From: "Edward A. James" <[email protected]>
This series adds two FSI-based device drivers; one for the SBEFIFO and one for
the On-Chip Controller (OCC).
IBM POWER9 processors contain some embedded hardware and software bits
collectively referred to as the self boot engine (SBE). One role of
the SBE is to act as a proxy that provides access to the registers of
the POWER chip from other (embedded) systems.
The POWER9 chip contains a hardware frontend for communicating with
the SBE from remote systems called the SBEFIFO. The SBEFIFO logic
is contained within an FSI CFAM and as such the driver implements an
FSI bus device.
The SBE expects to communicate using a defined wire protocol; however,
the driver knows nothing of the protocol and only provides raw access
to the fifo device to userspace applications wishing to communicate with
the SBE using the wire protocol.
The SBEFIFO consists of two hardware fifos. The upstream fifo is used
by the driver to transfer data to the SBE on the POWER chip, from the
system hosting the driver. The downstream fifo is used by the driver to
transfer data from the SBE on the power chip to the system hosting the
driver.
The OCC is a device embedded on a POWER processor that collects and
aggregates sensor data from the processor and system. The OCC can
provide the raw sensor data as well as perform thermal and power
management on the system.
This driver provides an atomic communications channel between a service
processor (e.g. a BMC) and the OCC. The driver is dependent on the FSI
SBEFIFO driver to get hardware access through the SBE to the OCC SRAM.
Commands are issued to the SBE to send or fetch data to the SRAM.
The format of communications to the OCC is writing a command to SRAM,
followed by a sending a "doorbell" or attention to the OCC, followed by
reading the response from OCC. All of this takes place atomically so
that multiple users don't collide in the SRAM.
Changes since v4:
* Use irqsave/irqrestore for spinlocks.
* Put the OCC driver in the same patchset as it doesn't build without SBEFIFO.
* Switch to reference counting for OCC clients.
Changes since v3:
* Add reset procedure and use it if there is data in the FIFO at probe time.
* Add timeout for waiting for data to appear in the FIFO; if the SBE isn't
running, then previously we would wait forever.
* Fix remove() order of operations for both drivers.
* Fix xfr memory leak for SBEFIFO.
* Formatting fixes.
Changes since v2:
* Rename occ.c and occ.h to fsi-occ.c and fsi-occ.h
* Improved remove() ordering in both drivers.
* Added cancel functionality to OCC driver to make sure no xfrs started during
remove().
* Fix spin_unlock with spin_unlock_irq in OCC driver.
* Fix list_first_entry with list_first_entry_or_null in OCC worker function.
* Add OCC response definitions to OCC include file.
* Handle probe() failures better.
Changes since v1:
* Split bindings into separate patch and added SBEFIFO device binding
* Fixed #includes
* Fix SBEFIFO race condition between write() and poll_timer().
* Followed Rob's suggestion to just create one platform device for hwmon
driver, instead of using the device tree.
* Also check for "command in progress" response from OCC and try a while
Edward A. James (8):
dt-bindings: fsi: Add SBEFIFO documentation
drivers/fsi: Add SBEFIFO FSI client device driver
drivers/fsi: sbefifo: Add miscdevice
drivers/fsi: sbefifo: Add in-kernel API
dt-bindings: fsi: Add OCC documentation
drivers/fsi: Add On-Chip Controller (OCC) driver
drivers/fsi: occ: Add miscdevice
drivers/fsi: occ: Add in-kernel API
.../devicetree/bindings/fsi/ibm,p9-occ.txt | 18 +
.../devicetree/bindings/fsi/ibm,p9-sbefifo.txt | 35 +
drivers/fsi/Kconfig | 17 +
drivers/fsi/Makefile | 2 +
drivers/fsi/fsi-occ.c | 876 +++++++++++++++++
drivers/fsi/fsi-sbefifo.c | 1027 ++++++++++++++++++++
include/linux/fsi-occ.h | 41 +
include/linux/fsi-sbefifo.h | 30 +
8 files changed, 2046 insertions(+)
create mode 100644 Documentation/devicetree/bindings/fsi/ibm,p9-occ.txt
create mode 100644 Documentation/devicetree/bindings/fsi/ibm,p9-sbefifo.txt
create mode 100644 drivers/fsi/fsi-occ.c
create mode 100644 drivers/fsi/fsi-sbefifo.c
create mode 100644 include/linux/fsi-occ.h
create mode 100644 include/linux/fsi-sbefifo.h
--
1.8.3.1
From 1584643918012931243@xxx Tue Nov 21 03:16:45 +0000 2017
X-GM-THRID: 1584342442584787748
X-Gmail-Labels: Inbox,Category Forums,HistoricalUnread
From: "Edward A. James" <[email protected]>
The OCC is a device embedded on a POWER processor that collects and
aggregates sensor data from the processor and system. The OCC can
provide the raw sensor data as well as perform thermal and power
management on the system.
This driver provides an atomic communications channel between a service
processor (e.g. a BMC) and the OCC. The driver is dependent on the FSI
SBEFIFO driver to get hardware access through the SBE to the OCC SRAM.
Commands are issued to the SBE to send or fetch data to the SRAM.
The format of communications to the OCC is writing a command to SRAM,
followed by a sending a "doorbell" or attention to the OCC, followed by
reading the response from OCC. All of this takes place atomically so
that multiple users don't collide in the SRAM.
Signed-off-by: Edward A. James <[email protected]>
---
drivers/fsi/Kconfig | 10 +
drivers/fsi/Makefile | 1 +
drivers/fsi/fsi-occ.c | 518 ++++++++++++++++++++++++++++++++++++++++++++++++++
3 files changed, 529 insertions(+)
create mode 100644 drivers/fsi/fsi-occ.c
diff --git a/drivers/fsi/Kconfig b/drivers/fsi/Kconfig
index 8c4d903..8ed4602 100644
--- a/drivers/fsi/Kconfig
+++ b/drivers/fsi/Kconfig
@@ -40,6 +40,16 @@ config FSI_SBEFIFO
a pipe-like FSI device for communicating with the self boot engine
(SBE) on POWER processors.
+config FSI_OCC
+ tristate "OCC SBEFIFO client device driver"
+ depends on FSI_SBEFIFO
+ ---help---
+ This option enables an SBEFIFO based On-Chip Controller (OCC) device
+ driver. The OCC is a device embedded on a POWER processor that collects
+ and aggregates sensor data from the processor and system. The OCC can
+ provide the raw sensor data as well as perform thermal and power
+ management on the system.
+
endif
endmenu
diff --git a/drivers/fsi/Makefile b/drivers/fsi/Makefile
index 851182e..75fdc6d 100644
--- a/drivers/fsi/Makefile
+++ b/drivers/fsi/Makefile
@@ -4,3 +4,4 @@ obj-$(CONFIG_FSI_MASTER_HUB) += fsi-master-hub.o
obj-$(CONFIG_FSI_MASTER_GPIO) += fsi-master-gpio.o
obj-$(CONFIG_FSI_SCOM) += fsi-scom.o
obj-$(CONFIG_FSI_SBEFIFO) += fsi-sbefifo.o
+obj-$(CONFIG_FSI_OCC) += fsi-occ.o
diff --git a/drivers/fsi/fsi-occ.c b/drivers/fsi/fsi-occ.c
new file mode 100644
index 0000000..99dbfa3
--- /dev/null
+++ b/drivers/fsi/fsi-occ.c
@@ -0,0 +1,518 @@
+/*
+ * Copyright 2017 IBM Corp.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <asm/unaligned.h>
+#include <linux/device.h>
+#include <linux/err.h>
+#include <linux/errno.h>
+#include <linux/fs.h>
+#include <linux/fsi-sbefifo.h>
+#include <linux/idr.h>
+#include <linux/kernel.h>
+#include <linux/list.h>
+#include <linux/module.h>
+#include <linux/mutex.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/spinlock.h>
+#include <linux/uaccess.h>
+#include <linux/wait.h>
+#include <linux/workqueue.h>
+
+#define OCC_SRAM_BYTES 4096
+#define OCC_CMD_DATA_BYTES 4090
+#define OCC_RESP_DATA_BYTES 4089
+
+#define OCC_RESP_CMD_IN_PRG 0xFF
+
+#define OCC_TIMEOUT_MS 1000
+#define OCC_CMD_IN_PRG_WAIT_MS 50
+
+struct occ {
+ struct device *sbefifo;
+ char name[32];
+ int idx;
+ struct list_head xfrs;
+ spinlock_t list_lock; /* lock access to the xfrs list */
+ struct mutex occ_lock; /* lock access to the hardware */
+ struct work_struct work;
+ bool cancel;
+};
+
+#define to_occ(x) container_of((x), struct occ, mdev)
+
+struct occ_response {
+ u8 seq_no;
+ u8 cmd_type;
+ u8 return_status;
+ __be16 data_length;
+ u8 data[OCC_RESP_DATA_BYTES];
+ __be16 checksum;
+} __packed;
+
+/*
+ * transfer flags are NOT mutually exclusive
+ *
+ * Initial flags are none; transfer is created and queued from write(). All
+ * flags are cleared when the transfer is completed by closing the file or
+ * reading all of the available response data.
+ * XFR_IN_PROGRESS is set when a transfer is started from occ_worker_putsram,
+ * and cleared if the transfer fails or occ_worker_getsram completes.
+ * XFR_COMPLETE is set when a transfer fails or finishes occ_worker_getsram.
+ * XFR_CANCELED is set when the transfer's client is released.
+ */
+enum {
+ XFR_IN_PROGRESS,
+ XFR_COMPLETE,
+ XFR_CANCELED,
+};
+
+struct occ_xfr {
+ struct list_head link;
+ int rc;
+ u8 buf[OCC_SRAM_BYTES];
+ size_t cmd_data_length;
+ size_t resp_data_length;
+ unsigned long flags;
+};
+
+/*
+ * client flags
+ *
+ * CLIENT_NONBLOCKING is set during open() if the file was opened with the
+ * O_NONBLOCK flag.
+ * CLIENT_XFR_PENDING is set during write() and cleared when all data has been
+ * read.
+ */
+enum {
+ CLIENT_NONBLOCKING,
+ CLIENT_XFR_PENDING,
+};
+
+struct occ_client {
+ struct kref kref;
+ struct occ *occ;
+ struct occ_xfr xfr;
+ spinlock_t lock; /* lock access to the client state */
+ wait_queue_head_t wait;
+ size_t read_offset;
+ unsigned long flags;
+};
+
+#define to_client(x) container_of((x), struct occ_client, xfr)
+
+static struct workqueue_struct *occ_wq;
+
+static DEFINE_IDA(occ_ida);
+
+static void occ_get_client(struct occ_client *client)
+{
+ kref_get(&client->kref);
+}
+
+static void occ_client_release(struct kref *kref)
+{
+ struct occ_client *client = container_of(kref, struct occ_client,
+ kref);
+
+ kfree(client);
+}
+
+static void occ_put_client(struct occ_client *client)
+{
+ kref_put(&client->kref, occ_client_release);
+}
+
+static int occ_write_sbefifo(struct sbefifo_client *client, const char *buf,
+ ssize_t len)
+{
+ int rc;
+ ssize_t total = 0;
+
+ do {
+ rc = sbefifo_drv_write(client, &buf[total], len - total);
+ if (rc < 0)
+ return rc;
+ else if (!rc)
+ break;
+
+ total += rc;
+ } while (total < len);
+
+ return (total == len) ? 0 : -ENOSPC;
+}
+
+static int occ_read_sbefifo(struct sbefifo_client *client, char *buf,
+ ssize_t len)
+{
+ int rc;
+ ssize_t total = 0;
+
+ do {
+ rc = sbefifo_drv_read(client, &buf[total], len - total);
+ if (rc < 0)
+ return rc;
+ else if (!rc)
+ break;
+
+ total += rc;
+ } while (total < len);
+
+ return (total == len) ? 0 : -ENODATA;
+}
+
+static int occ_getsram(struct device *sbefifo, u32 address, u8 *data,
+ ssize_t len)
+{
+ int rc;
+ u8 *resp;
+ __be32 buf[5];
+ u32 data_len = ((len + 7) / 8) * 8; /* must be multiples of 8 B */
+ struct sbefifo_client *client;
+
+ /*
+ * Magic sequence to do SBE getsram command. SBE will fetch data from
+ * specified SRAM address.
+ */
+ buf[0] = cpu_to_be32(0x5);
+ buf[1] = cpu_to_be32(0xa403);
+ buf[2] = cpu_to_be32(1);
+ buf[3] = cpu_to_be32(address);
+ buf[4] = cpu_to_be32(data_len);
+
+ client = sbefifo_drv_open(sbefifo, 0);
+ if (!client)
+ return -ENODEV;
+
+ rc = occ_write_sbefifo(client, (const char *)buf, sizeof(buf));
+ if (rc)
+ goto done;
+
+ resp = kzalloc(data_len, GFP_KERNEL);
+ if (!resp) {
+ rc = -ENOMEM;
+ goto done;
+ }
+
+ rc = occ_read_sbefifo(client, (char *)resp, data_len);
+ if (rc)
+ goto free;
+
+ /* check for good response */
+ rc = occ_read_sbefifo(client, (char *)buf, 8);
+ if (rc)
+ goto free;
+
+ if ((be32_to_cpu(buf[0]) == data_len) &&
+ (be32_to_cpu(buf[1]) == 0xC0DEA403))
+ memcpy(data, resp, len);
+ else
+ rc = -EBADMSG;
+
+free:
+ kfree(resp);
+
+done:
+ sbefifo_drv_release(client);
+ return rc;
+}
+
+static int occ_putsram(struct device *sbefifo, u32 address, u8 *data,
+ ssize_t len)
+{
+ int rc;
+ __be32 *buf;
+ u32 data_len = ((len + 7) / 8) * 8; /* must be multiples of 8 B */
+ size_t cmd_len = data_len + 20;
+ struct sbefifo_client *client;
+
+ buf = kzalloc(cmd_len, GFP_KERNEL);
+ if (!buf)
+ return -ENOMEM;
+
+ /*
+ * Magic sequence to do SBE putsram command. SBE will transfer
+ * data to specified SRAM address.
+ */
+ buf[0] = cpu_to_be32(0x5 + (data_len / 4));
+ buf[1] = cpu_to_be32(0xa404);
+ buf[2] = cpu_to_be32(1);
+ buf[3] = cpu_to_be32(address);
+ buf[4] = cpu_to_be32(data_len);
+
+ memcpy(&buf[5], data, len);
+
+ client = sbefifo_drv_open(sbefifo, 0);
+ if (!client) {
+ rc = -ENODEV;
+ goto free;
+ }
+
+ rc = occ_write_sbefifo(client, (const char *)buf, cmd_len);
+ if (rc)
+ goto done;
+
+ rc = occ_read_sbefifo(client, (char *)buf, 8);
+ if (rc)
+ goto done;
+
+ /* check for good response */
+ if ((be32_to_cpu(buf[0]) != data_len) ||
+ (be32_to_cpu(buf[1]) != 0xC0DEA404))
+ rc = -EBADMSG;
+
+done:
+ sbefifo_drv_release(client);
+free:
+ kfree(buf);
+ return rc;
+}
+
+static int occ_trigger_attn(struct device *sbefifo)
+{
+ int rc;
+ __be32 buf[6];
+ struct sbefifo_client *client;
+
+ /*
+ * Magic sequence to do SBE putscom command. SBE will write 8 bytes to
+ * specified SCOM address.
+ */
+ buf[0] = cpu_to_be32(0x6);
+ buf[1] = cpu_to_be32(0xa202);
+ buf[2] = 0;
+ buf[3] = cpu_to_be32(0x6D035);
+ buf[4] = cpu_to_be32(0x20010000); /* trigger occ attention */
+ buf[5] = 0;
+
+ client = sbefifo_drv_open(sbefifo, 0);
+ if (!client)
+ return -ENODEV;
+
+ rc = occ_write_sbefifo(client, (const char *)buf, sizeof(buf));
+ if (rc)
+ goto done;
+
+ rc = occ_read_sbefifo(client, (char *)buf, 8);
+ if (rc)
+ goto done;
+
+ /* check for good response */
+ if ((be32_to_cpu(buf[0]) != 0xC0DEA202) ||
+ (be32_to_cpu(buf[1]) & 0x0FFFFFFF))
+ rc = -EBADMSG;
+
+done:
+ sbefifo_drv_release(client);
+
+ return rc;
+}
+
+static void occ_worker(struct work_struct *work)
+{
+ int rc = 0, empty;
+ u16 resp_data_length;
+ unsigned long flags;
+ unsigned long start;
+ const unsigned long timeout = msecs_to_jiffies(OCC_TIMEOUT_MS);
+ const long int wait_time = msecs_to_jiffies(OCC_CMD_IN_PRG_WAIT_MS);
+ struct occ_xfr *xfr;
+ struct occ_response *resp;
+ struct occ_client *client;
+ struct occ *occ = container_of(work, struct occ, work);
+ struct device *sbefifo = occ->sbefifo;
+
+again:
+ if (occ->cancel)
+ return;
+
+ spin_lock_irqsave(&occ->list_lock, flags);
+
+ xfr = list_first_entry_or_null(&occ->xfrs, struct occ_xfr, link);
+ if (!xfr) {
+ spin_unlock_irqrestore(&occ->list_lock, flags);
+ return;
+ }
+
+ client = to_client(xfr);
+ occ_get_client(client);
+ resp = (struct occ_response *)xfr->buf;
+ set_bit(XFR_IN_PROGRESS, &xfr->flags);
+
+ spin_unlock_irqrestore(&occ->list_lock, flags);
+ mutex_lock(&occ->occ_lock);
+
+ start = jiffies;
+
+ /* write occ command */
+ rc = occ_putsram(sbefifo, 0xFFFBE000, xfr->buf,
+ xfr->cmd_data_length);
+ if (rc)
+ goto done;
+
+ rc = occ_trigger_attn(sbefifo);
+ if (rc)
+ goto done;
+
+ /* read occ response */
+ do {
+ rc = occ_getsram(sbefifo, 0xFFFBF000, xfr->buf, 8);
+ if (rc)
+ goto done;
+
+ if (resp->return_status == OCC_RESP_CMD_IN_PRG) {
+ rc = -EALREADY;
+
+ if (time_after(jiffies, start + timeout))
+ break;
+
+ set_current_state(TASK_INTERRUPTIBLE);
+ schedule_timeout(wait_time);
+ }
+ } while (rc);
+
+ resp_data_length = get_unaligned_be16(&resp->data_length);
+ if (resp_data_length > OCC_RESP_DATA_BYTES) {
+ rc = -EMSGSIZE;
+ goto done;
+ }
+
+ if (resp_data_length > 1) {
+ /* already got 3 bytes resp, also need 2 bytes checksum */
+ rc = occ_getsram(sbefifo, 0xFFFBF008, &xfr->buf[8],
+ resp_data_length - 1);
+ if (rc)
+ goto done;
+ }
+
+ xfr->resp_data_length = resp_data_length + 7;
+
+done:
+ mutex_unlock(&occ->occ_lock);
+
+ xfr->rc = rc;
+ set_bit(XFR_COMPLETE, &xfr->flags);
+
+ spin_lock_irqsave(&occ->list_lock, flags);
+
+ clear_bit(XFR_IN_PROGRESS, &xfr->flags);
+ list_del(&xfr->link);
+ empty = list_empty(&occ->xfrs);
+
+ spin_unlock_irqrestore(&occ->list_lock, flags);
+
+ wake_up_interruptible(&client->wait);
+ occ_put_client(client);
+
+ if (!empty)
+ goto again;
+}
+
+static int occ_probe(struct platform_device *pdev)
+{
+ int rc;
+ u32 reg;
+ struct occ *occ;
+ struct device *dev = &pdev->dev;
+
+ occ = devm_kzalloc(dev, sizeof(*occ), GFP_KERNEL);
+ if (!occ)
+ return -ENOMEM;
+
+ occ->sbefifo = dev->parent;
+ INIT_LIST_HEAD(&occ->xfrs);
+ spin_lock_init(&occ->list_lock);
+ mutex_init(&occ->occ_lock);
+ INIT_WORK(&occ->work, occ_worker);
+
+ if (dev->of_node) {
+ rc = of_property_read_u32(dev->of_node, "reg", ®);
+ if (!rc) {
+ /* make sure we don't have a duplicate from dts */
+ occ->idx = ida_simple_get(&occ_ida, reg, reg + 1,
+ GFP_KERNEL);
+ if (occ->idx < 0)
+ occ->idx = ida_simple_get(&occ_ida, 1, INT_MAX,
+ GFP_KERNEL);
+ } else {
+ occ->idx = ida_simple_get(&occ_ida, 1, INT_MAX,
+ GFP_KERNEL);
+ }
+ } else {
+ occ->idx = ida_simple_get(&occ_ida, 1, INT_MAX, GFP_KERNEL);
+ }
+
+ platform_set_drvdata(pdev, occ);
+
+ return 0;
+}
+
+static int occ_remove(struct platform_device *pdev)
+{
+ unsigned long flags;
+ struct occ *occ = platform_get_drvdata(pdev);
+ struct occ_xfr *xfr;
+ struct occ_client *client;
+
+ occ->cancel = true;
+
+ spin_lock_irqsave(&occ->list_lock, flags);
+ list_for_each_entry(xfr, &occ->xfrs, link) {
+ client = to_client(xfr);
+ wake_up_all(&client->wait);
+ }
+ spin_unlock_irqrestore(&occ->list_lock, flags);
+
+ cancel_work_sync(&occ->work);
+
+ ida_simple_remove(&occ_ida, occ->idx);
+
+ return 0;
+}
+
+static const struct of_device_id occ_match[] = {
+ { .compatible = "ibm,p9-occ" },
+ { },
+};
+
+static struct platform_driver occ_driver = {
+ .driver = {
+ .name = "occ",
+ .of_match_table = occ_match,
+ },
+ .probe = occ_probe,
+ .remove = occ_remove,
+};
+
+static int occ_init(void)
+{
+ occ_wq = create_singlethread_workqueue("occ");
+ if (!occ_wq)
+ return -ENOMEM;
+
+ return platform_driver_register(&occ_driver);
+}
+
+static void occ_exit(void)
+{
+ destroy_workqueue(occ_wq);
+
+ platform_driver_unregister(&occ_driver);
+
+ ida_destroy(&occ_ida);
+}
+
+module_init(occ_init);
+module_exit(occ_exit);
+
+MODULE_AUTHOR("Eddie James <[email protected]>");
+MODULE_DESCRIPTION("BMC P9 OCC driver");
+MODULE_LICENSE("GPL");
--
1.8.3.1
From 1584733314506989983@xxx Wed Nov 22 02:57:40 +0000 2017
X-GM-THRID: 1584733314506989983
X-Gmail-Labels: Inbox,Category Forums,HistoricalUnread
From: "Edward A. James" <[email protected]>
Add a miscdevice for the sbefifo to allow userspace access through
standard file operations.
Signed-off-by: Edward A. James <[email protected]>
---
drivers/fsi/fsi-sbefifo.c | 355 ++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 355 insertions(+)
diff --git a/drivers/fsi/fsi-sbefifo.c b/drivers/fsi/fsi-sbefifo.c
index 6ba190a..fed3739 100644
--- a/drivers/fsi/fsi-sbefifo.c
+++ b/drivers/fsi/fsi-sbefifo.c
@@ -19,6 +19,7 @@
#include <linux/kernel.h>
#include <linux/kref.h>
#include <linux/list.h>
+#include <linux/miscdevice.h>
#include <linux/module.h>
#include <linux/poll.h>
#include <linux/sched.h>
@@ -54,6 +55,7 @@
struct sbefifo {
struct timer_list poll_timer;
struct fsi_device *fsi_dev;
+ struct miscdevice mdev;
wait_queue_head_t wait;
struct list_head xfrs;
struct kref kref;
@@ -256,6 +258,99 @@ static void sbefifo_put(struct sbefifo *sbefifo)
kref_put(&sbefifo->kref, sbefifo_free);
}
+static struct sbefifo_xfr *sbefifo_enq_xfr(struct sbefifo_client *client)
+{
+ struct sbefifo *sbefifo = client->dev;
+ struct sbefifo_xfr *xfr;
+
+ if (READ_ONCE(sbefifo->rc))
+ return ERR_PTR(sbefifo->rc);
+
+ xfr = kzalloc(sizeof(*xfr), GFP_KERNEL);
+ if (!xfr)
+ return ERR_PTR(-ENOMEM);
+
+ xfr->rbuf = &client->rbuf;
+ xfr->wbuf = &client->wbuf;
+ list_add_tail(&xfr->xfrs, &sbefifo->xfrs);
+ list_add_tail(&xfr->client, &client->xfrs);
+
+ return xfr;
+}
+
+static bool sbefifo_xfr_rsp_pending(struct sbefifo_client *client)
+{
+ struct sbefifo_xfr *xfr = list_first_entry_or_null(&client->xfrs,
+ struct sbefifo_xfr,
+ client);
+
+ if (xfr && test_bit(SBEFIFO_XFR_RESP_PENDING, &xfr->flags))
+ return true;
+
+ return false;
+}
+
+static struct sbefifo_client *sbefifo_new_client(struct sbefifo *sbefifo)
+{
+ struct sbefifo_client *client;
+
+ client = kzalloc(sizeof(*client), GFP_KERNEL);
+ if (!client)
+ return NULL;
+
+ kref_init(&client->kref);
+ client->dev = sbefifo;
+ sbefifo_buf_init(&client->rbuf);
+ sbefifo_buf_init(&client->wbuf);
+ INIT_LIST_HEAD(&client->xfrs);
+
+ sbefifo_get(sbefifo);
+
+ return client;
+}
+
+static void sbefifo_client_release(struct kref *kref)
+{
+ struct sbefifo *sbefifo;
+ struct sbefifo_client *client;
+ struct sbefifo_xfr *xfr, *tmp;
+
+ client = container_of(kref, struct sbefifo_client, kref);
+ sbefifo = client->dev;
+
+ if (!READ_ONCE(sbefifo->rc)) {
+ list_for_each_entry_safe(xfr, tmp, &client->xfrs, client) {
+ if (test_bit(SBEFIFO_XFR_COMPLETE, &xfr->flags)) {
+ list_del(&xfr->client);
+ kfree(xfr);
+ continue;
+ }
+
+ /*
+ * The client left with pending or running xfrs.
+ * Cancel them.
+ */
+ set_bit(SBEFIFO_XFR_CANCEL, &xfr->flags);
+ sbefifo_get(sbefifo);
+ if (mod_timer(&client->dev->poll_timer, jiffies))
+ sbefifo_put(sbefifo);
+ }
+ }
+
+ sbefifo_put(sbefifo);
+ kfree(client);
+}
+
+static void sbefifo_get_client(struct sbefifo_client *client)
+{
+ kref_get(&client->kref);
+}
+
+static void sbefifo_put_client(struct sbefifo_client *client)
+{
+ kref_put(&client->kref, sbefifo_client_release);
+}
+
static struct sbefifo_xfr *sbefifo_next_xfr(struct sbefifo *sbefifo)
{
struct sbefifo_xfr *xfr, *tmp;
@@ -430,6 +525,252 @@ static void sbefifo_poll_timer(unsigned long data)
spin_unlock_irqrestore(&sbefifo->lock, flags);
}
+static int sbefifo_open(struct inode *inode, struct file *file)
+{
+ struct sbefifo *sbefifo = container_of(file->private_data,
+ struct sbefifo, mdev);
+ struct sbefifo_client *client;
+ int ret;
+
+ ret = READ_ONCE(sbefifo->rc);
+ if (ret)
+ return ret;
+
+ client = sbefifo_new_client(sbefifo);
+ if (!client)
+ return -ENOMEM;
+
+ file->private_data = client;
+
+ return 0;
+}
+
+static unsigned int sbefifo_poll(struct file *file, poll_table *wait)
+{
+ struct sbefifo_client *client = file->private_data;
+ struct sbefifo *sbefifo = client->dev;
+ unsigned int mask = 0;
+
+ poll_wait(file, &sbefifo->wait, wait);
+
+ if (READ_ONCE(sbefifo->rc))
+ mask |= POLLERR;
+
+ if (sbefifo_buf_nbreadable(&client->rbuf))
+ mask |= POLLIN;
+
+ if (sbefifo_buf_nbwriteable(&client->wbuf))
+ mask |= POLLOUT;
+
+ return mask;
+}
+
+static bool sbefifo_read_ready(struct sbefifo *sbefifo,
+ struct sbefifo_client *client, size_t *n,
+ size_t *ret)
+{
+ struct sbefifo_xfr *xfr = list_first_entry_or_null(&client->xfrs,
+ struct sbefifo_xfr,
+ client);
+
+ *n = sbefifo_buf_nbreadable(&client->rbuf);
+ *ret = READ_ONCE(sbefifo->rc);
+
+ return *ret || *n ||
+ (xfr && test_bit(SBEFIFO_XFR_COMPLETE, &xfr->flags));
+}
+
+static ssize_t sbefifo_read(struct file *file, char __user *buf, size_t len,
+ loff_t *offset)
+{
+ struct sbefifo_client *client = file->private_data;
+ struct sbefifo *sbefifo = client->dev;
+ struct sbefifo_xfr *xfr;
+ size_t n;
+ ssize_t ret = 0;
+
+ if ((len >> 2) << 2 != len)
+ return -EINVAL;
+
+ if ((file->f_flags & O_NONBLOCK) && !sbefifo_xfr_rsp_pending(client))
+ return -EAGAIN;
+
+ sbefifo_get_client(client);
+ if (wait_event_interruptible(sbefifo->wait,
+ sbefifo_read_ready(sbefifo, client, &n,
+ &ret))) {
+ ret = -ERESTARTSYS;
+ goto out;
+ }
+
+ if (ret) {
+ INIT_LIST_HEAD(&client->xfrs);
+ goto out;
+ }
+
+ n = min_t(size_t, n, len);
+
+ if (copy_to_user(buf, READ_ONCE(client->rbuf.rpos), n)) {
+ ret = -EFAULT;
+ goto out;
+ }
+
+ if (sbefifo_buf_readnb(&client->rbuf, n)) {
+ xfr = list_first_entry_or_null(&client->xfrs,
+ struct sbefifo_xfr, client);
+ if (!xfr) {
+ /* should be impossible to not have an xfr here */
+ WARN_ONCE(1, "no xfr in queue");
+ ret = -EPROTO;
+ goto out;
+ }
+
+ if (!test_bit(SBEFIFO_XFR_COMPLETE, &xfr->flags)) {
+ /* Fill the read buffer back up. */
+ sbefifo_get(sbefifo);
+ if (mod_timer(&client->dev->poll_timer, jiffies))
+ sbefifo_put(sbefifo);
+ } else {
+ list_del(&xfr->client);
+ kfree(xfr);
+ wake_up_interruptible(&sbefifo->wait);
+ }
+ }
+
+ ret = n;
+
+out:
+ sbefifo_put_client(client);
+ return ret;
+}
+
+static bool sbefifo_write_ready(struct sbefifo *sbefifo,
+ struct sbefifo_xfr *xfr,
+ struct sbefifo_client *client, size_t *n)
+{
+ struct sbefifo_xfr *next = list_first_entry_or_null(&client->xfrs,
+ struct sbefifo_xfr,
+ client);
+
+ *n = sbefifo_buf_nbwriteable(&client->wbuf);
+ return READ_ONCE(sbefifo->rc) || (next == xfr && *n);
+}
+
+static ssize_t sbefifo_write(struct file *file, const char __user *buf,
+ size_t len, loff_t *offset)
+{
+ unsigned long flags;
+ struct sbefifo_client *client = file->private_data;
+ struct sbefifo *sbefifo = client->dev;
+ struct sbefifo_xfr *xfr;
+ ssize_t ret = 0;
+ size_t n;
+
+ if ((len >> 2) << 2 != len)
+ return -EINVAL;
+
+ if (!len)
+ return 0;
+
+ sbefifo_get_client(client);
+ n = sbefifo_buf_nbwriteable(&client->wbuf);
+
+ spin_lock_irqsave(&sbefifo->lock, flags);
+ xfr = sbefifo_next_xfr(sbefifo); /* next xfr to be executed */
+
+ if ((file->f_flags & O_NONBLOCK) && xfr && n < len) {
+ spin_unlock_irqrestore(&sbefifo->lock, flags);
+ ret = -EAGAIN;
+ goto out;
+ }
+
+ xfr = sbefifo_enq_xfr(client); /* this xfr queued up */
+ if (IS_ERR(xfr)) {
+ spin_unlock_irqrestore(&sbefifo->lock, flags);
+ ret = PTR_ERR(xfr);
+ goto out;
+ }
+
+ spin_unlock_irqrestore(&sbefifo->lock, flags);
+
+ /*
+ * Partial writes are not really allowed in that EOT is sent exactly
+ * once per write.
+ */
+ while (len) {
+ if (wait_event_interruptible(sbefifo->wait,
+ sbefifo_write_ready(sbefifo, xfr,
+ client,
+ &n))) {
+ set_bit(SBEFIFO_XFR_CANCEL, &xfr->flags);
+ sbefifo_get(sbefifo);
+ if (mod_timer(&sbefifo->poll_timer, jiffies))
+ sbefifo_put(sbefifo);
+
+ ret = -ERESTARTSYS;
+ goto out;
+ }
+
+ if (sbefifo->rc) {
+ INIT_LIST_HEAD(&client->xfrs);
+ ret = sbefifo->rc;
+ goto out;
+ }
+
+ n = min_t(size_t, n, len);
+
+ if (copy_from_user(READ_ONCE(client->wbuf.wpos), buf, n)) {
+ set_bit(SBEFIFO_XFR_CANCEL, &xfr->flags);
+ sbefifo_get(sbefifo);
+ if (mod_timer(&sbefifo->poll_timer, jiffies))
+ sbefifo_put(sbefifo);
+ ret = -EFAULT;
+ goto out;
+ }
+
+ buf += n;
+
+ sbefifo_buf_wrotenb(&client->wbuf, n);
+ len -= n;
+ ret += n;
+
+ /*
+ * Set this before starting timer to avoid race condition on
+ * this flag with the timer function writer.
+ */
+ if (!len)
+ set_bit(SBEFIFO_XFR_WRITE_DONE, &xfr->flags);
+
+ /* Drain the write buffer. */
+ sbefifo_get(sbefifo);
+ if (mod_timer(&client->dev->poll_timer, jiffies))
+ sbefifo_put(sbefifo);
+ }
+
+out:
+ sbefifo_put_client(client);
+ return ret;
+}
+
+static int sbefifo_release(struct inode *inode, struct file *file)
+{
+ struct sbefifo_client *client = file->private_data;
+ struct sbefifo *sbefifo = client->dev;
+
+ sbefifo_put_client(client);
+
+ return READ_ONCE(sbefifo->rc);
+}
+
+static const struct file_operations sbefifo_fops = {
+ .owner = THIS_MODULE,
+ .open = sbefifo_open,
+ .read = sbefifo_read,
+ .write = sbefifo_write,
+ .poll = sbefifo_poll,
+ .release = sbefifo_release,
+};
+
static int sbefifo_request_reset(struct sbefifo *sbefifo)
{
int ret;
@@ -504,6 +845,18 @@ static int sbefifo_probe(struct device *dev)
setup_timer(&sbefifo->poll_timer, sbefifo_poll_timer,
(unsigned long)sbefifo);
+ sbefifo->mdev.minor = MISC_DYNAMIC_MINOR;
+ sbefifo->mdev.fops = &sbefifo_fops;
+ sbefifo->mdev.name = sbefifo->name;
+ sbefifo->mdev.parent = dev;
+ ret = misc_register(&sbefifo->mdev);
+ if (ret) {
+ dev_err(dev, "failed to register miscdevice: %d\n", ret);
+ ida_simple_remove(&sbefifo_ida, sbefifo->idx);
+ sbefifo_put(sbefifo);
+ return ret;
+ }
+
dev_set_drvdata(dev, sbefifo);
return 0;
@@ -529,6 +882,8 @@ static int sbefifo_remove(struct device *dev)
wake_up_all(&sbefifo->wait);
+ misc_deregister(&sbefifo->mdev);
+
ida_simple_remove(&sbefifo_ida, sbefifo->idx);
if (del_timer_sync(&sbefifo->poll_timer))
--
1.8.3.1
From 1584073410268572603@xxx Tue Nov 14 20:08:46 +0000 2017
X-GM-THRID: 1584073410268572603
X-Gmail-Labels: Inbox,Category Forums,HistoricalUnread
From: "Edward A. James" <[email protected]>
Add an in-kernel read/write API with exported functions. This is
necessary for other drivers which want to directly interact with the
OCC. Also add a child platform device node for the associated hwmon
device.
Signed-off-by: Edward A. James <[email protected]>
---
drivers/fsi/fsi-occ.c | 138 ++++++++++++++++++++++++++++++++++++++++--------
include/linux/fsi-occ.h | 41 ++++++++++++++
2 files changed, 157 insertions(+), 22 deletions(-)
create mode 100644 include/linux/fsi-occ.h
diff --git a/drivers/fsi/fsi-occ.c b/drivers/fsi/fsi-occ.c
index 1aeb67d..1713550 100644
--- a/drivers/fsi/fsi-occ.c
+++ b/drivers/fsi/fsi-occ.c
@@ -19,7 +19,9 @@
#include <linux/miscdevice.h>
#include <linux/module.h>
#include <linux/mutex.h>
+#include <linux/fsi-occ.h>
#include <linux/of.h>
+#include <linux/of_platform.h>
#include <linux/platform_device.h>
#include <linux/sched.h>
#include <linux/slab.h>
@@ -32,8 +34,6 @@
#define OCC_CMD_DATA_BYTES 4090
#define OCC_RESP_DATA_BYTES 4089
-#define OCC_RESP_CMD_IN_PRG 0xFF
-
#define OCC_TIMEOUT_MS 1000
#define OCC_CMD_IN_PRG_WAIT_MS 50
@@ -156,23 +156,34 @@ static void occ_put_client(struct occ_client *client)
kref_put(&client->kref, occ_client_release);
}
-static int occ_open(struct inode *inode, struct file *file)
+static struct occ_client *occ_open_common(struct occ *occ, unsigned long flags)
{
- struct miscdevice *mdev = file->private_data;
- struct occ *occ = to_occ(mdev);
struct occ_client *client = kzalloc(sizeof(*client), GFP_KERNEL);
if (!client)
- return -ENOMEM;
+ return NULL;
client->occ = occ;
kref_init(&client->kref);
spin_lock_init(&client->lock);
init_waitqueue_head(&client->wait);
- if (file->f_flags & O_NONBLOCK)
+ if (flags & O_NONBLOCK)
set_bit(CLIENT_NONBLOCKING, &client->flags);
+ return client;
+}
+
+static int occ_open(struct inode *inode, struct file *file)
+{
+ struct occ_client *client;
+ struct miscdevice *mdev = file->private_data;
+ struct occ *occ = to_occ(mdev);
+
+ client = occ_open_common(occ, file->f_flags);
+ if (!client)
+ return -ENOMEM;
+
file->private_data = client;
return 0;
@@ -184,15 +195,14 @@ static inline bool occ_read_ready(struct occ_xfr *xfr, struct occ *occ)
test_bit(XFR_CANCELED, &xfr->flags) || occ->cancel;
}
-static ssize_t occ_read(struct file *file, char __user *buf, size_t len,
- loff_t *offset)
+static ssize_t occ_read_common(struct occ_client *client, char __user *ubuf,
+ char *kbuf, size_t len)
{
int rc;
unsigned long flags;
size_t bytes;
struct occ_xfr *xfr;
struct occ *occ;
- struct occ_client *client = file->private_data;
if (!client)
return -ENODEV;
@@ -247,9 +257,14 @@ static ssize_t occ_read(struct file *file, char __user *buf, size_t len,
}
bytes = min(len, xfr->resp_data_length - client->read_offset);
- if (copy_to_user(buf, &xfr->buf[client->read_offset], bytes)) {
- rc = -EFAULT;
- goto done;
+ if (ubuf) {
+ if (copy_to_user(ubuf, &xfr->buf[client->read_offset],
+ bytes)) {
+ rc = -EFAULT;
+ goto done;
+ }
+ } else {
+ memcpy(kbuf, &xfr->buf[client->read_offset], bytes);
}
client->read_offset += bytes;
@@ -266,15 +281,23 @@ static ssize_t occ_read(struct file *file, char __user *buf, size_t len,
return rc;
}
-static ssize_t occ_write(struct file *file, const char __user *buf,
- size_t len, loff_t *offset)
+static ssize_t occ_read(struct file *file, char __user *buf, size_t len,
+ loff_t *offset)
+{
+ struct occ_client *client = file->private_data;
+
+ return occ_read_common(client, buf, NULL, len);
+}
+
+static ssize_t occ_write_common(struct occ_client *client,
+ const char __user *ubuf, const char *kbuf,
+ size_t len)
{
int rc;
unsigned long flags;
unsigned int i;
u16 data_length, checksum = 0;
struct occ_xfr *xfr;
- struct occ_client *client = file->private_data;
if (!client)
return -ENODEV;
@@ -301,9 +324,13 @@ static ssize_t occ_write(struct file *file, const char __user *buf,
* bytes 1-2: data length (msb first)
* bytes 3-n: data
*/
- if (copy_from_user(&xfr->buf[1], buf, len)) {
- rc = -EFAULT;
- goto done;
+ if (ubuf) {
+ if (copy_from_user(&xfr->buf[1], ubuf, len)) {
+ rc = -EFAULT;
+ goto done;
+ }
+ } else {
+ memcpy(&xfr->buf[1], kbuf, len);
}
data_length = (xfr->buf[2] << 8) + xfr->buf[3];
@@ -334,12 +361,19 @@ static ssize_t occ_write(struct file *file, const char __user *buf,
return rc;
}
-static int occ_release(struct inode *inode, struct file *file)
+static ssize_t occ_write(struct file *file, const char __user *buf,
+ size_t len, loff_t *offset)
+{
+ struct occ_client *client = file->private_data;
+
+ return occ_write_common(client, buf, NULL, len);
+}
+
+static int occ_release_common(struct occ_client *client)
{
unsigned long flags;
struct occ *occ;
struct occ_xfr *xfr;
- struct occ_client *client = file->private_data;
if (!client)
return -ENODEV;
@@ -372,6 +406,13 @@ static int occ_release(struct inode *inode, struct file *file)
return 0;
}
+static int occ_release(struct inode *inode, struct file *file)
+{
+ struct occ_client *client = file->private_data;
+
+ return occ_release_common(client);
+}
+
static const struct file_operations occ_fops = {
.owner = THIS_MODULE,
.open = occ_open,
@@ -665,12 +706,55 @@ static void occ_worker(struct work_struct *work)
goto again;
}
+struct occ_client *occ_drv_open(struct device *dev, unsigned long flags)
+{
+ struct occ *occ = dev_get_drvdata(dev);
+
+ if (!occ)
+ return NULL;
+
+ return occ_open_common(occ, flags);
+}
+EXPORT_SYMBOL_GPL(occ_drv_open);
+
+int occ_drv_read(struct occ_client *client, char *buf, size_t len)
+{
+ return occ_read_common(client, NULL, buf, len);
+}
+EXPORT_SYMBOL_GPL(occ_drv_read);
+
+int occ_drv_write(struct occ_client *client, const char *buf, size_t len)
+{
+ return occ_write_common(client, NULL, buf, len);
+}
+EXPORT_SYMBOL_GPL(occ_drv_write);
+
+void occ_drv_release(struct occ_client *client)
+{
+ occ_release_common(client);
+}
+EXPORT_SYMBOL_GPL(occ_drv_release);
+
+static int occ_unregister_child(struct device *dev, void *data)
+{
+ struct platform_device *child = to_platform_device(dev);
+
+ of_device_unregister(child);
+ if (dev->of_node)
+ of_node_clear_flag(dev->of_node, OF_POPULATED);
+
+ return 0;
+}
+
static int occ_probe(struct platform_device *pdev)
{
- int rc;
+ int rc, child_idx = 0;
u32 reg;
struct occ *occ;
+ struct device_node *np;
+ struct platform_device *child;
struct device *dev = &pdev->dev;
+ char child_name[32];
occ = devm_kzalloc(dev, sizeof(*occ), GFP_KERNEL);
if (!occ)
@@ -712,6 +796,15 @@ static int occ_probe(struct platform_device *pdev)
return rc;
}
+ /* create platform devs for dts child nodes (hwmon, etc) */
+ for_each_available_child_of_node(dev->of_node, np) {
+ snprintf(child_name, sizeof(child_name), "occ%d-dev%d",
+ occ->idx, child_idx++);
+ child = of_platform_device_create(np, child_name, dev);
+ if (!child)
+ dev_warn(dev, "failed to create child node dev\n");
+ }
+
platform_set_drvdata(pdev, occ);
return 0;
@@ -734,6 +827,7 @@ static int occ_remove(struct platform_device *pdev)
spin_unlock_irqrestore(&occ->list_lock, flags);
misc_deregister(&occ->mdev);
+ device_for_each_child(&pdev->dev, NULL, occ_unregister_child);
cancel_work_sync(&occ->work);
diff --git a/include/linux/fsi-occ.h b/include/linux/fsi-occ.h
new file mode 100644
index 0000000..0a4a54a
--- /dev/null
+++ b/include/linux/fsi-occ.h
@@ -0,0 +1,41 @@
+/*
+ * Copyright (C) IBM Corporation 2017
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERGCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ */
+
+#ifndef LINUX_FSI_OCC_H
+#define LINUX_FSI_OCC_H
+
+struct device;
+struct occ_client;
+
+#define OCC_RESP_CMD_IN_PRG 0xFF
+#define OCC_RESP_SUCCESS 0
+#define OCC_RESP_CMD_INVAL 0x11
+#define OCC_RESP_CMD_LEN_INVAL 0x12
+#define OCC_RESP_DATA_INVAL 0x13
+#define OCC_RESP_CHKSUM_ERR 0x14
+#define OCC_RESP_INT_ERR 0x15
+#define OCC_RESP_BAD_STATE 0x16
+#define OCC_RESP_CRIT_EXCEPT 0xE0
+#define OCC_RESP_CRIT_INIT 0xE1
+#define OCC_RESP_CRIT_WATCHDOG 0xE2
+#define OCC_RESP_CRIT_OCB 0xE3
+#define OCC_RESP_CRIT_HW 0xE4
+
+extern struct occ_client *occ_drv_open(struct device *dev,
+ unsigned long flags);
+extern int occ_drv_read(struct occ_client *client, char *buf, size_t len);
+extern int occ_drv_write(struct occ_client *client, const char *buf,
+ size_t len);
+extern void occ_drv_release(struct occ_client *client);
+
+#endif /* LINUX_FSI_OCC_H */
--
1.8.3.1
From 1583495565022227104@xxx Wed Nov 08 11:04:10 +0000 2017
X-GM-THRID: 1583329603309959733
X-Gmail-Labels: Inbox,Category Forums,HistoricalUnread