[RFC 07/22] dm: blk: add UCLASS_PARTITION
AKASHI Takahiro
takahiro.akashi at linaro.org
Mon Nov 15 02:43:19 CET 2021
Hi Simon,
On Sat, Nov 13, 2021 at 02:32:20PM -0700, Simon Glass wrote:
> Hi Heinrich,
>
> On Sat, 13 Nov 2021 at 11:42, Heinrich Schuchardt <xypron.glpk at gmx.de> wrote:
> >
> > Am 13. November 2021 19:14:32 MEZ schrieb Simon Glass <sjg at chromium.org>:
> > >Hi,
> > >
> > >On Mon, 8 Nov 2021 at 17:09, Simon Glass <sjg at chromium.org> wrote:
> > >>
> > >> Hi,
> > >>
> > >> On Mon, 8 Nov 2021 at 11:45, Ilias Apalodimas
> > >> <ilias.apalodimas at linaro.org> wrote:
> > >> >
> > >> > Hi chiming in a little late but
> > >> >
> > >> > On Mon, 8 Nov 2021 at 06:46, AKASHI Takahiro <takahiro.akashi at linaro.org> wrote:
> > >> > >
> > >> > > On Fri, Nov 05, 2021 at 10:12:16AM -0600, Simon Glass wrote:
> > >> > > > Hi Takahiro,
> > >> > > >
> > >> > > > On Thu, 4 Nov 2021 at 20:49, AKASHI Takahiro <takahiro.akashi at linaro.org> wrote:
> > >> > > > >
> > >> > > > > On Thu, Nov 04, 2021 at 08:02:05PM -0600, Simon Glass wrote:
> > >> > > > > > Hi,
> > >> > > > > >
> > >> > > > > > On Tue, 2 Nov 2021 at 01:43, Heinrich Schuchardt <xypron.glpk at gmx.de> wrote:
> > >> > > > > > >
> > >> > > > > > >
> > >> > > > > > >
> > >> > > > > > > On 11/1/21 03:14, Simon Glass wrote:
> > >> > > > > > > > Hi Takahiro,
> > >> > > > > > > >
> > >> > > > > > > > On Sun, 31 Oct 2021 at 19:52, AKASHI Takahiro
> > >> > > > > > > > <takahiro.akashi at linaro.org> wrote:
> > >> > > > > > > >>
> > >> > > > > > > >> On Sun, Oct 31, 2021 at 07:15:17PM -0600, Simon Glass wrote:
> > >> > > > > > > >>> Hi Takahiro,
> > >> > > > > > > >>>
> > >> > > > > > > >>> On Sun, 31 Oct 2021 at 18:36, AKASHI Takahiro
> > >> > > > > > > >>> <takahiro.akashi at linaro.org> wrote:
> > >> > > > > > > >>>>
> > >> > > > > > > >>>> On Sat, Oct 30, 2021 at 07:45:14AM +0200, Heinrich Schuchardt wrote:
> > >> > > > > > > >>>>>
> > >> > > > > > > >>>>>
> > >> > > > > > > >>>>> Am 29. Oktober 2021 23:17:56 MESZ schrieb Simon Glass <sjg at chromium.org>:
> > >> > > > > > > >>>>>> Hi,
> > >> > > > > > > >>>>>>
> > >> > > > > > > >>>>>> On Fri, 29 Oct 2021 at 13:26, Heinrich Schuchardt <xypron.glpk at gmx.de> wrote:
> > >> > > > > > > >>>>>>>
> > >> > > > > > > >>>>>>>
> > >> > > > > > > >>>>>>>
> > >> > > > > > > >>>>>>> Am 29. Oktober 2021 08:15:56 MESZ schrieb AKASHI Takahiro <takahiro.akashi at linaro.org>:
> > >> > > > > > > >>>>>>>> On Fri, Oct 29, 2021 at 06:57:24AM +0200, Heinrich Schuchardt wrote:
> > >> > > > > > > >>>>>>>>>
> > >> > > > > > > >>>>>>>>>
> > >> > > > > > > >>>>>>>>>> I agree with Heinrich that we are better to leave BLK as it is, both
> > >> > > > > > > >>>>>>>>>> in name and meaning. I think maybe I am missing the gist of your
> > >> > > > > > > >>>>>>>>>> argument.
> > >> > > > > > > >>>>>>>>>>
> > >> > > > > > > >>>>>>>>>> If we use UCLASS_PART, for example, can we have that refer to both s/w
> > >> > > > > > > >>>>>>>>>> and h/w partitions, as Herinch seems to allude to below? What would
> > >> > > > > > > >>>>>>>>>> the picture look like the, and would it get us closer to agreement?
> > >> > > > > > > >>>>>>>>>
> > >> > > > > > > >>>>>>>>> In the driver model:
> > >> > > > > > > >>>>>>>>>
> > >> > > > > > > >>>>>>>>> A UCLASS is a class of drivers that share the same interface.
> > >> > > > > > > >>>>>>>>> A UDEVICE is a logical device that belongs to exactly one UCLASS and is
> > >> > > > > > > >>>>>>>>> accessed through this UCLASS's interface.
> > >> > > > > > > >>>>>>>>
> > >> > > > > > > >>>>>>>> Please be careful about "accessed through" which is a quite confusing
> > >> > > > > > > >>>>>>>> expression. I don't always agree with this view.
> > >> > > > > > > >>>>>>>>
> > >> > > > > > > >>>>>>>>> A hardware partition is an object that exposes only a single interface
> > >> > > > > > > >>>>>>>>> for block IO.
> > >> > > > > > > >>>>>>>>>
> > >> > > > > > > >>>>>>>>> A software partition is an object that may expose two interfaces: one
> > >> > > > > > > >>>>>>>>> for block IO, the other for file IO.
> > >> > > > > > > >>>>>>>>
> > >> > > > > > > >>>>>>>> Are you talking about UEFI world or U-Boot?
> > >> > > > > > > >>>>>>>> Definitely, a hw partitions can provide a file system
> > >> > > > > > > >>>>>>>> if you want.
> > >> > > > > > > >>>>>>>> It's a matter of usage.
> > >> > > > > > > >>>>>>>>
> > >> > > > > > > >>>>>>>> I remember that we had some discussion about whether block devices
> > >> > > > > > > >>>>>>>> on UEFI system should always have a (sw) partition table or not.
> > >> > > > > > > >>>>>>>> But it is a different topic.
> > >> > > > > > > >>>>>>>>
> > >> > > > > > > >>>>>>>>> The UEFI model does not have a problem with this because on a handle you
> > >> > > > > > > >>>>>>>>> can install as many different protocols as you wish. But U-Boot's driver
> > >> > > > > > > >>>>>>>>> model only allows a single interface per device. Up to now U-Boot has
> > >> > > > > > > >>>>>>>>> overcome this limitation by creating child devices for the extra interfaces.
> > >> > > > > > > >>>>>>>>
> > >> > > > > > > >>>>>>>>> We have the following logical levels:
> > >> > > > > > > >>>>>>>>>
> > >> > > > > > > >>>>>>>>> Controller | Block device | Software Partition| File system
> > >> > > > > > > >>>>>>>>> ----------------+--------------+-------------------+------------
> > >> > > > > > > >>>>>>>>> NVMe Drive | Namespace | Partition 1..n | FAT, EXT4
> > >> > > > > > > >>>>>>>>> ATA Controller | ATA-Drive | |
> > >> > > > > > > >>>>>>>>> SCSI Controller | LUN | |
> > >> > > > > > > >>>>>>>>> MMC Controller | HW-Partition | |
> > >> > > > > > > >>>>>>>>> MMC Controller | SD-Card | |
> > >> > > > > > > >>>>>>>>> USB-Node | USB-Drive | |
> > >> > > > > > > >>>>>>>>>
> > >> > > > > > > >>>>>>>>> In the device tree this could be modeled as:
> > >> > > > > > > >>>>>>>>>
> > >> > > > > > > >>>>>>>>> |-- Controller (UCLASS_CTRL)
> > >> > > > > > > >>>>>>>>> | |-- Block device / HW Partition (UCLASS_BLK) (A)
> > >> > > > > > > >>>>>>>>> | | |-- Partition table (UCLASS_PARTITION_TABLE) (B)
> > >> > > > > > > >>>>>>>>> | | |-- Software Partition (UCLASS_BLK)
> > >> > > > > > > >>>>>>>>> | | |-- File system (UCLASS_FS)
> > >> > > > > > > >>>>>>>>> | |
> > >> > > > > > > >>>>>>>>> | |-- Block device (UCLASS_BLK)
> > >> > > > > > > >>>>>>>>> | |-- File system (UCLASS_FS)
> > >> > > > > > > >>>>>>>>
> > >> > > > > > > >>>>>>>> I don't know why we expect PARTITION_TABLE and FS to appear in DM tree.
> > >> > > > > > > >>>>>>>> What is the benefit?
> > >> > > > > > > >>>>>>>> (A) and (B) always have 1:1 relationship.
> > >> > > > > > > >>>>>>>
> > >> > > > > > > >>>>>>> No. You can have a bare device without a partition table.
> > >> > > > > > > >>>>>>
> > >> > > > > > > >>>>>> I can have a DOS partition that covers the whole device, without a
> > >> > > > > > > >>>>>> partition table. This is supported in U-Boot and Linux.
> > >> > > > > > > >>>>>>
> > >> > > > > > > >>>>>>>
> > >> > > > > > > >>>>>>> We have several partition table drivers: DOS, GPT, OSX, ... . In future we should also have one for the NOR Flash partitions. All of these drivers have a common interface. As the partition table type is mostly independent of the block device type we should use separate uclasses and udevices.
> > >> > > > > > > >>>>>>>
> > >> > > > > > > >>>>>>>> I also remember that you claimed that not all efi objects(handles and
> > >> > > > > > > >>>>>>>> protocols like SIMPE_FILE_SYSTEM_PROTOCOL) need to have corresponding
> > >> > > > > > > >>>>>>>> U-Boot counterparts in our 2019 discussion.
> > >> > > > > > > >>>>>>>>
> > >> > > > > > > >>>>>>>> If we *need* PARTITION_TALBLE, why don't we have HW_PARTITION_TABLE,
> > >> > > > > > > >>>>>>>> which should support other type of hw partitions as well?
> > >> > > > > > > >>>>>>>
> > >> > > > > > > >>>>>>> How hardware partitions, LUNs, ATA drives are enumerated is specific to the type of controller while the type of software partition table is independent of the block device.
> > >> > > > > > > >>>>>>>
> > >> > > > > > > >>>>>>>>
> > >> > > > > > > >>>>>>>> |-- eMMC controller (UCLASS_MMC)
> > >> > > > > > > >>>>>>>> | |-- eMMC device1 / Physical media (UCLASS_HW_PARTITION_TABLE?)
> > >> > > > > > > >>>>>>>> | |-- Block device / HW Partition:user data (UCLASS_BLK)
> > >> > > > > > > >>>>>>>> | | |-- Partition table (UCLASS_PARTITION_TABLE)
> > >> > > > > > > >>>>>>>> | | |-- Software Partition (UCLASS_BLK)
> > >> > > > > > > >>>>>>>> | | |-- File system (UCLASS_FS)
> > >> > > > > > > >>>>>>>> | |
> > >> > > > > > > >>>>>>>> | |-- Block device / HW Partition:boot0 (UCLASS_BLK)
> > >> > > > > > > >>>>>>>> | |-- Block device / HW Partition:boot1 (UCLASS_BLK)
> > >> > > > > > > >>>>>>>> ...
> > >> > > > > > > >>>>>>>> | |-- eMMC device2 / Physical media (UCLASS_HW_PARTITION_TABLE?)
> > >> > > > > > > >>>>>>>>
> > >> > > > > > > >>>>>>>> |-- scsi controller (UCLASS_SCSI)
> > >> > > > > > > >>>>>>>> | |-- scsi disk / Physical media (UCLASS_HW_PARTITION_TABLE?)
> > >> > > > > > > >>>>>>>> | |-- scsi LUN1 (UCLASS_HW_PARTITION_TABLE?)
> > >> > > > > > > >>>>>>>> | | |-- Partition table (UCLASS_PARTITION_TABLE)
> > >> > > > > > > >>>>>>>> | | |-- Software Partition (UCLASS_BLK)
> > >> > > > > > > >>>>>>>> | |-- scsi LUN2 (UCLASS_HW_PARTITION_TABLE?)
> > >> > > > > > > >>>>>>>> ...
> > >> > > > > > > >>>>>>>>
> > >> > > > > > > >>>>>>>> (Here I ignored scsi buses/channels which make things more complicated.)
> > >> > > > > > > >>>>>>>>
> > >> > > > > > > >>>>>>>> This kind of complex hierarchy doesn't benefit anybody.
> > >> > > > > > > >>>>>>>
> > >> > > > > > > >>>>>>> All these levels exist already. We simply do not model them yet in the DM way.
> > >> > > > > > > >>>>>>>
> > >> > > > > > > >>>>>>> The device tree depth is the outcome of the udevice exposing always only a single interface defined by the uclass.
> > >> > > > > > > >>>>>>>
> > >> > > > > > > >>>>>>> The UEFI design allows installing multiple protocol interfaces on a single handle. This may result in simpler device trees in some cases.
> > >> > > > > > > >>>>>>
> > >> > > > > > > >>>>>> Yes, the complexity has to go somewhere. With driver model I chose to
> > >> > > > > > > >>>>>> have a single interface per uclass, since it is simpler to understand,
> > >> > > > > > > >>>>>> no need to request a protocol for a device, etc.
> > >> > > > > > > >>>>>>
> > >> > > > > > > >>>>>> Our current setup is similar to this
> > >> > > > > > > >>>>>>
> > >> > > > > > > >>>>>> |-- Controller (UCLASS_MMC)
> > >> > > > > > > >>>>>> | |-- Block device (UCLASS_BLK) - 'usual' HW partition
> > >> > > > > > > >>>>>> | |-- Block device (UCLASS_BLK) - e.g. for a different HW partition*
> > >> > > > > > > >>>>>>
> > >> > > > > > > >>>>>> * although I don't think the MMC code actually supports it - SCSI does though
> > >> > > > > > > >>>>>>
> > >> > > > > > > >>>>>> We want to add devices for the partition table and the filesystem, so could do:
> > >> > > > > > > >>>>>>
> > >> > > > > > > >>>>>> |-- Controller (UCLASS_MMC)
> > >> > > > > > > >>>>>> | |-- Block device (UCLASS_BLK) - 'usual' HW partition (the whole device)
> > >> > > > > > > >>>>>> | | |-- Partition table (UCLASS_PART) - DOS partition (or EFI)
> > >> > > > > > > >>>>>> | | | |-- Block device (UCLASS_BLK) - partition 1
> > >> > > > > > > >>>>>> | | | | |-- Filesystem (UCLASS_FS) - DOS filesystem
> > >> > > > > > > >>>>>> | | | |-- Block device (UCLASS_BLK) - partition 2
> > >> > > > > > > >>>>>> | | | | |-- Filesystem (UCLASS_FS) - ext5 filesystem
> > >> > > > > > > >>>>>> | |-- Block device (UCLASS_BLK) - e.g. for a different HW
> > >> > > > > > > >>>>>> partition (the whole device)
> > >> > > > > > > >>>>>>
> > >> > > > > > > >>>>>> This is similar to Heinrich's, but without the top-level
> > >> > > > > > > >>>>>> UCLASS_HW_PARTITION_TABLE which I am not sure is necessary.
> > >> > > > > > > >>>>>
> > >> > > > > > > >>>>> Are further MMC hw partitions, multiple SCSI LUNs and multiple NVME namespaces already treated as separate BLK devices?
> > >> > > > > > > >>>>
> > >> > > > > > > >>>> Yes.
> > >> > > > > > > >>>> What I meant to say is that, if we don't need a partition table 'udevice'
> > >> > > > > > > >>>> for hw partitions, we don't need such a device for sw partitions neither.
> > >> > > > > > > >>>>
> > >> > > > > > > >>>> Meanwhile, what about UCLASS_FS? Why do we need this?
> > >> > > > > > > >>>
> > >> > > > > > > >>> We don't need it for our current discussion, but if we want to 'open'
> > >> > > > > > > >>> the filesystem and keep the metadata around, rather than reading it
> > >> > > > > > > >>> again every time we access a file, we might find it useful. Open files
> > >> > > > > > > >>> could be children of the FS uclass, perhaps, if we go a step further
> > >> > > > > > > >>> and create devices for them.
> > >> > > > > > > >>
> > >> > > > > > > >> Do you want to invent linux-like mount-point concepts or procfs?
> > >> > > > > > > >> I remember that you didn't want to have child nodes under BLK devices.
> > >> > > > > > > >> I'm getting confused about our goal.
> > >> > > > > > > >
> > >> > > > > > > > I think we are all a bit unsure.
> > >> > > > > > > >
> > >> > > > > > > > I think BLK devices can have children, sorry if I said the wrong thing
> > >> > > > > > > > somewhere along the way. For example, a partition would be under a BLK
> > >> > > > > > > > device, or a FS.
> > >> > > > > > > >
> > >> > > > > > > >> What should DM represent in U-Boot world?
> > >> > > > > > > >
> > >> > > > > > > > That is what we are trying to figure out.
> > >> > > > > > > >
> > >> > > > > > > > I think the minimum is to have a a way to represent partitions (s/w
> > >> > > > > > > > and hw/). As I understand it, that's what we've been discussing.
> > >> > > > > > >
> > >> > > > > > > The discovery of hardware partitions is specific to the block device
> > >> > > > > > > controller SCSI/MMC/ATA/NVMe. We currently do not provide any
> > >> > > > > > > manipulation commands to create hardware partitions (e.g. NVMe
> > >> > > > > > > namespaces, SCSI LUNs). This is why extracting a uclass for hardware
> > >> > > > > > > partitions does not seem necessary.
> > >> > > > > >
> > >> > > > > > I can see the reasoning here. It might not stand the test of time but
> > >> > > > > > how about we go with it for now? For MMC hardware partition we would
> > >> > > > > > just end up with multiple BLK devices, like we do with SCSI LUNs at
> > >> > > > > > present, which seems like it should work (with some code tweaks).
> > >> > > > > >
> > >> > > > > > >
> > >> > > > > > > Software partitioning (MBR, GPT, ...) is independent of the harboring
> > >> > > > > > > block device.
> > >> > > > > > >
> > >> > > > > > > We already have a set of drivers for software partition tables in disk/.
> > >> > > > > > > Currently the available methods of the drivers are defined in
> > >> > > > > > > U_BOOT_PART_TYPE referring to struct part_driver.
> > >> > > > > > >
> > >> > > > > > > Currently struct part_driver knows only the following methods:
> > >> > > > > > >
> > >> > > > > > > - get_info()
> > >> > > > > > > - print()
> > >> > > > > > > - test()
> > >> > > > > > >
> > >> > > > > > > These drivers should be ome a uclass.
> > >> > > > > > >
> > >> > > > > > > gpt.c and mbr.c allow to create and delete partitions. I think we should add
> > >> > > > > > >
> > >> > > > > > > - create_partition()
> > >> > > > > > > - delete_partition()
> > >> > > > > > >
> > >> > > > > > > to the uclass methods.
> > >> > > > > >
> > >> > > > > > That sounds good to me, although since it is a partition uclass, we
> > >> > > > > > can just use create() and delete().
> > >> > > > >
> > >> > > > > I don't know why we need a "partition table" device in the middle
> > >> > > > > of DM hierarchy.
> > >> > > > > I believe that it simply makes the view of DM tree complicated
> > >> > > > > without any explicit benefit.
> > >> > > >
> > >> > > > Well we clearly have an API here. The partition uclass can:
> > >> > > >
> > >> > > > - hold the partition table in dev_get_uclass_priv()
> > >> > > > - support a read() operation to read the partition
> > >> > > > - support create() to rewrite the partition table
> > >> > > > - support delete() to overwrite/erase the partition table
> > >> > > >
> > >> > > > Then it means that filesystems have the partition table as a parent
> > >> > > > (unless they are whole-device filesystems), which makes sense
> > >> > > >
> > >> > > > So that's why I like the idea.
> > >> > > >
> > >> > > > Other than the extra complexity, is there anything else wrong with it?
> > >> > >
> > >> > > - First of all, a partition table doesn't look like a 'device' at all.
> > >> > > - Second, a partition table is just static data for block devices.
> > >> > > IMO, even if we want to have this data, we can simply hold it
> > >> > > as some sort of attribute of the device, or maybe as a 'tag' which
> > >> > > I will introduce in the next version.
> > >> > >
> > >> > > -Takahiro Akashi
> > >> > >
> > >> >
> > >> > I don't know how this affect the code, but I agree with Akashi-san
> > >> > here. It's indeed useful to keep the partition table stored
> > >> > somewhere, but I think not showing them as part of the device tree is
> > >> > more intuitive.
> > >>
> > >> Well I think I'm easy either way. I just thought that Heinrich made a
> > >> good case for having a partition uclass.
> > >>
> > >> But as Takahiro says, we can use a tag to attach the partition table
> > >> to the device. But it should be attached to the device's children (the
> > >> BLK device) not the media device itself, right?
> > >
> > >As there has been no discussion in 5 days and Takahiro is writing
> > >this, let's go with no uclass for the partition table.
> > >
> >
> > Without uclass you cannot bring the partition table drivers into th driver model.
This transition may be able to be done later when really necessary
as long as we agree that a partition table be hold within a "raw" disk
object (with a tag support).
# I don't think we need it for now.
> > No clue what a tag should be in the driver model.
>
> A tag is a way to associate data with a device. At present we do this
> with varoius built-in mechanisms (priv data, uclass-priv, plat, etc.)
> but with tags you can add something else.
Since this discussion thread is getting too long, I would like
to respin my RFC. How should I deal with your "event notification"
proposal?
-Takahiro Akashi
> Regards,
> Simon
More information about the U-Boot
mailing list