Skip to contentSkip to conversation

9p: add support for root file systems

PATCH In review Changbin Du 4 participants 12 messages Jun 6, 2021 – Jun 20, 2021

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

v3 cover 9p: add support for root file systems

Just like cifs and nfs, this short series enables rootfs support for 9p. Bellow is an example which mounts v9fs with tag 'r' as rootfs in qemu guest via virtio transport.

  $ qemu-system-x86_64 -enable-kvm -cpu host -m 1024 \
        -virtfs local,path=$rootfs_dir,mount_tag=r,security_model=passthrough,id=r \
        -kernel /path/to/linux/arch/x86/boot/bzImage -nographic \
        -append "root=/dev/v9fs v9fsroot=r,trans=virtio rw console=ttyS0 3"

v3:
o rebase.

v2:
o use pr_err instead of printk.
o ROOT_DEV is only set after checking.
o cleanup DEFAULT_MNT_OPTS.
o do not retry mount for fd and virtio transport.

Changbin Du (3):
9p: add support for root file systems
9p: doc: move to a new dedicated folder
9p: doc: add v9fsroot description

Changbin Du (3):
9p: add support for root file systems
9p: doc: move to a new dedicated folder
9p: doc: add v9fsroot description

 Documentation/filesystems/index.rst         |  2 +-
 Documentation/filesystems/{ => v9fs}/9p.rst |  0
 Documentation/filesystems/v9fs/index.rst    | 12 ++++
 Documentation/filesystems/v9fs/v9fsroot.rst | 52 +++++++++++++++++
 MAINTAINERS                                 |  8 ++-
 fs/9p/Kconfig                               |  6 ++
 fs/9p/Makefile                              |  1 +
 fs/9p/v9fsroot.c                            | 64 +++++++++++++++++++++
 include/linux/root_dev.h                    |  1 +
 init/do_mounts.c                            | 55 ++++++++++++++++++
 10 files changed, 199 insertions(+), 2 deletions(-)
 rename Documentation/filesystems/{ => v9fs}/9p.rst (100%)
 create mode 100644 Documentation/filesystems/v9fs/index.rst
 create mode 100644 Documentation/filesystems/v9fs/v9fsroot.rst
 create mode 100644 fs/9p/v9fsroot.c
Signature · 1 line
2.30.2
#1 lore raw

PATCH v3 1/3 9p: add support for root file systems

This introduces a new kernel command-line option called 'v9fsroot=' which will tell the kernel to mount the root file system by utilizing the 9p protocol.

This allows us to mount host folder as rootfs for guest linux in qemu. Bellow is an example which mounts v9fs with tag 'r' as rootfs in qemu guest via virtio transport.

  $ qemu-system-x86_64 -enable-kvm -cpu host -m 1024 \
      -virtfs local,path=$rootfs_dir,mount_tag=r,security_model=passthrough,id=r \
      -kernel /path/to/linux/arch/x86/boot/bzImage -nographic \
      -append "root=/dev/v9fs v9fsroot=r,trans=virtio rw console=ttyS0 3"

---
v2:
o use pr_err instead of printk.
o ROOT_DEV is only set after checking.
o cleanup DEFAULT_MNT_OPTS.
o do not retry mount for fd and virtio transport.

---
 MAINTAINERS              |  5 ++++
 fs/9p/Kconfig            |  6 ++++
 fs/9p/Makefile           |  1 +
 fs/9p/v9fsroot.c         | 64 ++++++++++++++++++++++++++++++++++++++++
 include/linux/root_dev.h |  1 +
 init/do_mounts.c         | 55 ++++++++++++++++++++++++++++++++++
 6 files changed, 132 insertions(+)
 create mode 100644 fs/9p/v9fsroot.c
MAINTAINERS +5 -0
index b706dd20ff2b..35b2c8f614d0 100644
@@ -239,6 +239,11 @@ F:	include/trace/events/9p.h
 F:	include/uapi/linux/virtio_9p.h
 F:	net/9p/
 
+9P FILE SYSTEM ROOTFS SUPPORT
+R:	Changbin Du <[email protected]>
+S:	Supported
+F:	fs/9p/v9fsroot.c
+
 A8293 MEDIA DRIVER
 M:	Antti Palosaari <[email protected]>
 L:	[email protected]
fs/9p/Kconfig +6 -0
index 09fd4a185fd2..71c5a49f9a27 100644
@@ -42,3 +42,9 @@ config 9P_FS_SECURITY
 
 	  If you are not using a security module that requires using
 	  extended attributes for file security labels, say N.
+
+config 9P_FS_ROOT
+	bool "9p root file system"
+	depends on 9P_FS=y
+	help
+	  Enables root file system support over 9p protocol.
fs/9p/Makefile +1 -0
index e7800a5c7395..bc2a4ef10049 100644
@@ -15,3 +15,4 @@ obj-$(CONFIG_9P_FS) := 9p.o
 
 9p-$(CONFIG_9P_FSCACHE) += cache.o
 9p-$(CONFIG_9P_FS_POSIX_ACL) += acl.o
+9p-$(CONFIG_9P_FS_ROOT) += v9fsroot.o
fs/9p/v9fsroot.c new file +64 -0

64 lines not on this page · raw message

include/linux/root_dev.h +1 -0
index 4e78651371ba..becd0ee2ff87 100644
@@ -9,6 +9,7 @@
 enum {
 	Root_NFS = MKDEV(UNNAMED_MAJOR, 255),
 	Root_CIFS = MKDEV(UNNAMED_MAJOR, 254),
+	Root_V9FS = MKDEV(UNNAMED_MAJOR, 253),
 	Root_RAM0 = MKDEV(RAMDISK_MAJOR, 0),
 	Root_RAM1 = MKDEV(RAMDISK_MAJOR, 1),
 	Root_FD0 = MKDEV(FLOPPY_MAJOR, 0),
init/do_mounts.c +55 -0

73 lines not on this page · raw message

Signature · 1 line
2.30.2
#4 lore raw

PATCH v3 2/3 9p: doc: move to a new dedicated folder

In reply to Changbin Du

Later we will add another documentation for v9fs.

---
 Documentation/filesystems/index.rst         |  2 +-
 Documentation/filesystems/{ => v9fs}/9p.rst |  0
 Documentation/filesystems/v9fs/index.rst    | 11 +++++++++++
 MAINTAINERS                                 |  2 +-
 4 files changed, 13 insertions(+), 2 deletions(-)
 rename Documentation/filesystems/{ => v9fs}/9p.rst (100%)
 create mode 100644 Documentation/filesystems/v9fs/index.rst
Documentation/filesystems/index.rst +1 -1
index d4853cb919d2..e53992636a49 100644
@@ -63,7 +63,7 @@ Documentation for filesystem implementations.
 .. toctree::
    :maxdepth: 2
 
-   9p
+   v9fs/index
    adfs
    affs
    afs
Documentation/filesystems/v9fs/9p.rst renamed from Documentation/filesystems/9p.rst +0 -0
similarity index 100%
rename from Documentation/filesystems/9p.rst
rename to Documentation/filesystems/v9fs/9p.rst
Documentation/filesystems/v9fs/index.rst new file +11 -0
new file mode 100644
index 000000000000..a1e45b89e2a2
@@ -0,0 +1,11 @@
+.. SPDX-License-Identifier: GPL-2.0
+
+====
+v9fs
+====
+
+.. toctree::
+   :maxdepth: 6
+   :numbered:
+
+   9p
MAINTAINERS +1 -1
index 35b2c8f614d0..3da44eef1471 100644
@@ -232,7 +232,7 @@ W:	http://swik.net/v9fs
 Q:	http://patchwork.kernel.org/project/v9fs-devel/list/
 T:	git git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs.git
 T:	git git://github.com/martinetd/linux.git
-F:	Documentation/filesystems/9p.rst
+F:	Documentation/filesystems/v9fs/
 F:	fs/9p/
 F:	include/net/9p/
 F:	include/trace/events/9p.h
Signature · 1 line
2.30.2
#5 lore raw

PATCH v3 3/3 9p: doc: add v9fsroot description

In reply to Changbin Du

This documentation is modified from cifs/cifsroot.rst.

---
 Documentation/filesystems/v9fs/index.rst    |  1 +
 Documentation/filesystems/v9fs/v9fsroot.rst | 52 +++++++++++++++++++++
 MAINTAINERS                                 |  1 +
 3 files changed, 54 insertions(+)
 create mode 100644 Documentation/filesystems/v9fs/v9fsroot.rst
Documentation/filesystems/v9fs/index.rst +1 -0
index a1e45b89e2a2..65e1ceb04c9c 100644
@@ -9,3 +9,4 @@ v9fs
    :numbered:
 
    9p
+   v9fsroot
Documentation/filesystems/v9fs/v9fsroot.rst new file +52 -0

52 lines not on this page · raw message

MAINTAINERS +1 -0
index 3da44eef1471..dfcee6dfe182 100644
@@ -243,6 +243,7 @@ F:	net/9p/
 R:	Changbin Du <[email protected]>
 S:	Supported
 F:	fs/9p/v9fsroot.c
+F:	Documentation/filesystems/v9fs/v9fsroot.rst
 
 A8293 MEDIA DRIVER
 M:	Antti Palosaari <[email protected]>
Signature · 1 line
2.30.2
#6 lore raw
In reply to Changbin Du

Changbin Du wrote on Mon, Jun 07, 2021 at 07:09:19AM +0800:

Quoted 8 lines

Just like cifs and nfs, this short series enables rootfs support for 9p. Bellow is an example which mounts v9fs with tag 'r' as rootfs in qemu guest via virtio transport.

  $ qemu-system-x86_64 -enable-kvm -cpu host -m 1024 \
        -virtfs local,path=$rootfs_dir,mount_tag=r,security_model=passthrough,id=r \
        -kernel /path/to/linux/arch/x86/boot/bzImage -nographic \
        -append "root=/dev/v9fs v9fsroot=r,trans=virtio rw console=ttyS0 3"

Thanks for rebasing -- I haven't forgotten, just been otherwise busy.

Will make time for this and the netfs rework over the next couple of weeks.

Just a couple of notes:

linux-fsdevel@vger is still not in Cc of the patches, and they weren't for any other version except for one of my replies, so it might be worth resending just for them as it touches do_mounts.c / root_dev.h

Quoted 1 line
 MAINTAINERS                                 |  8 ++-

I don't think it's worth adding yourself as MAINTAINERS for this, will likely strip it out (I know I'm the one who asked for your help if people have problems with this, but we should keep contacts simple) I'll forward you requests when they come if it's not trivial and you're not subscribed to the v9fs-developer@sf list

Signature · 1 line
Dominique
#2 lore raw
In reply to Changbin Du
I love your patch! Perhaps something to improve:

Hi Changbin,

I love your patch! Perhaps something to improve:

[auto build test WARNING on lwn/docs-next]
[also build test WARNING on linus/master v5.13-rc5 next-20210604] [cannot apply to v9fs/for-next]
[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/Changbin-Du/9p-add-support-for-root-file-systems/20210607-071229
base:   git://git.lwn.net/linux-2.6 docs-next
config: arm-allyesconfig (attached as .config)
compiler: arm-linux-gnueabi-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/96098f751038703cc0fda4f018236d240a86930d
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Changbin-Du/9p-add-support-for-root-file-systems/20210607-071229 git checkout 96098f751038703cc0fda4f018236d240a86930d
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross ARCH=arm

If you fix the issue, kindly add following tag as appropriate

All warnings (new ones prefixed by >>):

>> fs/9p/v9fsroot.c:53:12: warning: no previous prototype for 'v9fs_root_data' [-Wmissing-prototypes]
      53 | int __init v9fs_root_data(char **dev, char **opts)
         |            ^~~~~~~~~~~~~~

vim +/v9fs_root_data +53 fs/9p/v9fsroot.c

    52	
  > 53	int __init v9fs_root_data(char **dev, char **opts)

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org

#3 lore raw
Quoted 25 lines · kernel test robot

Hi Changbin,

I love your patch! Perhaps something to improve:

[auto build test WARNING on lwn/docs-next]
[also build test WARNING on linus/master v5.13-rc5 next-20210604] [cannot apply to v9fs/for-next]
[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/Changbin-Du/9p-add-support-for-root-file-systems/20210607-071229
base:   git://git.lwn.net/linux-2.6 docs-next
config: arm-allyesconfig (attached as .config)
compiler: arm-linux-gnueabi-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/96098f751038703cc0fda4f018236d240a86930d
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Changbin-Du/9p-add-support-for-root-file-systems/20210607-071229 git checkout 96098f751038703cc0fda4f018236d240a86930d
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross ARCH=arm

Quoted 9 lines

If you fix the issue, kindly add following tag as appropriate

All warnings (new ones prefixed by >>):

Quoted 1 line
Quoted 1 line

fs/9p/v9fsroot.c:53:12: warning: no previous prototype for 'v9fs_root_data' [-Wmissing-prototypes]

      53 | int __init v9fs_root_data(char **dev, char **opts)
         |            ^~~~~~~~~~~~~~

This just follows the existing rootfs support manner. This function doesn't have a dedicated header file to place. So I think we can ignore this warning.

Quoted 9 lines

vim +/v9fs_root_data +53 fs/9p/v9fsroot.c

    52	
Quoted 1 line

53 int __init v9fs_root_data(char **dev, char **opts)

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org

Signature · 2 lines
Cheers,
Changbin Du
#7 lore raw
In reply to Changbin Du
Quoted 8 lines · Changbin Du

Just like cifs and nfs, this short series enables rootfs support for 9p. Bellow is an example which mounts v9fs with tag 'r' as rootfs in qemu guest via virtio transport.

  $ qemu-system-x86_64 -enable-kvm -cpu host -m 1024 \
        -virtfs local,path=$rootfs_dir,mount_tag=r,security_model=passthrough,id=r \
        -kernel /path/to/linux/arch/x86/boot/bzImage -nographic \
        -append "root=/dev/v9fs v9fsroot=r,trans=virtio rw console=ttyS0 3"

Rather than inventing a pseudo-device /dev/v9fs for this, would it potentially work to use the existing rootfstype and rootflags options for this? rootfstype already determines what filesystem should be used to mount the root, and rootflags already provides options for that filesystem.

For instance, for the above example:
rootfstype=9p root=r rootflags=trans=virtio

That would require a bit of fiddling to make rootfstype=9p allow a root that's just the mount_tag. If that isn't an option, then even with root=/dev/v9fs I think it still makes sense to use the existing rootflags for "trans=virtio" rather than creating a new "v9fsroot" option for that.

- Josh Triplett

#8 lore raw

Josh Triplett wrote on Mon, Jun 14, 2021 at 01:57:54AM -0700:

Quoted 16 lines · Changbin Du
Quoted 8 lines

Just like cifs and nfs, this short series enables rootfs support for 9p. Bellow is an example which mounts v9fs with tag 'r' as rootfs in qemu guest via virtio transport.

  $ qemu-system-x86_64 -enable-kvm -cpu host -m 1024 \
        -virtfs local,path=$rootfs_dir,mount_tag=r,security_model=passthrough,id=r \
        -kernel /path/to/linux/arch/x86/boot/bzImage -nographic \
        -append "root=/dev/v9fs v9fsroot=r,trans=virtio rw console=ttyS0 3"

Rather than inventing a pseudo-device /dev/v9fs for this, would it potentially work to use the existing rootfstype and rootflags options for this? rootfstype already determines what filesystem should be used to mount the root, and rootflags already provides options for that filesystem.

Quoted 8 lines

For instance, for the above example:
rootfstype=9p root=r rootflags=trans=virtio

That would require a bit of fiddling to make rootfstype=9p allow a root that's just the mount_tag. If that isn't an option, then even with root=/dev/v9fs I think it still makes sense to use the existing rootflags for "trans=virtio" rather than creating a new "v9fsroot" option for that.

This doesn't work as is because of the way the code is written, if there's no block device associated with a root=x option right now it will lead to kernel panic.

I replied with folks in Cc but there's another thread on linux-fsdevel@ with a more generic approach that will build a list of filesystems which don't require such a block device (either hardcoded with virtiofs and 9p or based on FS_REQUIRES_DEV), thread started there but there's a second patch hidden an more discussion below:
https://lore.kernel.org/linux-fsdevel/[email protected]/

My preferred approach right now would be to go with their approach, and adjust the documentation Changbin Du wrote here, to have the best of both worlds.

Signature · 1 line
Dominique
#9 lore raw
Quoted 30 lines · Dominique Martinet

Josh Triplett wrote on Mon, Jun 14, 2021 at 01:57:54AM -0700:

Quoted 16 lines · Changbin Du
Quoted 8 lines

Just like cifs and nfs, this short series enables rootfs support for 9p. Bellow is an example which mounts v9fs with tag 'r' as rootfs in qemu guest via virtio transport.

  $ qemu-system-x86_64 -enable-kvm -cpu host -m 1024 \
        -virtfs local,path=$rootfs_dir,mount_tag=r,security_model=passthrough,id=r \
        -kernel /path/to/linux/arch/x86/boot/bzImage -nographic \
        -append "root=/dev/v9fs v9fsroot=r,trans=virtio rw console=ttyS0 3"

Rather than inventing a pseudo-device /dev/v9fs for this, would it potentially work to use the existing rootfstype and rootflags options for this? rootfstype already determines what filesystem should be used to mount the root, and rootflags already provides options for that filesystem.

Quoted 8 lines

For instance, for the above example:
rootfstype=9p root=r rootflags=trans=virtio

That would require a bit of fiddling to make rootfstype=9p allow a root that's just the mount_tag. If that isn't an option, then even with root=/dev/v9fs I think it still makes sense to use the existing rootflags for "trans=virtio" rather than creating a new "v9fsroot" option for that.

This doesn't work as is because of the way the code is written, if there's no block device associated with a root=x option right now it will lead to kernel panic.

Quoted 6 lines

I replied with folks in Cc but there's another thread on linux-fsdevel@ with a more generic approach that will build a list of filesystems which don't require such a block device (either hardcoded with virtiofs and 9p or based on FS_REQUIRES_DEV), thread started there but there's a second patch hidden an more discussion below:
https://lore.kernel.org/linux-fsdevel/[email protected]/

The patch later on in that thread (either using a list of non-block-device filesystems or the version referenced elsewhere that uses a flag in the filesystem definition) looks really appealing! That's exactly what I was hoping for. That gets us closer to directly translating mount -t type -o options rootdesc into rootfstype=type rootflags=options root=rootdesc in the general case, rather than having special cases for different filesystems.

#10 lore raw
Quoted 38 lines · Josh Triplett
Quoted 30 lines · Dominique Martinet

Josh Triplett wrote on Mon, Jun 14, 2021 at 01:57:54AM -0700:

Quoted 16 lines · Changbin Du
Quoted 8 lines

Just like cifs and nfs, this short series enables rootfs support for 9p. Bellow is an example which mounts v9fs with tag 'r' as rootfs in qemu guest via virtio transport.

  $ qemu-system-x86_64 -enable-kvm -cpu host -m 1024 \
        -virtfs local,path=$rootfs_dir,mount_tag=r,security_model=passthrough,id=r \
        -kernel /path/to/linux/arch/x86/boot/bzImage -nographic \
        -append "root=/dev/v9fs v9fsroot=r,trans=virtio rw console=ttyS0 3"

Rather than inventing a pseudo-device /dev/v9fs for this, would it potentially work to use the existing rootfstype and rootflags options for this? rootfstype already determines what filesystem should be used to mount the root, and rootflags already provides options for that filesystem.

Quoted 8 lines

For instance, for the above example:
rootfstype=9p root=r rootflags=trans=virtio

That would require a bit of fiddling to make rootfstype=9p allow a root that's just the mount_tag. If that isn't an option, then even with root=/dev/v9fs I think it still makes sense to use the existing rootflags for "trans=virtio" rather than creating a new "v9fsroot" option for that.

This doesn't work as is because of the way the code is written, if there's no block device associated with a root=x option right now it will lead to kernel panic.

Quoted 6 lines

I replied with folks in Cc but there's another thread on linux-fsdevel@ with a more generic approach that will build a list of filesystems which don't require such a block device (either hardcoded with virtiofs and 9p or based on FS_REQUIRES_DEV), thread started there but there's a second patch hidden an more discussion below:
https://lore.kernel.org/linux-fsdevel/[email protected]/

Quoted 7 lines

The patch later on in that thread (either using a list of non-block-device filesystems or the version referenced elsewhere that uses a flag in the filesystem definition) looks really appealing! That's exactly what I was hoping for. That gets us closer to directly translating mount -t type -o options rootdesc into rootfstype=type rootflags=options root=rootdesc in the general case, rather than having special cases for different filesystems.

Bellow is all the parameters of non-block rootfs support.

 nfs:    nfsroot=[<server-ip>:]<root-dir>[,<nfs-options>]
 cifs:   cifsroot=//<server-ip>/<share>[,options]
 v9fs:   v9fsroot=<tag/ip>[,options]
 mtd:    root=mtd:<identifier> or root=ubi:<identifier>

virtiofs: root=fstag:<tag>

The main problem is we lack a generic handing for non-block rootdev. I think maybe we can unify all of above.
non-block: root=<type>:<identifier>
blockdev: root=<bock-dev-path> or root=<blockdev major/minor>

Then:

 nfs:    root=nfs:[<server-ip>:]<root-dir>  rootflags=[nfs-options]
 cifs:   root=cifs://<server-ip>/<share>    rootflags=[options]
 v9fs:   root=9p:<tag/ip>                   rootflags=[options]
 mtd:    root=mtd:<identifier>              rootflags=[options]
 ubi:    root=ubi:<identifier>              rootflags=[options]
 virtiofs: root=virtiofs:<tag>              rootflags=[options]

And maybe we can also remove all the special fs code out of init/do_mounts.c.

Signature · 2 lines
Cheers,
Changbin Du
#11 lore raw

Changbin Du wrote on Sun, Jun 20, 2021 at 11:36:59AM +0800:

Quoted 2 lines

The main problem is we lack a generic handing for non-block rootdev. I think maybe we can unify all of above.

We're already going in that direction, please have a look at the threads on fsdevel:
(new patch by Christoph)
https://lore.kernel.org/linux-fsdevel/[email protected]/

(older threads I linked earlier)
https://lore.kernel.org/linux-fsdevel/[email protected]/

I think it's getting there, Christoph should send a v2 addressing Vivek remarks that will likely get picked up.

Signature · 1 line
Dominique