* [RFC PATCH 2/2] erofs: dax support for non-tailpacking regular file
@ 2021-07-04 13:50 ` Gao Xiang
0 siblings, 0 replies; 13+ messages in thread
From: Gao Xiang @ 2021-07-04 13:50 UTC (permalink / raw)
To: linux-erofs
Cc: nvdimm, Darrick J. Wong, LKML, Joseph Qi, Liu Bo, linux-fsdevel,
Gao Xiang, Liu Jiang
DAX is quite useful for some VM use cases in order to save guest
memory extremely with minimal lightweight EROFS.
In order to prepare for such use cases, add preliminary dax support
for non-tailpacking regular files for now.
Tested with the DRAM-emulated PMEM and the EROFS image generated by
"mkfs.erofs -Enoinline_data enwik9.fsdax.img enwik9"
Cc: nvdimm@lists.linux.dev
Cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
---
fs/erofs/data.c | 41 ++++++++++++++++++++++++++++++++++++++++-
fs/erofs/inode.c | 5 +++++
fs/erofs/internal.h | 2 ++
fs/erofs/super.c | 20 ++++++++++++++++++--
4 files changed, 65 insertions(+), 3 deletions(-)
diff --git a/fs/erofs/data.c b/fs/erofs/data.c
index 38e9439c2510..edeee8ccb22a 100644
--- a/fs/erofs/data.c
+++ b/fs/erofs/data.c
@@ -6,7 +6,7 @@
#include "internal.h"
#include <linux/prefetch.h>
#include <linux/iomap.h>
-
+#include <linux/dax.h>
#include <trace/events/erofs.h>
static void erofs_readendio(struct bio *bio)
@@ -323,6 +323,7 @@ static int erofs_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
return ret;
iomap->bdev = inode->i_sb->s_bdev;
+ iomap->dax_dev = EROFS_I_SB(inode)->dax_dev;
iomap->offset = map.m_la;
iomap->length = map.m_llen;
@@ -382,6 +383,11 @@ static ssize_t erofs_file_read_iter(struct kiocb *iocb, struct iov_iter *to)
if (!iov_iter_count(to))
return 0;
+#ifdef CONFIG_FS_DAX
+ if (IS_DAX(iocb->ki_filp->f_mapping->host))
+ return dax_iomap_rw(iocb, to, &erofs_iomap_ops);
+#endif
+
if (iocb->ki_flags & IOCB_DIRECT) {
int err = erofs_prepare_dio(iocb, to);
@@ -410,6 +416,39 @@ const struct address_space_operations erofs_raw_access_aops = {
.direct_IO = noop_direct_IO,
};
+#ifdef CONFIG_FS_DAX
+static vm_fault_t erofs_dax_huge_fault(struct vm_fault *vmf,
+ enum page_entry_size pe_size)
+{
+ return dax_iomap_fault(vmf, pe_size, NULL, NULL, &erofs_iomap_ops);
+}
+
+static vm_fault_t erofs_dax_fault(struct vm_fault *vmf)
+{
+ return erofs_dax_huge_fault(vmf, PE_SIZE_PTE);
+}
+
+static const struct vm_operations_struct erofs_dax_vm_ops = {
+ .fault = erofs_dax_fault,
+ .huge_fault = erofs_dax_huge_fault,
+};
+
+int erofs_file_mmap(struct file *file, struct vm_area_struct *vma)
+{
+ if (!IS_DAX(file_inode(file)))
+ return generic_file_readonly_mmap(file, vma);
+
+ if ((vma->vm_flags & VM_SHARED) && (vma->vm_flags & VM_MAYWRITE))
+ return -EINVAL;
+
+ vma->vm_ops = &erofs_dax_vm_ops;
+ vma->vm_flags |= VM_HUGEPAGE;
+ return 0;
+}
+#else
+#define erofs_file_mmap generic_file_readonly_mmap
+#endif
+
const struct file_operations erofs_file_fops = {
.llseek = generic_file_llseek,
.read_iter = erofs_file_read_iter,
diff --git a/fs/erofs/inode.c b/fs/erofs/inode.c
index 00edb7562fea..695b97acb9a6 100644
--- a/fs/erofs/inode.c
+++ b/fs/erofs/inode.c
@@ -174,6 +174,11 @@ static struct page *erofs_read_inode(struct inode *inode,
inode->i_mtime.tv_nsec = inode->i_ctime.tv_nsec;
inode->i_atime.tv_nsec = inode->i_ctime.tv_nsec;
+ inode->i_flags &= ~S_DAX;
+ if (test_opt(&sbi->ctx, DAX) && S_ISREG(inode->i_mode) &&
+ vi->datalayout == EROFS_INODE_FLAT_PLAIN)
+ inode->i_flags |= S_DAX;
+
if (!nblks)
/* measure inode.i_blocks as generic filesystems */
inode->i_blocks = roundup(inode->i_size, EROFS_BLKSIZ) >> 9;
diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h
index 2669c785d548..8b0542d35148 100644
--- a/fs/erofs/internal.h
+++ b/fs/erofs/internal.h
@@ -83,6 +83,7 @@ struct erofs_sb_info {
struct erofs_sb_lz4_info lz4;
#endif /* CONFIG_EROFS_FS_ZIP */
+ struct dax_device *dax_dev;
u32 blocks;
u32 meta_blkaddr;
#ifdef CONFIG_EROFS_FS_XATTR
@@ -115,6 +116,7 @@ struct erofs_sb_info {
/* Mount flags set via mount options or defaults */
#define EROFS_MOUNT_XATTR_USER 0x00000010
#define EROFS_MOUNT_POSIX_ACL 0x00000020
+#define EROFS_MOUNT_DAX 0x00000040
#define clear_opt(ctx, option) ((ctx)->mount_opt &= ~EROFS_MOUNT_##option)
#define set_opt(ctx, option) ((ctx)->mount_opt |= EROFS_MOUNT_##option)
diff --git a/fs/erofs/super.c b/fs/erofs/super.c
index 8fc6c04b54f4..9aa385e22e1a 100644
--- a/fs/erofs/super.c
+++ b/fs/erofs/super.c
@@ -11,6 +11,7 @@
#include <linux/crc32c.h>
#include <linux/fs_context.h>
#include <linux/fs_parser.h>
+#include <linux/dax.h>
#include "xattr.h"
#define CREATE_TRACE_POINTS
@@ -355,6 +356,7 @@ enum {
Opt_user_xattr,
Opt_acl,
Opt_cache_strategy,
+ Opt_dax,
Opt_err
};
@@ -370,6 +372,7 @@ static const struct fs_parameter_spec erofs_fs_parameters[] = {
fsparam_flag_no("acl", Opt_acl),
fsparam_enum("cache_strategy", Opt_cache_strategy,
erofs_param_cache_strategy),
+ fsparam_flag("dax", Opt_dax),
{}
};
@@ -410,6 +413,14 @@ static int erofs_fc_parse_param(struct fs_context *fc,
ctx->cache_strategy = result.uint_32;
#else
errorfc(fc, "compression not supported, cache_strategy ignored");
+#endif
+ break;
+ case Opt_dax:
+#ifdef CONFIG_FS_DAX
+ warnfc(fc, "DAX enabled. Warning: EXPERIMENTAL, use at your own risk");
+ set_opt(ctx, DAX);
+#else
+ errorfc(fc, "dax options not supported");
#endif
break;
default:
@@ -496,6 +507,7 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)
return -ENOMEM;
sb->s_fs_info = sbi;
+ sbi->dax_dev = fs_dax_get_by_bdev(sb->s_bdev);
err = erofs_read_superblock(sb);
if (err)
return err;
@@ -609,6 +621,8 @@ static void erofs_kill_sb(struct super_block *sb)
sbi = EROFS_SB(sb);
if (!sbi)
return;
+ if (sbi->dax_dev)
+ fs_put_dax(sbi->dax_dev);
kfree(sbi);
sb->s_fs_info = NULL;
}
@@ -711,8 +725,8 @@ static int erofs_statfs(struct dentry *dentry, struct kstatfs *buf)
static int erofs_show_options(struct seq_file *seq, struct dentry *root)
{
- struct erofs_sb_info *sbi __maybe_unused = EROFS_SB(root->d_sb);
- struct erofs_fs_context *ctx __maybe_unused = &sbi->ctx;
+ struct erofs_sb_info *sbi = EROFS_SB(root->d_sb);
+ struct erofs_fs_context *ctx = &sbi->ctx;
#ifdef CONFIG_EROFS_FS_XATTR
if (test_opt(ctx, XATTR_USER))
@@ -734,6 +748,8 @@ static int erofs_show_options(struct seq_file *seq, struct dentry *root)
else if (ctx->cache_strategy == EROFS_ZIP_CACHE_READAROUND)
seq_puts(seq, ",cache_strategy=readaround");
#endif
+ if (test_opt(ctx, DAX))
+ seq_puts(seq, ",dax");
return 0;
}
--
2.24.4
^ permalink raw reply related [flat|nested] 13+ messages in thread
* Re: [RFC PATCH 2/2] erofs: dax support for non-tailpacking regular file
2021-07-04 13:50 ` Gao Xiang
(?)
@ 2021-07-04 16:00 ` kernel test robot
-1 siblings, 0 replies; 13+ messages in thread
From: kernel test robot @ 2021-07-04 16:00 UTC (permalink / raw)
To: kbuild-all
[-- Attachment #1: Type: text/plain, Size: 2812 bytes --]
Hi Gao,
[FYI, it's a private test report for your RFC patch.]
[auto build test WARNING on xiang-erofs/dev-test]
[also build test WARNING on v5.13 next-20210701]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Gao-Xiang/erofs-dio-dax-support-for-non-tailpacking-cases/20210704-215229
base: https://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs.git dev-test
config: csky-randconfig-r005-20210704 (attached as .config)
compiler: csky-linux-gcc (GCC) 9.3.0
reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# https://github.com/0day-ci/linux/commit/e5df28a5e6dbbcf409d0f7a7976a6263ac3de64a
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Gao-Xiang/erofs-dio-dax-support-for-non-tailpacking-cases/20210704-215229
git checkout e5df28a5e6dbbcf409d0f7a7976a6263ac3de64a
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross ARCH=csky
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
All warnings (new ones prefixed by >>):
>> fs/erofs/data.c:436:5: warning: no previous prototype for 'erofs_file_mmap' [-Wmissing-prototypes]
436 | int erofs_file_mmap(struct file *file, struct vm_area_struct *vma)
| ^~~~~~~~~~~~~~~
Kconfig warnings: (for reference only)
WARNING: unmet direct dependencies detected for LOCKDEP
Depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT && (FRAME_POINTER || MIPS || PPC || S390 || MICROBLAZE || ARM || ARC || X86)
Selected by
- PROVE_LOCKING && DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT
- LOCK_STAT && DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT
- DEBUG_LOCK_ALLOC && DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT
vim +/erofs_file_mmap +436 fs/erofs/data.c
435
> 436 int erofs_file_mmap(struct file *file, struct vm_area_struct *vma)
437 {
438 if (!IS_DAX(file_inode(file)))
439 return generic_file_readonly_mmap(file, vma);
440
441 if ((vma->vm_flags & VM_SHARED) && (vma->vm_flags & VM_MAYWRITE))
442 return -EINVAL;
443
444 vma->vm_ops = &erofs_dax_vm_ops;
445 vma->vm_flags |= VM_HUGEPAGE;
446 return 0;
447 }
448 #else
449 #define erofs_file_mmap generic_file_readonly_mmap
450 #endif
451
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org
[-- Attachment #2: config.gz --]
[-- Type: application/gzip, Size: 23816 bytes --]
^ permalink raw reply [flat|nested] 13+ messages in thread
* [RFC PATCH v1.1 2/2] erofs: dax support for non-tailpacking regular file
2021-07-04 13:50 ` Gao Xiang
@ 2021-07-05 13:21 ` Gao Xiang
-1 siblings, 0 replies; 13+ messages in thread
From: Gao Xiang @ 2021-07-05 13:21 UTC (permalink / raw)
To: linux-erofs
Cc: linux-fsdevel, LKML, nvdimm, Liu Bo, Darrick J. Wong, Joseqh Qi,
Liu Jiang, Gao Xiang
DAX is quite useful for some VM use cases in order to save guest
memory extremely with minimal lightweight EROFS.
In order to prepare for such use cases, add preliminary dax support
for non-tailpacking regular files for now.
Tested with the DRAM-emulated PMEM and the EROFS image generated by
"mkfs.erofs -Enoinline_data enwik9.fsdax.img enwik9"
Cc: nvdimm@lists.linux.dev
Cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
---
change since v1:
- update missing hunks due to patch spliting...
bdev_dax_supported(...)
erofs_file_mmap(...)
fs/erofs/data.c | 43 +++++++++++++++++++++++++++++++++++++++++--
fs/erofs/inode.c | 5 +++++
fs/erofs/internal.h | 2 ++
fs/erofs/super.c | 26 ++++++++++++++++++++++++--
4 files changed, 72 insertions(+), 4 deletions(-)
diff --git a/fs/erofs/data.c b/fs/erofs/data.c
index 0f82b4cb474c..c188c629be45 100644
--- a/fs/erofs/data.c
+++ b/fs/erofs/data.c
@@ -6,7 +6,7 @@
#include "internal.h"
#include <linux/prefetch.h>
#include <linux/iomap.h>
-
+#include <linux/dax.h>
#include <trace/events/erofs.h>
static void erofs_readendio(struct bio *bio)
@@ -323,6 +323,7 @@ static int erofs_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
return ret;
iomap->bdev = inode->i_sb->s_bdev;
+ iomap->dax_dev = EROFS_I_SB(inode)->dax_dev;
iomap->offset = map.m_la;
iomap->length = map.m_llen;
@@ -382,6 +383,11 @@ static ssize_t erofs_file_read_iter(struct kiocb *iocb, struct iov_iter *to)
if (!iov_iter_count(to))
return 0;
+#ifdef CONFIG_FS_DAX
+ if (IS_DAX(iocb->ki_filp->f_mapping->host))
+ return dax_iomap_rw(iocb, to, &erofs_iomap_ops);
+#endif
+
if (iocb->ki_flags & IOCB_DIRECT) {
int err = erofs_prepare_dio(iocb, to);
@@ -410,9 +416,42 @@ const struct address_space_operations erofs_raw_access_aops = {
.direct_IO = noop_direct_IO,
};
+#ifdef CONFIG_FS_DAX
+static vm_fault_t erofs_dax_huge_fault(struct vm_fault *vmf,
+ enum page_entry_size pe_size)
+{
+ return dax_iomap_fault(vmf, pe_size, NULL, NULL, &erofs_iomap_ops);
+}
+
+static vm_fault_t erofs_dax_fault(struct vm_fault *vmf)
+{
+ return erofs_dax_huge_fault(vmf, PE_SIZE_PTE);
+}
+
+static const struct vm_operations_struct erofs_dax_vm_ops = {
+ .fault = erofs_dax_fault,
+ .huge_fault = erofs_dax_huge_fault,
+};
+
+static int erofs_file_mmap(struct file *file, struct vm_area_struct *vma)
+{
+ if (!IS_DAX(file_inode(file)))
+ return generic_file_readonly_mmap(file, vma);
+
+ if ((vma->vm_flags & VM_SHARED) && (vma->vm_flags & VM_MAYWRITE))
+ return -EINVAL;
+
+ vma->vm_ops = &erofs_dax_vm_ops;
+ vma->vm_flags |= VM_HUGEPAGE;
+ return 0;
+}
+#else
+#define erofs_file_mmap generic_file_readonly_mmap
+#endif
+
const struct file_operations erofs_file_fops = {
.llseek = generic_file_llseek,
.read_iter = erofs_file_read_iter,
- .mmap = generic_file_readonly_mmap,
+ .mmap = erofs_file_mmap,
.splice_read = generic_file_splice_read,
};
diff --git a/fs/erofs/inode.c b/fs/erofs/inode.c
index 00edb7562fea..695b97acb9a6 100644
--- a/fs/erofs/inode.c
+++ b/fs/erofs/inode.c
@@ -174,6 +174,11 @@ static struct page *erofs_read_inode(struct inode *inode,
inode->i_mtime.tv_nsec = inode->i_ctime.tv_nsec;
inode->i_atime.tv_nsec = inode->i_ctime.tv_nsec;
+ inode->i_flags &= ~S_DAX;
+ if (test_opt(&sbi->ctx, DAX) && S_ISREG(inode->i_mode) &&
+ vi->datalayout == EROFS_INODE_FLAT_PLAIN)
+ inode->i_flags |= S_DAX;
+
if (!nblks)
/* measure inode.i_blocks as generic filesystems */
inode->i_blocks = roundup(inode->i_size, EROFS_BLKSIZ) >> 9;
diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h
index 2669c785d548..8b0542d35148 100644
--- a/fs/erofs/internal.h
+++ b/fs/erofs/internal.h
@@ -83,6 +83,7 @@ struct erofs_sb_info {
struct erofs_sb_lz4_info lz4;
#endif /* CONFIG_EROFS_FS_ZIP */
+ struct dax_device *dax_dev;
u32 blocks;
u32 meta_blkaddr;
#ifdef CONFIG_EROFS_FS_XATTR
@@ -115,6 +116,7 @@ struct erofs_sb_info {
/* Mount flags set via mount options or defaults */
#define EROFS_MOUNT_XATTR_USER 0x00000010
#define EROFS_MOUNT_POSIX_ACL 0x00000020
+#define EROFS_MOUNT_DAX 0x00000040
#define clear_opt(ctx, option) ((ctx)->mount_opt &= ~EROFS_MOUNT_##option)
#define set_opt(ctx, option) ((ctx)->mount_opt |= EROFS_MOUNT_##option)
diff --git a/fs/erofs/super.c b/fs/erofs/super.c
index 8fc6c04b54f4..b44a964ab24f 100644
--- a/fs/erofs/super.c
+++ b/fs/erofs/super.c
@@ -11,6 +11,7 @@
#include <linux/crc32c.h>
#include <linux/fs_context.h>
#include <linux/fs_parser.h>
+#include <linux/dax.h>
#include "xattr.h"
#define CREATE_TRACE_POINTS
@@ -355,6 +356,7 @@ enum {
Opt_user_xattr,
Opt_acl,
Opt_cache_strategy,
+ Opt_dax,
Opt_err
};
@@ -370,6 +372,7 @@ static const struct fs_parameter_spec erofs_fs_parameters[] = {
fsparam_flag_no("acl", Opt_acl),
fsparam_enum("cache_strategy", Opt_cache_strategy,
erofs_param_cache_strategy),
+ fsparam_flag("dax", Opt_dax),
{}
};
@@ -410,6 +413,14 @@ static int erofs_fc_parse_param(struct fs_context *fc,
ctx->cache_strategy = result.uint_32;
#else
errorfc(fc, "compression not supported, cache_strategy ignored");
+#endif
+ break;
+ case Opt_dax:
+#ifdef CONFIG_FS_DAX
+ warnfc(fc, "DAX enabled. Warning: EXPERIMENTAL, use at your own risk");
+ set_opt(ctx, DAX);
+#else
+ errorfc(fc, "dax options not supported");
#endif
break;
default:
@@ -496,10 +507,17 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)
return -ENOMEM;
sb->s_fs_info = sbi;
+ sbi->dax_dev = fs_dax_get_by_bdev(sb->s_bdev);
err = erofs_read_superblock(sb);
if (err)
return err;
+ if (test_opt(ctx, DAX) &&
+ !bdev_dax_supported(sb->s_bdev, EROFS_BLKSIZ)) {
+ errorfc(fc, "DAX unsupported by block device. Turning off DAX.");
+ clear_opt(ctx, DAX);
+ }
+
sb->s_flags |= SB_RDONLY | SB_NOATIME;
sb->s_maxbytes = MAX_LFS_FILESIZE;
sb->s_time_gran = 1;
@@ -609,6 +627,8 @@ static void erofs_kill_sb(struct super_block *sb)
sbi = EROFS_SB(sb);
if (!sbi)
return;
+ if (sbi->dax_dev)
+ fs_put_dax(sbi->dax_dev);
kfree(sbi);
sb->s_fs_info = NULL;
}
@@ -711,8 +731,8 @@ static int erofs_statfs(struct dentry *dentry, struct kstatfs *buf)
static int erofs_show_options(struct seq_file *seq, struct dentry *root)
{
- struct erofs_sb_info *sbi __maybe_unused = EROFS_SB(root->d_sb);
- struct erofs_fs_context *ctx __maybe_unused = &sbi->ctx;
+ struct erofs_sb_info *sbi = EROFS_SB(root->d_sb);
+ struct erofs_fs_context *ctx = &sbi->ctx;
#ifdef CONFIG_EROFS_FS_XATTR
if (test_opt(ctx, XATTR_USER))
@@ -734,6 +754,8 @@ static int erofs_show_options(struct seq_file *seq, struct dentry *root)
else if (ctx->cache_strategy == EROFS_ZIP_CACHE_READAROUND)
seq_puts(seq, ",cache_strategy=readaround");
#endif
+ if (test_opt(ctx, DAX))
+ seq_puts(seq, ",dax");
return 0;
}
--
2.24.4
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [RFC PATCH v1.1 2/2] erofs: dax support for non-tailpacking regular file
@ 2021-07-05 13:21 ` Gao Xiang
0 siblings, 0 replies; 13+ messages in thread
From: Gao Xiang @ 2021-07-05 13:21 UTC (permalink / raw)
To: linux-erofs
Cc: nvdimm, Darrick J. Wong, LKML, Joseqh Qi, Liu Bo, linux-fsdevel,
Gao Xiang, Liu Jiang
DAX is quite useful for some VM use cases in order to save guest
memory extremely with minimal lightweight EROFS.
In order to prepare for such use cases, add preliminary dax support
for non-tailpacking regular files for now.
Tested with the DRAM-emulated PMEM and the EROFS image generated by
"mkfs.erofs -Enoinline_data enwik9.fsdax.img enwik9"
Cc: nvdimm@lists.linux.dev
Cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
---
change since v1:
- update missing hunks due to patch spliting...
bdev_dax_supported(...)
erofs_file_mmap(...)
fs/erofs/data.c | 43 +++++++++++++++++++++++++++++++++++++++++--
fs/erofs/inode.c | 5 +++++
fs/erofs/internal.h | 2 ++
fs/erofs/super.c | 26 ++++++++++++++++++++++++--
4 files changed, 72 insertions(+), 4 deletions(-)
diff --git a/fs/erofs/data.c b/fs/erofs/data.c
index 0f82b4cb474c..c188c629be45 100644
--- a/fs/erofs/data.c
+++ b/fs/erofs/data.c
@@ -6,7 +6,7 @@
#include "internal.h"
#include <linux/prefetch.h>
#include <linux/iomap.h>
-
+#include <linux/dax.h>
#include <trace/events/erofs.h>
static void erofs_readendio(struct bio *bio)
@@ -323,6 +323,7 @@ static int erofs_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
return ret;
iomap->bdev = inode->i_sb->s_bdev;
+ iomap->dax_dev = EROFS_I_SB(inode)->dax_dev;
iomap->offset = map.m_la;
iomap->length = map.m_llen;
@@ -382,6 +383,11 @@ static ssize_t erofs_file_read_iter(struct kiocb *iocb, struct iov_iter *to)
if (!iov_iter_count(to))
return 0;
+#ifdef CONFIG_FS_DAX
+ if (IS_DAX(iocb->ki_filp->f_mapping->host))
+ return dax_iomap_rw(iocb, to, &erofs_iomap_ops);
+#endif
+
if (iocb->ki_flags & IOCB_DIRECT) {
int err = erofs_prepare_dio(iocb, to);
@@ -410,9 +416,42 @@ const struct address_space_operations erofs_raw_access_aops = {
.direct_IO = noop_direct_IO,
};
+#ifdef CONFIG_FS_DAX
+static vm_fault_t erofs_dax_huge_fault(struct vm_fault *vmf,
+ enum page_entry_size pe_size)
+{
+ return dax_iomap_fault(vmf, pe_size, NULL, NULL, &erofs_iomap_ops);
+}
+
+static vm_fault_t erofs_dax_fault(struct vm_fault *vmf)
+{
+ return erofs_dax_huge_fault(vmf, PE_SIZE_PTE);
+}
+
+static const struct vm_operations_struct erofs_dax_vm_ops = {
+ .fault = erofs_dax_fault,
+ .huge_fault = erofs_dax_huge_fault,
+};
+
+static int erofs_file_mmap(struct file *file, struct vm_area_struct *vma)
+{
+ if (!IS_DAX(file_inode(file)))
+ return generic_file_readonly_mmap(file, vma);
+
+ if ((vma->vm_flags & VM_SHARED) && (vma->vm_flags & VM_MAYWRITE))
+ return -EINVAL;
+
+ vma->vm_ops = &erofs_dax_vm_ops;
+ vma->vm_flags |= VM_HUGEPAGE;
+ return 0;
+}
+#else
+#define erofs_file_mmap generic_file_readonly_mmap
+#endif
+
const struct file_operations erofs_file_fops = {
.llseek = generic_file_llseek,
.read_iter = erofs_file_read_iter,
- .mmap = generic_file_readonly_mmap,
+ .mmap = erofs_file_mmap,
.splice_read = generic_file_splice_read,
};
diff --git a/fs/erofs/inode.c b/fs/erofs/inode.c
index 00edb7562fea..695b97acb9a6 100644
--- a/fs/erofs/inode.c
+++ b/fs/erofs/inode.c
@@ -174,6 +174,11 @@ static struct page *erofs_read_inode(struct inode *inode,
inode->i_mtime.tv_nsec = inode->i_ctime.tv_nsec;
inode->i_atime.tv_nsec = inode->i_ctime.tv_nsec;
+ inode->i_flags &= ~S_DAX;
+ if (test_opt(&sbi->ctx, DAX) && S_ISREG(inode->i_mode) &&
+ vi->datalayout == EROFS_INODE_FLAT_PLAIN)
+ inode->i_flags |= S_DAX;
+
if (!nblks)
/* measure inode.i_blocks as generic filesystems */
inode->i_blocks = roundup(inode->i_size, EROFS_BLKSIZ) >> 9;
diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h
index 2669c785d548..8b0542d35148 100644
--- a/fs/erofs/internal.h
+++ b/fs/erofs/internal.h
@@ -83,6 +83,7 @@ struct erofs_sb_info {
struct erofs_sb_lz4_info lz4;
#endif /* CONFIG_EROFS_FS_ZIP */
+ struct dax_device *dax_dev;
u32 blocks;
u32 meta_blkaddr;
#ifdef CONFIG_EROFS_FS_XATTR
@@ -115,6 +116,7 @@ struct erofs_sb_info {
/* Mount flags set via mount options or defaults */
#define EROFS_MOUNT_XATTR_USER 0x00000010
#define EROFS_MOUNT_POSIX_ACL 0x00000020
+#define EROFS_MOUNT_DAX 0x00000040
#define clear_opt(ctx, option) ((ctx)->mount_opt &= ~EROFS_MOUNT_##option)
#define set_opt(ctx, option) ((ctx)->mount_opt |= EROFS_MOUNT_##option)
diff --git a/fs/erofs/super.c b/fs/erofs/super.c
index 8fc6c04b54f4..b44a964ab24f 100644
--- a/fs/erofs/super.c
+++ b/fs/erofs/super.c
@@ -11,6 +11,7 @@
#include <linux/crc32c.h>
#include <linux/fs_context.h>
#include <linux/fs_parser.h>
+#include <linux/dax.h>
#include "xattr.h"
#define CREATE_TRACE_POINTS
@@ -355,6 +356,7 @@ enum {
Opt_user_xattr,
Opt_acl,
Opt_cache_strategy,
+ Opt_dax,
Opt_err
};
@@ -370,6 +372,7 @@ static const struct fs_parameter_spec erofs_fs_parameters[] = {
fsparam_flag_no("acl", Opt_acl),
fsparam_enum("cache_strategy", Opt_cache_strategy,
erofs_param_cache_strategy),
+ fsparam_flag("dax", Opt_dax),
{}
};
@@ -410,6 +413,14 @@ static int erofs_fc_parse_param(struct fs_context *fc,
ctx->cache_strategy = result.uint_32;
#else
errorfc(fc, "compression not supported, cache_strategy ignored");
+#endif
+ break;
+ case Opt_dax:
+#ifdef CONFIG_FS_DAX
+ warnfc(fc, "DAX enabled. Warning: EXPERIMENTAL, use at your own risk");
+ set_opt(ctx, DAX);
+#else
+ errorfc(fc, "dax options not supported");
#endif
break;
default:
@@ -496,10 +507,17 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)
return -ENOMEM;
sb->s_fs_info = sbi;
+ sbi->dax_dev = fs_dax_get_by_bdev(sb->s_bdev);
err = erofs_read_superblock(sb);
if (err)
return err;
+ if (test_opt(ctx, DAX) &&
+ !bdev_dax_supported(sb->s_bdev, EROFS_BLKSIZ)) {
+ errorfc(fc, "DAX unsupported by block device. Turning off DAX.");
+ clear_opt(ctx, DAX);
+ }
+
sb->s_flags |= SB_RDONLY | SB_NOATIME;
sb->s_maxbytes = MAX_LFS_FILESIZE;
sb->s_time_gran = 1;
@@ -609,6 +627,8 @@ static void erofs_kill_sb(struct super_block *sb)
sbi = EROFS_SB(sb);
if (!sbi)
return;
+ if (sbi->dax_dev)
+ fs_put_dax(sbi->dax_dev);
kfree(sbi);
sb->s_fs_info = NULL;
}
@@ -711,8 +731,8 @@ static int erofs_statfs(struct dentry *dentry, struct kstatfs *buf)
static int erofs_show_options(struct seq_file *seq, struct dentry *root)
{
- struct erofs_sb_info *sbi __maybe_unused = EROFS_SB(root->d_sb);
- struct erofs_fs_context *ctx __maybe_unused = &sbi->ctx;
+ struct erofs_sb_info *sbi = EROFS_SB(root->d_sb);
+ struct erofs_fs_context *ctx = &sbi->ctx;
#ifdef CONFIG_EROFS_FS_XATTR
if (test_opt(ctx, XATTR_USER))
@@ -734,6 +754,8 @@ static int erofs_show_options(struct seq_file *seq, struct dentry *root)
else if (ctx->cache_strategy == EROFS_ZIP_CACHE_READAROUND)
seq_puts(seq, ",cache_strategy=readaround");
#endif
+ if (test_opt(ctx, DAX))
+ seq_puts(seq, ",dax");
return 0;
}
--
2.24.4
^ permalink raw reply related [flat|nested] 13+ messages in thread
* Re: [RFC PATCH v1.1 2/2] erofs: dax support for non-tailpacking regular file
2021-07-05 13:21 ` Gao Xiang
@ 2021-07-09 1:47 ` Darrick J. Wong
-1 siblings, 0 replies; 13+ messages in thread
From: Darrick J. Wong @ 2021-07-09 1:47 UTC (permalink / raw)
To: Gao Xiang
Cc: linux-erofs, linux-fsdevel, LKML, nvdimm, Liu Bo, Joseqh Qi, Liu Jiang
On Mon, Jul 05, 2021 at 09:21:53PM +0800, Gao Xiang wrote:
> DAX is quite useful for some VM use cases in order to save guest
> memory extremely with minimal lightweight EROFS.
>
> In order to prepare for such use cases, add preliminary dax support
> for non-tailpacking regular files for now.
>
> Tested with the DRAM-emulated PMEM and the EROFS image generated by
> "mkfs.erofs -Enoinline_data enwik9.fsdax.img enwik9"
>
> Cc: nvdimm@lists.linux.dev
> Cc: linux-fsdevel@vger.kernel.org
> Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
> ---
> change since v1:
> - update missing hunks due to patch spliting...
> bdev_dax_supported(...)
> erofs_file_mmap(...)
>
> fs/erofs/data.c | 43 +++++++++++++++++++++++++++++++++++++++++--
> fs/erofs/inode.c | 5 +++++
> fs/erofs/internal.h | 2 ++
> fs/erofs/super.c | 26 ++++++++++++++++++++++++--
> 4 files changed, 72 insertions(+), 4 deletions(-)
>
> diff --git a/fs/erofs/data.c b/fs/erofs/data.c
> index 0f82b4cb474c..c188c629be45 100644
> --- a/fs/erofs/data.c
> +++ b/fs/erofs/data.c
> @@ -6,7 +6,7 @@
> #include "internal.h"
> #include <linux/prefetch.h>
> #include <linux/iomap.h>
> -
> +#include <linux/dax.h>
> #include <trace/events/erofs.h>
>
> static void erofs_readendio(struct bio *bio)
> @@ -323,6 +323,7 @@ static int erofs_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
> return ret;
>
> iomap->bdev = inode->i_sb->s_bdev;
> + iomap->dax_dev = EROFS_I_SB(inode)->dax_dev;
> iomap->offset = map.m_la;
> iomap->length = map.m_llen;
>
> @@ -382,6 +383,11 @@ static ssize_t erofs_file_read_iter(struct kiocb *iocb, struct iov_iter *to)
> if (!iov_iter_count(to))
> return 0;
>
> +#ifdef CONFIG_FS_DAX
> + if (IS_DAX(iocb->ki_filp->f_mapping->host))
> + return dax_iomap_rw(iocb, to, &erofs_iomap_ops);
> +#endif
> +
> if (iocb->ki_flags & IOCB_DIRECT) {
> int err = erofs_prepare_dio(iocb, to);
>
> @@ -410,9 +416,42 @@ const struct address_space_operations erofs_raw_access_aops = {
> .direct_IO = noop_direct_IO,
> };
>
> +#ifdef CONFIG_FS_DAX
> +static vm_fault_t erofs_dax_huge_fault(struct vm_fault *vmf,
> + enum page_entry_size pe_size)
> +{
> + return dax_iomap_fault(vmf, pe_size, NULL, NULL, &erofs_iomap_ops);
> +}
> +
> +static vm_fault_t erofs_dax_fault(struct vm_fault *vmf)
> +{
> + return erofs_dax_huge_fault(vmf, PE_SIZE_PTE);
> +}
> +
> +static const struct vm_operations_struct erofs_dax_vm_ops = {
> + .fault = erofs_dax_fault,
> + .huge_fault = erofs_dax_huge_fault,
> +};
> +
> +static int erofs_file_mmap(struct file *file, struct vm_area_struct *vma)
> +{
> + if (!IS_DAX(file_inode(file)))
> + return generic_file_readonly_mmap(file, vma);
> +
> + if ((vma->vm_flags & VM_SHARED) && (vma->vm_flags & VM_MAYWRITE))
> + return -EINVAL;
> +
> + vma->vm_ops = &erofs_dax_vm_ops;
> + vma->vm_flags |= VM_HUGEPAGE;
> + return 0;
> +}
> +#else
> +#define erofs_file_mmap generic_file_readonly_mmap
> +#endif
> +
> const struct file_operations erofs_file_fops = {
> .llseek = generic_file_llseek,
> .read_iter = erofs_file_read_iter,
> - .mmap = generic_file_readonly_mmap,
> + .mmap = erofs_file_mmap,
> .splice_read = generic_file_splice_read,
> };
> diff --git a/fs/erofs/inode.c b/fs/erofs/inode.c
> index 00edb7562fea..695b97acb9a6 100644
> --- a/fs/erofs/inode.c
> +++ b/fs/erofs/inode.c
> @@ -174,6 +174,11 @@ static struct page *erofs_read_inode(struct inode *inode,
> inode->i_mtime.tv_nsec = inode->i_ctime.tv_nsec;
> inode->i_atime.tv_nsec = inode->i_ctime.tv_nsec;
>
> + inode->i_flags &= ~S_DAX;
> + if (test_opt(&sbi->ctx, DAX) && S_ISREG(inode->i_mode) &&
> + vi->datalayout == EROFS_INODE_FLAT_PLAIN)
> + inode->i_flags |= S_DAX;
> +
> if (!nblks)
> /* measure inode.i_blocks as generic filesystems */
> inode->i_blocks = roundup(inode->i_size, EROFS_BLKSIZ) >> 9;
> diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h
> index 2669c785d548..8b0542d35148 100644
> --- a/fs/erofs/internal.h
> +++ b/fs/erofs/internal.h
> @@ -83,6 +83,7 @@ struct erofs_sb_info {
>
> struct erofs_sb_lz4_info lz4;
> #endif /* CONFIG_EROFS_FS_ZIP */
> + struct dax_device *dax_dev;
> u32 blocks;
> u32 meta_blkaddr;
> #ifdef CONFIG_EROFS_FS_XATTR
> @@ -115,6 +116,7 @@ struct erofs_sb_info {
> /* Mount flags set via mount options or defaults */
> #define EROFS_MOUNT_XATTR_USER 0x00000010
> #define EROFS_MOUNT_POSIX_ACL 0x00000020
> +#define EROFS_MOUNT_DAX 0x00000040
>
> #define clear_opt(ctx, option) ((ctx)->mount_opt &= ~EROFS_MOUNT_##option)
> #define set_opt(ctx, option) ((ctx)->mount_opt |= EROFS_MOUNT_##option)
> diff --git a/fs/erofs/super.c b/fs/erofs/super.c
> index 8fc6c04b54f4..b44a964ab24f 100644
> --- a/fs/erofs/super.c
> +++ b/fs/erofs/super.c
> @@ -11,6 +11,7 @@
> #include <linux/crc32c.h>
> #include <linux/fs_context.h>
> #include <linux/fs_parser.h>
> +#include <linux/dax.h>
> #include "xattr.h"
>
> #define CREATE_TRACE_POINTS
> @@ -355,6 +356,7 @@ enum {
> Opt_user_xattr,
> Opt_acl,
> Opt_cache_strategy,
> + Opt_dax,
> Opt_err
> };
>
> @@ -370,6 +372,7 @@ static const struct fs_parameter_spec erofs_fs_parameters[] = {
> fsparam_flag_no("acl", Opt_acl),
> fsparam_enum("cache_strategy", Opt_cache_strategy,
> erofs_param_cache_strategy),
> + fsparam_flag("dax", Opt_dax),
> {}
> };
>
> @@ -410,6 +413,14 @@ static int erofs_fc_parse_param(struct fs_context *fc,
> ctx->cache_strategy = result.uint_32;
> #else
> errorfc(fc, "compression not supported, cache_strategy ignored");
> +#endif
> + break;
> + case Opt_dax:
> +#ifdef CONFIG_FS_DAX
> + warnfc(fc, "DAX enabled. Warning: EXPERIMENTAL, use at your own risk");
> + set_opt(ctx, DAX);
You might want to allow 'dax=always' and 'dax=never' to maintain parity
with xfs/ext4's mount options...
--D
> +#else
> + errorfc(fc, "dax options not supported");
> #endif
> break;
> default:
> @@ -496,10 +507,17 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)
> return -ENOMEM;
>
> sb->s_fs_info = sbi;
> + sbi->dax_dev = fs_dax_get_by_bdev(sb->s_bdev);
> err = erofs_read_superblock(sb);
> if (err)
> return err;
>
> + if (test_opt(ctx, DAX) &&
> + !bdev_dax_supported(sb->s_bdev, EROFS_BLKSIZ)) {
> + errorfc(fc, "DAX unsupported by block device. Turning off DAX.");
> + clear_opt(ctx, DAX);
> + }
> +
> sb->s_flags |= SB_RDONLY | SB_NOATIME;
> sb->s_maxbytes = MAX_LFS_FILESIZE;
> sb->s_time_gran = 1;
> @@ -609,6 +627,8 @@ static void erofs_kill_sb(struct super_block *sb)
> sbi = EROFS_SB(sb);
> if (!sbi)
> return;
> + if (sbi->dax_dev)
> + fs_put_dax(sbi->dax_dev);
> kfree(sbi);
> sb->s_fs_info = NULL;
> }
> @@ -711,8 +731,8 @@ static int erofs_statfs(struct dentry *dentry, struct kstatfs *buf)
>
> static int erofs_show_options(struct seq_file *seq, struct dentry *root)
> {
> - struct erofs_sb_info *sbi __maybe_unused = EROFS_SB(root->d_sb);
> - struct erofs_fs_context *ctx __maybe_unused = &sbi->ctx;
> + struct erofs_sb_info *sbi = EROFS_SB(root->d_sb);
> + struct erofs_fs_context *ctx = &sbi->ctx;
>
> #ifdef CONFIG_EROFS_FS_XATTR
> if (test_opt(ctx, XATTR_USER))
> @@ -734,6 +754,8 @@ static int erofs_show_options(struct seq_file *seq, struct dentry *root)
> else if (ctx->cache_strategy == EROFS_ZIP_CACHE_READAROUND)
> seq_puts(seq, ",cache_strategy=readaround");
> #endif
> + if (test_opt(ctx, DAX))
> + seq_puts(seq, ",dax");
> return 0;
> }
>
> --
> 2.24.4
>
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [RFC PATCH v1.1 2/2] erofs: dax support for non-tailpacking regular file
@ 2021-07-09 1:47 ` Darrick J. Wong
0 siblings, 0 replies; 13+ messages in thread
From: Darrick J. Wong @ 2021-07-09 1:47 UTC (permalink / raw)
To: Gao Xiang
Cc: nvdimm, LKML, Joseqh Qi, Liu Bo, linux-fsdevel, Liu Jiang, linux-erofs
On Mon, Jul 05, 2021 at 09:21:53PM +0800, Gao Xiang wrote:
> DAX is quite useful for some VM use cases in order to save guest
> memory extremely with minimal lightweight EROFS.
>
> In order to prepare for such use cases, add preliminary dax support
> for non-tailpacking regular files for now.
>
> Tested with the DRAM-emulated PMEM and the EROFS image generated by
> "mkfs.erofs -Enoinline_data enwik9.fsdax.img enwik9"
>
> Cc: nvdimm@lists.linux.dev
> Cc: linux-fsdevel@vger.kernel.org
> Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
> ---
> change since v1:
> - update missing hunks due to patch spliting...
> bdev_dax_supported(...)
> erofs_file_mmap(...)
>
> fs/erofs/data.c | 43 +++++++++++++++++++++++++++++++++++++++++--
> fs/erofs/inode.c | 5 +++++
> fs/erofs/internal.h | 2 ++
> fs/erofs/super.c | 26 ++++++++++++++++++++++++--
> 4 files changed, 72 insertions(+), 4 deletions(-)
>
> diff --git a/fs/erofs/data.c b/fs/erofs/data.c
> index 0f82b4cb474c..c188c629be45 100644
> --- a/fs/erofs/data.c
> +++ b/fs/erofs/data.c
> @@ -6,7 +6,7 @@
> #include "internal.h"
> #include <linux/prefetch.h>
> #include <linux/iomap.h>
> -
> +#include <linux/dax.h>
> #include <trace/events/erofs.h>
>
> static void erofs_readendio(struct bio *bio)
> @@ -323,6 +323,7 @@ static int erofs_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
> return ret;
>
> iomap->bdev = inode->i_sb->s_bdev;
> + iomap->dax_dev = EROFS_I_SB(inode)->dax_dev;
> iomap->offset = map.m_la;
> iomap->length = map.m_llen;
>
> @@ -382,6 +383,11 @@ static ssize_t erofs_file_read_iter(struct kiocb *iocb, struct iov_iter *to)
> if (!iov_iter_count(to))
> return 0;
>
> +#ifdef CONFIG_FS_DAX
> + if (IS_DAX(iocb->ki_filp->f_mapping->host))
> + return dax_iomap_rw(iocb, to, &erofs_iomap_ops);
> +#endif
> +
> if (iocb->ki_flags & IOCB_DIRECT) {
> int err = erofs_prepare_dio(iocb, to);
>
> @@ -410,9 +416,42 @@ const struct address_space_operations erofs_raw_access_aops = {
> .direct_IO = noop_direct_IO,
> };
>
> +#ifdef CONFIG_FS_DAX
> +static vm_fault_t erofs_dax_huge_fault(struct vm_fault *vmf,
> + enum page_entry_size pe_size)
> +{
> + return dax_iomap_fault(vmf, pe_size, NULL, NULL, &erofs_iomap_ops);
> +}
> +
> +static vm_fault_t erofs_dax_fault(struct vm_fault *vmf)
> +{
> + return erofs_dax_huge_fault(vmf, PE_SIZE_PTE);
> +}
> +
> +static const struct vm_operations_struct erofs_dax_vm_ops = {
> + .fault = erofs_dax_fault,
> + .huge_fault = erofs_dax_huge_fault,
> +};
> +
> +static int erofs_file_mmap(struct file *file, struct vm_area_struct *vma)
> +{
> + if (!IS_DAX(file_inode(file)))
> + return generic_file_readonly_mmap(file, vma);
> +
> + if ((vma->vm_flags & VM_SHARED) && (vma->vm_flags & VM_MAYWRITE))
> + return -EINVAL;
> +
> + vma->vm_ops = &erofs_dax_vm_ops;
> + vma->vm_flags |= VM_HUGEPAGE;
> + return 0;
> +}
> +#else
> +#define erofs_file_mmap generic_file_readonly_mmap
> +#endif
> +
> const struct file_operations erofs_file_fops = {
> .llseek = generic_file_llseek,
> .read_iter = erofs_file_read_iter,
> - .mmap = generic_file_readonly_mmap,
> + .mmap = erofs_file_mmap,
> .splice_read = generic_file_splice_read,
> };
> diff --git a/fs/erofs/inode.c b/fs/erofs/inode.c
> index 00edb7562fea..695b97acb9a6 100644
> --- a/fs/erofs/inode.c
> +++ b/fs/erofs/inode.c
> @@ -174,6 +174,11 @@ static struct page *erofs_read_inode(struct inode *inode,
> inode->i_mtime.tv_nsec = inode->i_ctime.tv_nsec;
> inode->i_atime.tv_nsec = inode->i_ctime.tv_nsec;
>
> + inode->i_flags &= ~S_DAX;
> + if (test_opt(&sbi->ctx, DAX) && S_ISREG(inode->i_mode) &&
> + vi->datalayout == EROFS_INODE_FLAT_PLAIN)
> + inode->i_flags |= S_DAX;
> +
> if (!nblks)
> /* measure inode.i_blocks as generic filesystems */
> inode->i_blocks = roundup(inode->i_size, EROFS_BLKSIZ) >> 9;
> diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h
> index 2669c785d548..8b0542d35148 100644
> --- a/fs/erofs/internal.h
> +++ b/fs/erofs/internal.h
> @@ -83,6 +83,7 @@ struct erofs_sb_info {
>
> struct erofs_sb_lz4_info lz4;
> #endif /* CONFIG_EROFS_FS_ZIP */
> + struct dax_device *dax_dev;
> u32 blocks;
> u32 meta_blkaddr;
> #ifdef CONFIG_EROFS_FS_XATTR
> @@ -115,6 +116,7 @@ struct erofs_sb_info {
> /* Mount flags set via mount options or defaults */
> #define EROFS_MOUNT_XATTR_USER 0x00000010
> #define EROFS_MOUNT_POSIX_ACL 0x00000020
> +#define EROFS_MOUNT_DAX 0x00000040
>
> #define clear_opt(ctx, option) ((ctx)->mount_opt &= ~EROFS_MOUNT_##option)
> #define set_opt(ctx, option) ((ctx)->mount_opt |= EROFS_MOUNT_##option)
> diff --git a/fs/erofs/super.c b/fs/erofs/super.c
> index 8fc6c04b54f4..b44a964ab24f 100644
> --- a/fs/erofs/super.c
> +++ b/fs/erofs/super.c
> @@ -11,6 +11,7 @@
> #include <linux/crc32c.h>
> #include <linux/fs_context.h>
> #include <linux/fs_parser.h>
> +#include <linux/dax.h>
> #include "xattr.h"
>
> #define CREATE_TRACE_POINTS
> @@ -355,6 +356,7 @@ enum {
> Opt_user_xattr,
> Opt_acl,
> Opt_cache_strategy,
> + Opt_dax,
> Opt_err
> };
>
> @@ -370,6 +372,7 @@ static const struct fs_parameter_spec erofs_fs_parameters[] = {
> fsparam_flag_no("acl", Opt_acl),
> fsparam_enum("cache_strategy", Opt_cache_strategy,
> erofs_param_cache_strategy),
> + fsparam_flag("dax", Opt_dax),
> {}
> };
>
> @@ -410,6 +413,14 @@ static int erofs_fc_parse_param(struct fs_context *fc,
> ctx->cache_strategy = result.uint_32;
> #else
> errorfc(fc, "compression not supported, cache_strategy ignored");
> +#endif
> + break;
> + case Opt_dax:
> +#ifdef CONFIG_FS_DAX
> + warnfc(fc, "DAX enabled. Warning: EXPERIMENTAL, use at your own risk");
> + set_opt(ctx, DAX);
You might want to allow 'dax=always' and 'dax=never' to maintain parity
with xfs/ext4's mount options...
--D
> +#else
> + errorfc(fc, "dax options not supported");
> #endif
> break;
> default:
> @@ -496,10 +507,17 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)
> return -ENOMEM;
>
> sb->s_fs_info = sbi;
> + sbi->dax_dev = fs_dax_get_by_bdev(sb->s_bdev);
> err = erofs_read_superblock(sb);
> if (err)
> return err;
>
> + if (test_opt(ctx, DAX) &&
> + !bdev_dax_supported(sb->s_bdev, EROFS_BLKSIZ)) {
> + errorfc(fc, "DAX unsupported by block device. Turning off DAX.");
> + clear_opt(ctx, DAX);
> + }
> +
> sb->s_flags |= SB_RDONLY | SB_NOATIME;
> sb->s_maxbytes = MAX_LFS_FILESIZE;
> sb->s_time_gran = 1;
> @@ -609,6 +627,8 @@ static void erofs_kill_sb(struct super_block *sb)
> sbi = EROFS_SB(sb);
> if (!sbi)
> return;
> + if (sbi->dax_dev)
> + fs_put_dax(sbi->dax_dev);
> kfree(sbi);
> sb->s_fs_info = NULL;
> }
> @@ -711,8 +731,8 @@ static int erofs_statfs(struct dentry *dentry, struct kstatfs *buf)
>
> static int erofs_show_options(struct seq_file *seq, struct dentry *root)
> {
> - struct erofs_sb_info *sbi __maybe_unused = EROFS_SB(root->d_sb);
> - struct erofs_fs_context *ctx __maybe_unused = &sbi->ctx;
> + struct erofs_sb_info *sbi = EROFS_SB(root->d_sb);
> + struct erofs_fs_context *ctx = &sbi->ctx;
>
> #ifdef CONFIG_EROFS_FS_XATTR
> if (test_opt(ctx, XATTR_USER))
> @@ -734,6 +754,8 @@ static int erofs_show_options(struct seq_file *seq, struct dentry *root)
> else if (ctx->cache_strategy == EROFS_ZIP_CACHE_READAROUND)
> seq_puts(seq, ",cache_strategy=readaround");
> #endif
> + if (test_opt(ctx, DAX))
> + seq_puts(seq, ",dax");
> return 0;
> }
>
> --
> 2.24.4
>
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [RFC PATCH v1.1 2/2] erofs: dax support for non-tailpacking regular file
2021-07-09 1:47 ` Darrick J. Wong
@ 2021-07-09 2:28 ` Gao Xiang
-1 siblings, 0 replies; 13+ messages in thread
From: Gao Xiang @ 2021-07-09 2:28 UTC (permalink / raw)
To: Darrick J. Wong
Cc: linux-erofs, linux-fsdevel, LKML, nvdimm, Liu Bo, Joseqh Qi, Liu Jiang
Hi Darrick,
On Thu, Jul 08, 2021 at 06:47:19PM -0700, Darrick J. Wong wrote:
> On Mon, Jul 05, 2021 at 09:21:53PM +0800, Gao Xiang wrote:
...
> > Opt_cache_strategy,
> > + Opt_dax,
> > Opt_err
> > };
> >
> > @@ -370,6 +372,7 @@ static const struct fs_parameter_spec erofs_fs_parameters[] = {
> > fsparam_flag_no("acl", Opt_acl),
> > fsparam_enum("cache_strategy", Opt_cache_strategy,
> > erofs_param_cache_strategy),
> > + fsparam_flag("dax", Opt_dax),
> > {}
> > };
> >
> > @@ -410,6 +413,14 @@ static int erofs_fc_parse_param(struct fs_context *fc,
> > ctx->cache_strategy = result.uint_32;
> > #else
> > errorfc(fc, "compression not supported, cache_strategy ignored");
> > +#endif
> > + break;
> > + case Opt_dax:
> > +#ifdef CONFIG_FS_DAX
> > + warnfc(fc, "DAX enabled. Warning: EXPERIMENTAL, use at your own risk");
> > + set_opt(ctx, DAX);
>
> You might want to allow 'dax=always' and 'dax=never' to maintain parity
> with xfs/ext4's mount options...
Yeah, thanks for your suggestion. Will revise in the next version..
(Also, more use case details and development status about this scenario
will be shown in the following months...)
Thanks,
Gao Xiang
>
> --D
>
> > +#else
> > + errorfc(fc, "dax options not supported");
> > #endif
> > break;
> > default:
> > @@ -496,10 +507,17 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)
> > return -ENOMEM;
> >
> > sb->s_fs_info = sbi;
> > + sbi->dax_dev = fs_dax_get_by_bdev(sb->s_bdev);
> > err = erofs_read_superblock(sb);
> > if (err)
> > return err;
> >
> > + if (test_opt(ctx, DAX) &&
> > + !bdev_dax_supported(sb->s_bdev, EROFS_BLKSIZ)) {
> > + errorfc(fc, "DAX unsupported by block device. Turning off DAX.");
> > + clear_opt(ctx, DAX);
> > + }
> > +
> > sb->s_flags |= SB_RDONLY | SB_NOATIME;
> > sb->s_maxbytes = MAX_LFS_FILESIZE;
> > sb->s_time_gran = 1;
> > @@ -609,6 +627,8 @@ static void erofs_kill_sb(struct super_block *sb)
> > sbi = EROFS_SB(sb);
> > if (!sbi)
> > return;
> > + if (sbi->dax_dev)
> > + fs_put_dax(sbi->dax_dev);
> > kfree(sbi);
> > sb->s_fs_info = NULL;
> > }
> > @@ -711,8 +731,8 @@ static int erofs_statfs(struct dentry *dentry, struct kstatfs *buf)
> >
> > static int erofs_show_options(struct seq_file *seq, struct dentry *root)
> > {
> > - struct erofs_sb_info *sbi __maybe_unused = EROFS_SB(root->d_sb);
> > - struct erofs_fs_context *ctx __maybe_unused = &sbi->ctx;
> > + struct erofs_sb_info *sbi = EROFS_SB(root->d_sb);
> > + struct erofs_fs_context *ctx = &sbi->ctx;
> >
> > #ifdef CONFIG_EROFS_FS_XATTR
> > if (test_opt(ctx, XATTR_USER))
> > @@ -734,6 +754,8 @@ static int erofs_show_options(struct seq_file *seq, struct dentry *root)
> > else if (ctx->cache_strategy == EROFS_ZIP_CACHE_READAROUND)
> > seq_puts(seq, ",cache_strategy=readaround");
> > #endif
> > + if (test_opt(ctx, DAX))
> > + seq_puts(seq, ",dax");
> > return 0;
> > }
> >
> > --
> > 2.24.4
> >
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [RFC PATCH v1.1 2/2] erofs: dax support for non-tailpacking regular file
@ 2021-07-09 2:28 ` Gao Xiang
0 siblings, 0 replies; 13+ messages in thread
From: Gao Xiang @ 2021-07-09 2:28 UTC (permalink / raw)
To: Darrick J. Wong
Cc: nvdimm, LKML, Joseqh Qi, Liu Bo, linux-fsdevel, Liu Jiang, linux-erofs
Hi Darrick,
On Thu, Jul 08, 2021 at 06:47:19PM -0700, Darrick J. Wong wrote:
> On Mon, Jul 05, 2021 at 09:21:53PM +0800, Gao Xiang wrote:
...
> > Opt_cache_strategy,
> > + Opt_dax,
> > Opt_err
> > };
> >
> > @@ -370,6 +372,7 @@ static const struct fs_parameter_spec erofs_fs_parameters[] = {
> > fsparam_flag_no("acl", Opt_acl),
> > fsparam_enum("cache_strategy", Opt_cache_strategy,
> > erofs_param_cache_strategy),
> > + fsparam_flag("dax", Opt_dax),
> > {}
> > };
> >
> > @@ -410,6 +413,14 @@ static int erofs_fc_parse_param(struct fs_context *fc,
> > ctx->cache_strategy = result.uint_32;
> > #else
> > errorfc(fc, "compression not supported, cache_strategy ignored");
> > +#endif
> > + break;
> > + case Opt_dax:
> > +#ifdef CONFIG_FS_DAX
> > + warnfc(fc, "DAX enabled. Warning: EXPERIMENTAL, use at your own risk");
> > + set_opt(ctx, DAX);
>
> You might want to allow 'dax=always' and 'dax=never' to maintain parity
> with xfs/ext4's mount options...
Yeah, thanks for your suggestion. Will revise in the next version..
(Also, more use case details and development status about this scenario
will be shown in the following months...)
Thanks,
Gao Xiang
>
> --D
>
> > +#else
> > + errorfc(fc, "dax options not supported");
> > #endif
> > break;
> > default:
> > @@ -496,10 +507,17 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)
> > return -ENOMEM;
> >
> > sb->s_fs_info = sbi;
> > + sbi->dax_dev = fs_dax_get_by_bdev(sb->s_bdev);
> > err = erofs_read_superblock(sb);
> > if (err)
> > return err;
> >
> > + if (test_opt(ctx, DAX) &&
> > + !bdev_dax_supported(sb->s_bdev, EROFS_BLKSIZ)) {
> > + errorfc(fc, "DAX unsupported by block device. Turning off DAX.");
> > + clear_opt(ctx, DAX);
> > + }
> > +
> > sb->s_flags |= SB_RDONLY | SB_NOATIME;
> > sb->s_maxbytes = MAX_LFS_FILESIZE;
> > sb->s_time_gran = 1;
> > @@ -609,6 +627,8 @@ static void erofs_kill_sb(struct super_block *sb)
> > sbi = EROFS_SB(sb);
> > if (!sbi)
> > return;
> > + if (sbi->dax_dev)
> > + fs_put_dax(sbi->dax_dev);
> > kfree(sbi);
> > sb->s_fs_info = NULL;
> > }
> > @@ -711,8 +731,8 @@ static int erofs_statfs(struct dentry *dentry, struct kstatfs *buf)
> >
> > static int erofs_show_options(struct seq_file *seq, struct dentry *root)
> > {
> > - struct erofs_sb_info *sbi __maybe_unused = EROFS_SB(root->d_sb);
> > - struct erofs_fs_context *ctx __maybe_unused = &sbi->ctx;
> > + struct erofs_sb_info *sbi = EROFS_SB(root->d_sb);
> > + struct erofs_fs_context *ctx = &sbi->ctx;
> >
> > #ifdef CONFIG_EROFS_FS_XATTR
> > if (test_opt(ctx, XATTR_USER))
> > @@ -734,6 +754,8 @@ static int erofs_show_options(struct seq_file *seq, struct dentry *root)
> > else if (ctx->cache_strategy == EROFS_ZIP_CACHE_READAROUND)
> > seq_puts(seq, ",cache_strategy=readaround");
> > #endif
> > + if (test_opt(ctx, DAX))
> > + seq_puts(seq, ",dax");
> > return 0;
> > }
> >
> > --
> > 2.24.4
> >
^ permalink raw reply [flat|nested] 13+ messages in thread