Re: [RFC PATCH v3 07/16] cxl/mem: Implement polled mode mailbox

From: Jonathan Cameron
Date: Wed Jan 13 2021 - 13:28:09 EST


On Mon, 11 Jan 2021 14:51:11 -0800
Ben Widawsky <ben.widawsky@xxxxxxxxx> wrote:

> Provide enough functionality to utilize the mailbox of a memory device.
> The mailbox is used to interact with the firmware running on the memory
> device.
>
> The CXL specification defines separate capabilities for the mailbox and
> the memory device. While we can confirm the mailbox is ready, in order
> to actually interact with the memory device, you must also confirm the
> device's firmware is ready.
>
> Create a function to handle sending a command, optionally with a
> payload, to the memory device, polling on a result, and then optionally
> copying out the payload. The algorithm for doing this comes straight out
> of the CXL 2.0 specification.
>
> Primary mailboxes are capable of generating an interrupt when submitting
> a command in the background. That implementation is saved for a later
> time.
>
> Secondary mailboxes aren't implemented at this time.
>
> The flow is proven with one implemented command, "identify". Because the
> class code has already told the driver this is a memory device and the
> identify command is mandatory, it's safe to assume for sane devices that
> everything here will work.
>
> Signed-off-by: Ben Widawsky <ben.widawsky@xxxxxxxxx>

Hi Ben,

A few trivial things inline for this one but basically LGTM.

Jonathan

> ---
> drivers/cxl/cxl.h | 43 +++++++
> drivers/cxl/mem.c | 312 ++++++++++++++++++++++++++++++++++++++++++++++
> 2 files changed, 355 insertions(+)
>
> diff --git a/drivers/cxl/cxl.h b/drivers/cxl/cxl.h
> index a77286d04ce4..ca3fa496e21c 100644
> --- a/drivers/cxl/cxl.h
> +++ b/drivers/cxl/cxl.h
> @@ -32,9 +32,40 @@
> #define CXLDEV_MB_CAP_PAYLOAD_SIZE_MASK GENMASK(4, 0)
> #define CXLDEV_MB_CAP_PAYLOAD_SIZE_SHIFT 0
> #define CXLDEV_MB_CTRL_OFFSET 0x04
> +#define CXLDEV_MB_CTRL_DOORBELL BIT(0)
> #define CXLDEV_MB_CMD_OFFSET 0x08
> +#define CXLDEV_MB_CMD_COMMAND_OPCODE_SHIFT 0
> +#define CXLDEV_MB_CMD_COMMAND_OPCODE_MASK GENMASK(15, 0)
> +#define CXLDEV_MB_CMD_PAYLOAD_LENGTH_SHIFT 16
> +#define CXLDEV_MB_CMD_PAYLOAD_LENGTH_MASK GENMASK(36, 16)
> #define CXLDEV_MB_STATUS_OFFSET 0x10
> +#define CXLDEV_MB_STATUS_RET_CODE_SHIFT 32

Given use of FIELD_GET etc we shouldn't need both shift and the mask
in these defines (hopefully) and you aren't using it in this patch.

> +#define CXLDEV_MB_STATUS_RET_CODE_MASK GENMASK(47, 32)
> #define CXLDEV_MB_BG_CMD_STATUS_OFFSET 0x18
> +#define CXLDEV_MB_PAYLOAD_OFFSET 0x20
> +
> +/* Memory Device (CXL 2.0 - 8.2.8.5.1.1) */
> +#define CXLMDEV_STATUS_OFFSET 0x0
> +#define CXLMDEV_DEV_FATAL BIT(0)
> +#define CXLMDEV_FW_HALT BIT(1)
> +#define CXLMDEV_STATUS_MEDIA_STATUS_SHIFT 2
> +#define CXLMDEV_STATUS_MEDIA_STATUS_MASK GENMASK(3, 2)
> +#define CXLMDEV_MS_NOT_READY 0
> +#define CXLMDEV_MS_READY 1
> +#define CXLMDEV_MS_ERROR 2
> +#define CXLMDEV_MS_DISABLED 3
> +#define CXLMDEV_READY(status) \
> + (CXL_GET_FIELD(status, CXLMDEV_STATUS_MEDIA_STATUS) == CXLMDEV_MS_READY)
> +#define CXLMDEV_MBOX_IF_READY BIT(4)
> +#define CXLMDEV_RESET_NEEDED_SHIFT 5
> +#define CXLMDEV_RESET_NEEDED_MASK GENMASK(7, 5)
> +#define CXLMDEV_RESET_NEEDED_NOT 0
> +#define CXLMDEV_RESET_NEEDED_COLD 1
> +#define CXLMDEV_RESET_NEEDED_WARM 2
> +#define CXLMDEV_RESET_NEEDED_HOT 3
> +#define CXLMDEV_RESET_NEEDED_CXL 4
> +#define CXLMDEV_RESET_NEEDED(status) \
> + (CXL_GET_FIELD(status, CXLMDEV_RESET_NEEDED) != CXLMDEV_RESET_NEEDED_NOT)
>
> /**
> * struct cxl_mem - A CXL memory device
> @@ -45,6 +76,16 @@ struct cxl_mem {
> struct pci_dev *pdev;
> void __iomem *regs;
>
> + struct {
> + struct range range;
> + } pmem;
> +
> + struct {
> + struct range range;
> + } ram;
> +
> + char firmware_version[0x10];
> +
> /* Cap 0001h - CXL_CAP_CAP_ID_DEVICE_STATUS */
> struct {
> void __iomem *regs;
> @@ -52,6 +93,7 @@ struct cxl_mem {
>
> /* Cap 0002h - CXL_CAP_CAP_ID_PRIMARY_MAILBOX */
> struct {
> + struct mutex mutex; /* Protects device mailbox and firmware */
> void __iomem *regs;
> size_t payload_size;
> } mbox;
> @@ -90,6 +132,7 @@ struct cxl_mem {
>
> cxl_reg(status);
> cxl_reg(mbox);
> +cxl_reg(mem);
>
> #define cxl_payload_regs(cxlm) \
> ((void __iomem *)(cxlm)->mbox.regs + CXLDEV_MB_PAYLOAD_OFFSET)
> diff --git a/drivers/cxl/mem.c b/drivers/cxl/mem.c
> index 8da9f4a861ea..e9ba97bbd7b9 100644
> --- a/drivers/cxl/mem.c
> +++ b/drivers/cxl/mem.c
> @@ -1,5 +1,6 @@
> // SPDX-License-Identifier: GPL-2.0-only
> /* Copyright(c) 2020 Intel Corporation. All rights reserved. */
> +#include <linux/sched/clock.h>
> #include <linux/module.h>
> #include <linux/pci.h>
> #include <linux/io.h>
> @@ -7,6 +8,248 @@
> #include "pci.h"
> #include "cxl.h"
>
> +#define cxl_doorbell_busy(cxlm) \
> + (cxl_read_mbox_reg32(cxlm, CXLDEV_MB_CTRL_OFFSET) & \
> + CXLDEV_MB_CTRL_DOORBELL)
> +
> +#define CXL_MAILBOX_TIMEOUT_US 2000
> +
> +enum opcode {
> + CXL_MBOX_OP_IDENTIFY = 0x4000,
> + CXL_MBOX_OP_MAX = 0x10000
> +};
> +
> +/**
> + * struct mbox_cmd - A command to be submitted to hardware.
> + * @opcode: (input) The command set and command submitted to hardware.
> + * @payload: (input/output) Pointer to the input and output payload.
> + * Payload can be NULL if the caller wants to populate the payload
> + * registers themselves (potentially avoiding a copy).
> + * @size_in: (input) Number of bytes to load from @payload.
> + * @size_out:
> + * - (input) Number of bytes allocated to load into @payload.
> + * - (output) Number of bytes loaded into @payload.
> + * @return_code: (output) Error code returned from hardware.
> + *
> + * This is the primary mechanism used to send commands to the hardware.
> + * All the fields except @payload correspond exactly to the fields described in
> + * Command Register section of the CXL 2.0 spec (8.2.8.4.5). @payload
> + * corresponds to the Command Payload Registers (8.2.8.4.8).
> + */
> +struct mbox_cmd {
> + u16 opcode;
> + void *payload;
> + size_t size_in;
> + size_t size_out;
> + u16 return_code;
> +};
> +
> +static int cxl_mem_wait_for_doorbell(struct cxl_mem *cxlm)
> +{
> + const int timeout = msecs_to_jiffies(CXL_MAILBOX_TIMEOUT_US);
> + const unsigned long start = jiffies;
> + unsigned long end = start;
> +
> + while (cxl_doorbell_busy(cxlm)) {
> + end = jiffies;
> +
> + if (time_after(end, start + timeout)) {
> + /* Check again in case preempted before timeout test */
> + if (!cxl_doorbell_busy(cxlm))
> + break;
> + return -ETIMEDOUT;
> + }
> + cpu_relax();
> + }
> +
> + dev_dbg(&cxlm->pdev->dev, "Doorbell wait took %dms",
> + jiffies_to_msecs(end) - jiffies_to_msecs(start));
> + return 0;
> +}
> +
> +static void cxl_mem_mbox_timeout(struct cxl_mem *cxlm,
> + struct mbox_cmd *mbox_cmd)
> +{
> + dev_warn(&cxlm->pdev->dev, "Mailbox command timed out\n");
> + dev_info(&cxlm->pdev->dev,
> + "\topcode: 0x%04x\n"
> + "\tpayload size: %zub\n",
> + mbox_cmd->opcode, mbox_cmd->size_in);
> + print_hex_dump_debug("Payload ", DUMP_PREFIX_OFFSET, 16, 1,
> + mbox_cmd->payload, mbox_cmd->size_in, true);
> +
> + /* Here's a good place to figure out if a device reset is needed */
> +}
> +
> +/**
> + * cxl_mem_mbox_send_cmd() - Send a mailbox command to a memory device.
> + * @cxlm: The CXL memory device to communicate with.
> + * @mbox_cmd: Command to send to the memory device.
> + *
> + * Context: Any context. Expects mbox_lock to be held.
> + * Return: -ETIMEDOUT if timeout occurred waiting for completion. 0 on success.
> + * Caller should check the return code in @mbox_cmd to make sure it
> + * succeeded.
> + *
> + * This is a generic form of the CXL mailbox send command, thus the only I/O
> + * operations used are cxl_read_mbox_reg(). Memory devices, and perhaps other
> + * types of CXL devices may have further information available upon error
> + * conditions.
> + *
> + * FIXME: As stated above, references to &struct cxl_mem should be changed to a
> + * more generic cxl structure when needed.

I'd flip that to a TODO. If this driver is ready to go before other usecases
perhaps need to clear this doesn't 'need' to be fixed.

> + */
> +static int cxl_mem_mbox_send_cmd(struct cxl_mem *cxlm,
> + struct mbox_cmd *mbox_cmd)
> +{
> + u64 cmd_reg, status_reg;
> + size_t out_len;
> + int rc;
> +
> + lockdep_assert_held(&cxlm->mbox.mutex);
> +
> + /*
> + * Here are the steps from 8.2.8.4 of the CXL 2.0 spec.
> + * 1. Caller reads MB Control Register to verify doorbell is clear
> + * 2. Caller writes Command Register
> + * 3. Caller writes Command Payload Registers if input payload is non-empty
> + * 4. Caller writes MB Control Register to set doorbell
> + * 5. Caller either polls for doorbell to be clear or waits for interrupt if configured
> + * 6. Caller reads MB Status Register to fetch Return code
> + * 7. If command successful, Caller reads Command Register to get Payload Length
> + * 8. If output payload is non-empty, host reads Command Payload Registers
> + */
> +
> + /* #1 */
> + WARN_ON(cxl_doorbell_busy(cxlm));

If it was busy, should we proceed anyway?

> +
> + cmd_reg = CXL_SET_FIELD(mbox_cmd->opcode, CXLDEV_MB_CMD_COMMAND_OPCODE);
> + if (mbox_cmd->size_in) {
> + cmd_reg |= CXL_SET_FIELD(mbox_cmd->size_in,
> + CXLDEV_MB_CMD_PAYLOAD_LENGTH);
> + if (mbox_cmd->payload)
> + memcpy_toio(cxl_payload_regs(cxlm), mbox_cmd->payload,
> + mbox_cmd->size_in);

I'm sure this is fine, but technically isn't this #3 then #2?

> + }
> +
> + /* #2, #3 */
> + cxl_write_mbox_reg64(cxlm, CXLDEV_MB_CMD_OFFSET, cmd_reg);
> +
> + /* #4 */
> + dev_dbg(&cxlm->pdev->dev, "Sending command\n");
> + cxl_write_mbox_reg32(cxlm, CXLDEV_MB_CTRL_OFFSET,
> + CXLDEV_MB_CTRL_DOORBELL);
> +
> + /* #5 */
> + rc = cxl_mem_wait_for_doorbell(cxlm);
> + if (rc == -ETIMEDOUT) {
> + cxl_mem_mbox_timeout(cxlm, mbox_cmd);
> + return rc;
> + }
> +
> + /* #6 */
> + status_reg = cxl_read_mbox_reg64(cxlm, CXLDEV_MB_STATUS_OFFSET);
> + mbox_cmd->return_code =
> + CXL_GET_FIELD(status_reg, CXLDEV_MB_STATUS_RET_CODE);
> +
> + if (mbox_cmd->return_code != 0) {
> + dev_dbg(&cxlm->pdev->dev, "Mailbox operation had an error\n");
> + return 0;
> + }
> +
> + /* #7 */
> + cmd_reg = cxl_read_mbox_reg64(cxlm, CXLDEV_MB_CMD_OFFSET);
> + out_len = CXL_GET_FIELD(cmd_reg, CXLDEV_MB_CMD_PAYLOAD_LENGTH);
> + mbox_cmd->size_out = out_len;
> +
> + /* #8 */
> + if (out_len && mbox_cmd->payload)
> + memcpy_fromio(mbox_cmd->payload, cxl_payload_regs(cxlm),
> + mbox_cmd->size_out);
> +
> + return 0;
> +}
> +
> +/**
> + * cxl_mem_mbox_get() - Acquire exclusive access to the mailbox.
> + * @cxlm: The memory device to gain access to.
> + *
> + * Context: Any context. Takes the mbox_lock.
> + * Return: 0 if exclusive access was acquired.
> + */
> +static int cxl_mem_mbox_get(struct cxl_mem *cxlm)
> +{
> + u64 md_status;
> + int rc = -EBUSY;

Trivial: Always set below so no need to init here.

> +
> + mutex_lock_io(&cxlm->mbox.mutex);
> +
> + /*
> + * XXX: There is some amount of ambiguity in the 2.0 version of the spec
> + * around the mailbox interface ready (8.2.8.5.1.1). The purpose of the
> + * bit is to allow firmware running on the device to notify us that it's
> + * ready to receive commands. It is unclear if the bit needs to be read
> + * every time one tries to use the mailbox, ie. the firmware can switch
> + * it on and off as needed. Second, there is no defined timeout for
> + * mailbox ready, like there is for the doorbell interface.
> + *
> + * As such, we make the following assumptions:
> + * 1. The firmware might toggle the Mailbox Interface Ready bit, and so
> + * we check it for every command.
> + * 2. If the doorbell is clear, the firmware should have first set the
> + * Mailbox Interface Ready bit. Therefore, waiting for the doorbell
> + * to be ready is a sufficient amount of time.
> + */
> + rc = cxl_mem_wait_for_doorbell(cxlm);
> + if (rc) {
> + dev_warn(&cxlm->pdev->dev, "Mailbox interface not ready\n");
> + goto out;
> + }
> +
> + md_status = cxl_read_mem_reg64(cxlm, CXLMDEV_STATUS_OFFSET);
> + if (md_status & CXLMDEV_MBOX_IF_READY && CXLMDEV_READY(md_status)) {
> + /*
> + * Hardware shouldn't allow a ready status but also have failure
> + * bits set. Spit out an error, this should be a bug report
> + */
> + if (md_status & CXLMDEV_DEV_FATAL) {
> + dev_err(&cxlm->pdev->dev,
> + "CXL device reporting ready and fatal\n");
> + rc = -EFAULT;
> + goto out;
> + }
> + if (md_status & CXLMDEV_FW_HALT) {
> + dev_err(&cxlm->pdev->dev,
> + "CXL device reporting ready and halted\n");
> + rc = -EFAULT;
> + goto out;
> + }
> + if (CXLMDEV_RESET_NEEDED(md_status)) {
> + dev_err(&cxlm->pdev->dev,
> + "CXL device reporting ready and reset needed\n");
> + rc = -EFAULT;
> + goto out;
> + }
> +
> + return 0;
> + }
> +
> +out:
> + mutex_unlock(&cxlm->mbox.mutex);
> + return rc;
> +}
> +
> +/**
> + * cxl_mem_mbox_put() - Release exclusive access to the mailbox.
> + * @cxlm: The CXL memory device to communicate with.
> + *
> + * Context: Any context. Expects mbox_lock to be held.
> + */
> +static void cxl_mem_mbox_put(struct cxl_mem *cxlm)
> +{
> + mutex_unlock(&cxlm->mbox.mutex);
> +}
> +
> /**
> * cxl_mem_setup_regs() - Setup necessary MMIO.
> * @cxlm: The CXL memory device to communicate with.
> @@ -135,6 +378,8 @@ static struct cxl_mem *cxl_mem_create(struct pci_dev *pdev, u32 reg_lo,
> return NULL;
> }
>
> + mutex_init(&cxlm->mbox.mutex);
> +
> regs = pcim_iomap_table(pdev)[bar];
> cxlm->pdev = pdev;
> cxlm->regs = regs + offset;
> @@ -167,6 +412,69 @@ static int cxl_mem_dvsec(struct pci_dev *pdev, int dvsec)
> return 0;
> }
>
> +/**
> + * cxl_mem_identify() - Send the IDENTIFY command to the device.
> + * @cxlm: The device to identify.
> + *
> + * Return: 0 if identify was executed successfully.
> + *
> + * This will dispatch the identify command to the device and on success populate
> + * structures to be exported to sysfs.
> + */
> +static int cxl_mem_identify(struct cxl_mem *cxlm)
> +{
> + struct cxl_mbox_identify {
> + char fw_revision[0x10];
> + __le64 total_capacity;
> + __le64 volatile_capacity;
> + __le64 persistent_capacity;
> + __le64 partition_align;
> + __le16 info_event_log_size;
> + __le16 warning_event_log_size;
> + __le16 failure_event_log_size;
> + __le16 fatal_event_log_size;
> + __le32 lsa_size;
> + u8 poison_list_max_mer[3];
> + __le16 inject_poison_limit;
> + u8 poison_caps;
> + u8 qos_telemetry_caps;
> + } __packed id;
> + struct mbox_cmd mbox_cmd;
> + int rc;
> +
> + /* Retrieve initial device memory map */
> + rc = cxl_mem_mbox_get(cxlm);
> + if (rc)
> + return rc;
> +
> + mbox_cmd = (struct mbox_cmd){
> + .opcode = CXL_MBOX_OP_IDENTIFY,
> + .payload = &id,
> + .size_in = 0,
> + };
> + rc = cxl_mem_mbox_send_cmd(cxlm, &mbox_cmd);

Description of this function has:
"Caller should check the return code in @mbox_cmd to make sure it succeeded."

I'm not seeing that here.

> + cxl_mem_mbox_put(cxlm);
> + if (rc)
> + return rc;
> +
> + if (mbox_cmd.size_out != sizeof(id))
> + return -ENXIO;
> +
> + /*
> + * TODO: enumerate DPA map, as 'ram' and 'pmem' do not alias.
> + * For now, only the capacity is exported in sysfs
> + */
> + cxlm->ram.range.start = 0;
> + cxlm->ram.range.end = le64_to_cpu(id.volatile_capacity) - 1;
> +
> + cxlm->pmem.range.start = 0;
> + cxlm->pmem.range.end = le64_to_cpu(id.persistent_capacity) - 1;
> +
> + memcpy(cxlm->firmware_version, id.fw_revision, sizeof(id.fw_revision));
> +
> + return rc;
> +}
> +
> static int cxl_mem_probe(struct pci_dev *pdev, const struct pci_device_id *id)
> {
> struct device *dev = &pdev->dev;
> @@ -222,6 +530,10 @@ static int cxl_mem_probe(struct pci_dev *pdev, const struct pci_device_id *id)
> if (rc)
> goto err;
>
> + rc = cxl_mem_identify(cxlm);
> + if (rc)
> + goto err;
> +
> pci_set_drvdata(pdev, cxlm);
> return 0;
>