashfs/page_actor.h b/fs/squashfs/page_actor.h
index 97d4983559b1..36fbc93d8907 100644
--- a/fs/squashfs/page_actor.h
+++ b/fs/squashfs/page_actor.h
@@ -25,17 +25,15 @@ struct squashfs_page_actor {
pgoff_t next_index;
};
-extern struct squashfs_page_actor *squashfs_page_actor_init(void **buffer,
+extern void squashfs_page_actor_init(struct squashfs_page_actor *actor, void **buffer,
int pages, int length);
-extern struct squashfs_page_actor *squashfs_page_actor_init_special(
+extern void squashfs_page_actor_init_special(struct squashfs_context *context,
struct squashfs_sb_info *msblk,
- struct page **page, int pages, int length);
+ int pages, int length);
static inline struct page *squashfs_page_actor_free(struct squashfs_page_actor *actor)
{
struct page *last_page = actor->last_page;
- kfree(actor->tmp_buffer);
- kfree(actor);
return last_page;
}
static inline void *squashfs_first_page(struct squashfs_page_actor *actor)
diff --git a/fs/squashfs/read_context.c b/fs/squashfs/read_context.c
new file mode 100644
index 000000000000..35a52ed572d7
--- /dev/null
+++ b/fs/squashfs/read_context.c
@@ -0,0 +1,181 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Squashfs - a compressed read only filesystem for Linux
+ *
+ * Copyright (c) 2023
+ * Phillip Lougher <phillip@squashfs.org.uk>
+ *
+ * read_context.c
+ */
+
+#include <linux/fs.h>
+#include <linux/vfs.h>
+#include <linux/mutex.h>
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
+#include <linux/sched.h>
+#include <linux/wait.h>
+#include <linux/pagemap.h>
+#include <linux/bio.h>
+
+#include "squashfs_fs.h"
+#include "squashfs_fs_sb.h"
+#include "squashfs.h"
+#include "page_actor.h"
+
+
+void squashfs_delete_context(struct squashfs_context *context)
+{
+ int i;
+
+ for (i = 0; i < context->bio_pages; i++)
+ if (context->bio_page[i])
+ __free_page(context->bio_page[i]);
+
+ kfree(context->bio);
+ kfree(context->page);
+ kfree(context->actor);
+ kfree(context->hole);
+ kfree(context);
+}
+
+
+struct squashfs_context *squashfs_create_context(int length, int alloc_hole,
+ int alloc_array)
+{
+ int pages = (length + PAGE_SIZE - 1) >> PAGE_SHIFT;
+ int bio_pages = pages + 1, i;
+ struct squashfs_context *context;
+
+ context = kzalloc(sizeof(struct squashfs_context), GFP_KERNEL);
+ if (context == NULL)
+ return ERR_PTR(-ENOMEM);
+
+ context->bio_page = kcalloc(bio_pages, sizeof(struct bio_pages *),
+ GFP_KERNEL);
+
+ if (context->bio_page == NULL)
+ goto cleanup;
+
+ context->bio_pages = bio_pages;
+
+ for (i = 0; i < bio_pages; i++) {
+ context->bio_page[i] = alloc_page(GFP_KERNEL);
+
+ if (context->bio_page[i] == NULL)
+ goto cleanup;
+ }
+
+ context->bio = bio_kmalloc(bio_pages, GFP_KERNEL);
+ if (context->bio == NULL)
+ goto cleanup;
+
+ context->actor = kmalloc(sizeof(struct squashfs_page_actor),
+ GFP_KERNEL);
+ if (context->actor == NULL)
+ goto cleanup;
+
+ if (alloc_array) {
+ context->page = kmalloc_array(pages, sizeof(struct page *),
+ GFP_KERNEL);
+ if (context->page == NULL)
+ goto cleanup;
+ }
+
+ if (alloc_hole) {
+ context->hole = kmalloc(PAGE_SIZE, GFP_KERNEL);
+ if (context->hole == NULL)
+ goto cleanup;
+ }
+
+ return context;
+
+cleanup:
+ squashfs_delete_context(context);
+ return ERR_PTR(-ENOMEM);
+}
+
+
+struct squashfs_context *squashfs_get_context(struct squashfs_contexts *contexts)
+{
+ struct squashfs_context *context;
+
+ while (1) {
+ mutex_lock(&contexts->mutex);
+
+ if (!list_empty(&contexts->list)) {
+ context = list_entry(contexts->list.prev,
+ struct squashfs_context, list);
+ list_del(&context->list);
+ mutex_unlock(&contexts->mutex);
+ break;
+ }
+
+ mutex_unlock(&contexts->mutex);
+ wait_event(contexts->wait_queue, !list_empty(&contexts->list));
+ }
+
+ return context;
+}
+
+
+void squashfs_put_context(struct squashfs_contexts *contexts,
+ struct squashfs_context *context)
+{
+
+ mutex_lock(&contexts->mutex);
+ list_add(&context->list, &contexts->list);
+ mutex_unlock(&contexts->mutex);
+ wake_up(&contexts->wait_queue);
+}
+
+
+void squashfs_delete_contexts(struct squashfs_contexts *contexts)
+{
+ struct squashfs_context *context;
+
+ if (contexts == NULL)
+ return;
+
+ while (!list_empty(&contexts->list)) {
+ context = list_entry(contexts->list.prev,
+ struct squashfs_context, list);
+ list_del(&context->list);
+ squashfs_delete_context(context);
+ }
+
+ kfree(contexts);
+}
+
+
+struct squashfs_contexts *squashfs_create_contexts(int block_size, int entries,
+ int alloc_hole, int alloc_array)
+{
+ int i;
+ struct squashfs_contexts *contexts;
+
+ contexts = kzalloc(sizeof(struct squashfs_contexts), GFP_KERNEL);
+ if (contexts == NULL)
+ return ERR_PTR(-ENOMEM);
+
+ mutex_init(&contexts->mutex);
+ INIT_LIST_HEAD(&contexts->list);
+ init_waitqueue_head(&contexts->wait_queue);
+
+ for (i = 0; i < entries; i++) {
+ struct squashfs_context *context;
+
+ context = squashfs_create_context(block_size, alloc_hole,
+ alloc_array);
+ if (context == NULL)
+ goto cleanup;
+
+ list_add(&context->list, &contexts->list);
+ }
+
+ return contexts;
+
+cleanup:
+ squashfs_delete_contexts(contexts);
+ return ERR_PTR(-ENOMEM);
+}
diff --git a/fs/squashfs/read_context.h b/fs/squashfs/read_context.h
new file mode 100644
index 000000000000..a9b93a43893b
--- /dev/null
+++ b/fs/squashfs/read_context.h
@@ -0,0 +1,31 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+#ifndef READ_CONTEXT_H
+#define READ_CONTEXT_H
+/*
+ * Copyright (c) 2023
+ * Phillip Lougher <phillip@squashfs.org.uk>
+ */
+
+struct squashfs_context {
+ int bio_pages;
+ struct bio *bio;
+ struct page **bio_page;
+ struct squashfs_page_actor *actor;
+ struct page **page;
+ void *hole;
+ struct list_head list;
+};
+
+struct squashfs_contexts {
+ struct mutex mutex;
+ struct list_head list;
+ wait_queue_head_t wait_queue;
+};
+
+extern void squashfs_delete_context(struct squashfs_context *context);
+extern struct squashfs_context *squashfs_create_context(int length, int alloc_hole, int alloc_array);
+extern struct squashfs_context *squashfs_get_context(struct squashfs_contexts *contexts);
+extern void squashfs_put_context(struct squashfs_contexts *contexts, struct squashfs_context *context);
+extern void squashfs_delete_contexts(struct squashfs_contexts *contexts);
+extern struct squashfs_contexts *squashfs_create_contexts(int block_size, int entries, int alloc_hole, int alloc_array);
+#endif
diff --git a/fs/squashfs/squashfs.h b/fs/squashfs/squashfs.h
index a6164fdf9435..c5db13907e9f 100644
--- a/fs/squashfs/squashfs.h
+++ b/fs/squashfs/squashfs.h
@@ -8,6 +8,9 @@
* squashfs.h
*/
+#include "read_context.h"
+#include "page_actor.h"
+
#define TRACE(s, args...) pr_debug("SQUASHFS: "s, ## args)
#define ERROR(s, args...) pr_err("SQUASHFS error: "s, ## args)
@@ -16,19 +19,20 @@
/* block.c */
extern int squashfs_read_data(struct super_block *, u64, int, u64 *,
- struct squashfs_page_actor *);
+ struct squashfs_context *);
/* cache.c */
-extern struct squashfs_cache *squashfs_cache_init(char *, int, int);
+extern struct squashfs_cache *squashfs_cache_init(char *, int, int, int);
extern void squashfs_cache_delete(struct squashfs_cache *);
extern struct squashfs_cache_entry *squashfs_cache_get(struct super_block *,
- struct squashfs_cache *, u64, int);
+ struct squashfs_cache *, struct squashfs_contexts *,
+ struct squashfs_context *, u64, int);
extern void squashfs_cache_put(struct squashfs_cache_entry *);
extern int squashfs_copy_data(void *, struct squashfs_cache_entry *, int, int);
extern int squashfs_read_metadata(struct super_block *, void *, u64 *,
int *, int);
extern struct squashfs_cache_entry *squashfs_get_fragment(struct super_block *,
- u64, int);
+ u64, int, struct squashfs_context *);
extern struct squashfs_cache_entry *squashfs_get_datablock(struct super_block *,
u64, int);
extern void *squashfs_read_table(struct super_block *, u64, int);
diff --git a/fs/squashfs/squashfs_fs_sb.h b/fs/squashfs/squashfs_fs_sb.h
index 72f6f4b37863..6cee243be9bb 100644
--- a/fs/squashfs/squashfs_fs_sb.h
+++ b/fs/squashfs/squashfs_fs_sb.h
@@ -23,6 +23,7 @@ struct squashfs_cache {
int pages;
spinlock_t lock;
wait_queue_head_t wait_queue;
+ struct squashfs_contexts *contexts;
struct squashfs_cache_entry *entry;
};
@@ -37,7 +38,6 @@ struct squashfs_cache_entry {
wait_queue_head_t wait_queue;
struct squashfs_cache *cache;
void **data;
- struct squashfs_page_actor *actor;
};
struct squashfs_sb_info {
@@ -47,6 +47,7 @@ struct squashfs_sb_info {
struct squashfs_cache *block_cache;
struct squashfs_cache *fragment_cache;
struct squashfs_cache *read_page;
+ struct squashfs_contexts *contexts;
int next_meta_index;
__le64 *id_table;
__le64 *fragment_index;
diff --git a/fs/squashfs/super.c b/fs/squashfs/super.c
index e090fae48e68..3ba5c2b5c9e0 100644
--- a/fs/squashfs/super.c
+++ b/fs/squashfs/super.c
@@ -33,6 +33,7 @@
#include "squashfs_fs.h"
#include "squashfs_fs_sb.h"
#include "squashfs_fs_i.h"
+#include "read_context.h"
#include "squashfs.h"
#include "decompressor.h"
#include "xattr.h"
@@ -317,18 +318,23 @@ static int squashfs_fill_super(struct super_block *sb, struct fs_context *fc)
err = -ENOMEM;
msblk->block_cache = squashfs_cache_init("metadata",
- SQUASHFS_CACHED_BLKS, SQUASHFS_METADATA_SIZE);
+ SQUASHFS_CACHED_BLKS, SQUASHFS_METADATA_SIZE, 1);
if (msblk->block_cache == NULL)
goto failed_mount;
/* Allocate read_page block */
msblk->read_page = squashfs_cache_init("data",
- msblk->max_thread_num, msblk->block_size);
+ msblk->max_thread_num, msblk->block_size, 0);
if (msblk->read_page == NULL) {
errorf(fc, "Failed to allocate read_page block");
goto failed_mount;
}
+ msblk->contexts = squashfs_create_contexts(msblk->block_size,
+ msblk->max_thread_num, msblk->decompressor->alloc_buffer, 1);
+ if (msblk->contexts == NULL)
+ goto failed_mount;
+
msblk->stream = squashfs_decompressor_setup(sb, flags);
if (IS_ERR(msblk->stream)) {
err = PTR_ERR(msblk->stream);
@@ -392,7 +398,7 @@ static int squashfs_fill_super(struct super_block *sb, struct fs_context *fc)
goto check_directory_table;
msblk->fragment_cache = squashfs_cache_init("fragment",
- SQUASHFS_CACHED_FRAGMENTS, msblk->block_size);
+ SQUASHFS_CACHED_FRAGMENTS, msblk->block_size, 0);
if (msblk->fragment_cache == NULL) {
err = -ENOMEM;
goto failed_mount;
@@ -454,6 +460,7 @@ static int squashfs_fill_super(struct super_block *sb, struct fs_context *fc)
squashfs_cache_delete(msblk->block_cache);
squashfs_cache_delete(msblk->fragment_cache);
squashfs_cache_delete(msblk->read_page);
+ squashfs_delete_contexts(msblk->contexts);
msblk->thread_ops->destroy(msblk);
kfree(msblk->inode_lookup_table);
kfree(msblk->fragment_index);
diff --git a/fs/squashfs/symlink.c b/fs/squashfs/symlink.c
index 2bf977a52c2c..8471a3a5491b 100644
--- a/fs/squashfs/symlink.c
+++ b/fs/squashfs/symlink.c
@@ -69,7 +69,8 @@ static int squashfs_symlink_read_folio(struct file *file, struct folio *folio)
* blocks, we may need to call squashfs_cache_get multiple times.
*/
for (bytes = 0; bytes < length; offset = 0, bytes += copied) {
- entry = squashfs_cache_get(sb, msblk->block_cache, block, 0);
+ entry = squashfs_cache_get(sb, msblk->block_cache, NULL, NULL,
+ block, 0);
if (entry->error) {
ERROR("Unable to read symlink [%llx:%x]\n",
squashfs_i(inode)->start,
--
2.35.1
next prev parent reply other threads:[~2023-05-07 21:07 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-04-26 5:10 [PATCH 0/1] mm/oom_kill: system enters a state something like hang when running stress-ng Hui Wang
2023-04-26 5:10 ` [PATCH 1/1] mm/oom_kill: trigger the oom killer if oom occurs without __GFP_FS Hui Wang
2023-04-26 8:33 ` Michal Hocko
2023-04-26 11:07 ` Hui Wang
2023-04-26 16:44 ` Phillip Lougher
2023-04-26 17:38 ` Phillip Lougher
2023-04-26 18:26 ` Yang Shi
2023-04-26 19:06 ` Phillip Lougher
2023-04-26 19:34 ` Phillip Lougher
2023-04-27 0:42 ` Hui Wang
2023-04-27 1:37 ` Phillip Lougher
2023-04-27 5:22 ` Hui Wang
2023-04-27 1:18 ` Gao Xiang
2023-04-27 3:47 ` Hui Wang
2023-04-27 4:17 ` Gao Xiang
2023-04-27 7:03 ` Colin King (gmail)
2023-04-27 7:49 ` Hui Wang
2023-04-28 19:53 ` Michal Hocko
2023-05-03 11:49 ` Hui Wang
2023-05-03 12:20 ` Michal Hocko
2023-05-03 18:41 ` Phillip Lougher
2023-05-03 19:10 ` Phillip Lougher
2023-05-03 19:38 ` Hui Wang
2023-05-07 21:07 ` Phillip Lougher [this message]
2023-05-08 10:05 ` Hui Wang
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=b8f57b19-89f5-6a17-d5de-8ab488b96ac6@squashfs.org.uk \
--to=phillip@squashfs.org.uk \
--cc=akpm@linux-foundation.org \
--cc=colin.i.king@gmail.com \
--cc=hannes@cmpxchg.org \
--cc=hch@infradead.org \
--cc=hsiangkao@linux.alibaba.com \
--cc=hui.wang@canonical.com \
--cc=linux-mm@kvack.org \
--cc=mgorman@suse.de \
--cc=mhocko@suse.com \
--cc=shy828301@gmail.com \
--cc=surenb@google.com \
--cc=vbabka@suse.cz \
/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