Mike Rapoport wrote:
> From: "Mike Rapoport (Microsoft)" <[email protected]>
>
> There are use cases, for example virtual machine hosts, that create
> "persistent" memory regions using memmap= option on x86 or dummy
> pmem-region device tree nodes on DT based systems.
>
> Both these options are inflexible because they create static regions and
> the layout of the "persistent" memory cannot be adjusted without reboot
> and sometimes they even require firmware update.
>
> Add a ramdax driver that allows creation of DIMM devices on top of
> E820_TYPE_PRAM regions and devicetree pmem-region nodes.
>
> The DIMMs support label space management on the "device" and provide a
> flexible way to access RAM using fsdax and devdax.
>
> Signed-off-by: Mike Rapoport (Mircosoft) <[email protected]>
> ---
> drivers/nvdimm/Kconfig | 17 +++
> drivers/nvdimm/Makefile | 1 +
> drivers/nvdimm/ramdax.c | 272 ++++++++++++++++++++++++++++++++++++++++
> 3 files changed, 290 insertions(+)
> create mode 100644 drivers/nvdimm/ramdax.c
>
> diff --git a/drivers/nvdimm/Kconfig b/drivers/nvdimm/Kconfig
> index fde3e17c836c..9ac96a7cd773 100644
> --- a/drivers/nvdimm/Kconfig
> +++ b/drivers/nvdimm/Kconfig
> @@ -97,6 +97,23 @@ config OF_PMEM
>
> Select Y if unsure.
>
> +config RAMDAX
> + tristate "Support persistent memory interfaces on RAM carveouts"
> + depends on OF || X86
I see no compile time dependency for CONFIG_OF. The one call to
dev_of_node() looks like it still builds in the CONFIG_OF=n case. For
CONFIG_X86 the situation is different because the kernel needs
infrastructure to build the device.
So maybe change the dependency to drop OF and make it:
depends on X86_PMEM_LEGACY if X86
> + select X86_PMEM_LEGACY_DEVICE
...and drop this select.
> + default LIBNVDIMM
> + help
> + Allows creation of DAX devices on RAM carveouts.
> +
> + Memory ranges that are manually specified by the
> + 'memmap=nn[KMG]!ss[KMG]' kernel command line or defined by dummy
> + pmem-region device tree nodes would be managed by this driver as DIMM
> + devices with support for dynamic layout of namespaces.
> + The driver can be bound to e820_pmem or pmem-region platform
> + devices using 'driver_override' device attribute.
Maybe some notes for details like:
* 128K stolen at the end of the memmap range
* supports 509 namespaces (see 'ndctl create-namespace --help')
* must be force bound via driver_override
[..]
> +static int ramdax_probe(struct platform_device *pdev)
> +{
> + static struct nvdimm_bus_descriptor nd_desc;
> + struct device *dev = &pdev->dev;
> + struct nvdimm_bus *nvdimm_bus;
> + struct device_node *np;
> + int rc = -ENXIO;
> +
> + nd_desc.provider_name = "ramdax";
> + nd_desc.module = THIS_MODULE;
> + nd_desc.ndctl = ramdax_ctl;
> + nvdimm_bus = nvdimm_bus_register(dev, &nd_desc);
> + if (!nvdimm_bus)
> + goto err;
> +
> + np = dev_of_node(&pdev->dev);
> + if (np)
> + rc = ramdax_probe_of(pdev, nvdimm_bus, np);
Hmm, I do not see any confirmation that this node is actually a
"pmem-region". If you attach the kernel to the wrong device I think you
get fireworks that could be avoided with a manual of_match_node() check
of the same device_id list as the of_pmem driver.
That still would not require a "depends on OF" given of_match_node()
compiles away in the CONFIG_OF=n case.
[..]
This looks good to me. With the above comments addressed you can add:
Reviewed-by: Dan Williams <[email protected]>
