Re: [PATCH v6 07/23] drivers/fsi: Implement slave initialisation

From: Christopher Bostic
Date: Tue May 02 2017 - 14:59:12 EST




On 5/2/17 1:24 AM, Joel Stanley wrote:
On Tue, Apr 11, 2017 at 5:16 AM, Christopher Bostic
<cbostic@xxxxxxxxxxxxxxxxxx> wrote:
From: Jeremy Kerr <jk@xxxxxxxxxx>

Implement fsi_slave_init: if we can read a chip ID, create fsi_slave
devices and register with the driver core.

Includes changes from Chris Bostic <cbostic@xxxxxxxxxxxxxxxxxx>.

Signed-off-by: Jeremy Kerr <jk@xxxxxxxxxx>
Signed-off-by: Chris Bostic <cbostic@xxxxxxxxxxxxxxxxxx>
Signed-off-by: Joel Stanley <joel@xxxxxxxxx>
---
drivers/fsi/fsi-core.c | 66 ++++++++++++++++++++++++++++++++++++++++++++++++--
1 file changed, 64 insertions(+), 2 deletions(-)

diff --git a/drivers/fsi/fsi-core.c b/drivers/fsi/fsi-core.c
index 6e1cfdf..c705ca2 100644
--- a/drivers/fsi/fsi-core.c
+++ b/drivers/fsi/fsi-core.c
@@ -17,9 +17,12 @@
#include <linux/fsi.h>
#include <linux/idr.h>
#include <linux/module.h>
+#include <linux/slab.h>

#include "fsi-master.h"

+#define FSI_SLAVE_SIZE_23b 0x800000
+
static DEFINE_IDA(master_ida);

struct fsi_slave {
@@ -114,11 +117,70 @@ static int fsi_slave_write(struct fsi_slave *slave, uint32_t addr,
addr, val, size);
}

+static void fsi_slave_release(struct device *dev)
+{
+ struct fsi_slave *slave = to_fsi_slave(dev);
+
+ kfree(slave);
+}
+
static int fsi_slave_init(struct fsi_master *master, int link, uint8_t id)
{
- /* todo: initialise slave device, perform engine scan */
+ struct fsi_slave *slave;
+ uint32_t chip_id;
+ uint8_t crc;
+ int rc;
+
+ /* Currently, we only support single slaves on a link, and use the
+ * full 23-bit address range
+ */
+ if (id != 0)
+ return -EINVAL;
+
+ rc = fsi_master_read(master, link, id, 0, &chip_id, sizeof(chip_id));
+ if (rc) {
+ dev_warn(&master->dev, "can't read slave %02x:%02x %d\n",
+ link, id, rc);
When I boot a system with this driver loaded, I get his warning:

[ 9.740000] usbhid: USB HID core driver
[ 9.840000] fsi0: can't read slave 00:00 -5
[ 9.840000] NET: Registered protocol family 10

Two things:

There's a space in front of "fsi0".

Hi Joel,

Will correct.

This warning isn't useful at that point. The slave is not readable as
the FSI master is not present (the P9 hasn't been turned on). Can we
avoid printing the warning at boot?

Yes agreed. Will remove the warning in that case.

Thanks,
Chris

Cheers,

Joel

+ return -ENODEV;
+ }
+ chip_id = be32_to_cpu(chip_id);
+
+ crc = fsi_crc4(0, chip_id, 32);
+ if (crc) {
+ dev_warn(&master->dev, "slave %02x:%02x invalid chip id CRC!\n",
+ link, id);
+ return -EIO;
+ }
+
+ dev_info(&master->dev, "fsi: found chip %08x at %02x:%02x:%02x\n",
+ chip_id, master->idx, link, id);
+
+ /* We can communicate with a slave; create the slave device and
+ * register.
+ */
+ slave = kzalloc(sizeof(*slave), GFP_KERNEL);
+ if (!slave)
+ return -ENOMEM;
+
+ slave->master = master;
+ slave->dev.parent = &master->dev;
+ slave->dev.release = fsi_slave_release;
+ slave->link = link;
+ slave->id = id;
+ slave->size = FSI_SLAVE_SIZE_23b;
+
+ dev_set_name(&slave->dev, "slave@%02x:%02x", link, id);
+ rc = device_register(&slave->dev);
+ if (rc < 0) {
+ dev_warn(&master->dev, "failed to create slave device: %d\n",
+ rc);
+ put_device(&slave->dev);
+ return rc;
+ }
+
+ /* todo: perform engine scan */

- return -ENODEV;
+ return rc;
}

/* FSI master support */
--
1.8.2.2