2007-05-25 08:39:35

by Geert Uytterhoeven

[permalink] [raw]
Subject: [patch 5/7] ps3: Disk Storage Driver

Add a Disk Storage Driver for the PS3:
- Implemented as a block device driver with a dynamic major
- Disk names (and partitions) are of the format ps3d%c(%u)
- Uses software scatter-gather with a 64 KiB bounce buffer as the hypervisor
doesn't support scatter-gather

Signed-off-by: Geert Uytterhoeven <[email protected]>
---
arch/powerpc/platforms/ps3/Kconfig | 11 +
drivers/block/Makefile | 1
drivers/block/ps3disk.c | 402 +++++++++++++++++++++++++++++++++++++
3 files changed, 414 insertions(+)

--- a/arch/powerpc/platforms/ps3/Kconfig
+++ b/arch/powerpc/platforms/ps3/Kconfig
@@ -101,4 +101,15 @@ config PS3_STORAGE
depends on PPC_PS3
tristate

+config PS3_DISK
+ tristate "PS3 Disk Storage Driver"
+ depends on PPC_PS3 && BLOCK
+ select PS3_STORAGE
+ default y
+ help
+ Include support for the PS3 Disk Storage.
+
+ This support is required to access the PS3 hard disk.
+ In general, all users will say Y or M.
+
endmenu
--- a/drivers/block/Makefile
+++ b/drivers/block/Makefile
@@ -29,3 +29,4 @@ obj-$(CONFIG_VIODASD) += viodasd.o
obj-$(CONFIG_BLK_DEV_SX8) += sx8.o
obj-$(CONFIG_BLK_DEV_UB) += ub.o

+obj-$(CONFIG_PS3_DISK) += ps3disk.o
--- /dev/null
+++ b/drivers/block/ps3disk.c
@@ -0,0 +1,402 @@
+/*
+ * PS3 Disk Storage Driver
+ *
+ * Copyright (C) 2007 Sony Computer Entertainment Inc.
+ * Copyright 2007 Sony 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; version 2 of the License.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#include <linux/dma-mapping.h>
+#include <linux/blkdev.h>
+#include <linux/freezer.h>
+#include <linux/hdreg.h>
+#include <linux/interrupt.h>
+#include <linux/kthread.h>
+
+#include <asm/ps3stor.h>
+
+
+#define DEVICE_NAME "ps3disk"
+
+#define BOUNCE_SIZE (64*1024)
+
+// FIXME Use a fixed major assigned by LANANA?
+#define PS3DISK_MAJOR 0
+
+#define PS3DISK_MAX_DISKS 16
+#define PS3DISK_MINORS 16
+
+#define KERNEL_SECTOR_SIZE 512
+
+
+#define PS3DISK_NAME "ps3d%c"
+
+#define LV1_STORAGE_ATA_HDDOUT (0x23)
+
+
+struct ps3disk_private {
+ spinlock_t lock;
+ struct task_struct *thread;
+ struct request_queue *queue;
+ struct gendisk *gendisk;
+ unsigned int blocking_factor;
+};
+#define ps3disk_priv(dev) ((dev)->sbd.core.driver_data)
+
+static int ps3disk_major = PS3DISK_MAJOR;
+
+static int ps3disk_open(struct inode *inode, struct file *file)
+{
+ struct ps3_storage_device *dev = inode->i_bdev->bd_disk->private_data;
+
+ file->private_data = dev;
+ return 0;
+}
+
+
+
+static struct block_device_operations ps3disk_fops = {
+ .owner = THIS_MODULE,
+ .open = ps3disk_open,
+};
+
+static void ps3disk_scatter_gather(struct ps3_storage_device *dev,
+ struct request *req, int gather)
+{
+ unsigned int sectors = 0, offset = 0;
+ struct bio *bio;
+ sector_t sector;
+ struct bio_vec *bvec;
+ unsigned int i = 0, j;
+ size_t size;
+ void *buf;
+
+ rq_for_each_bio(bio, req) {
+ sector = bio->bi_sector;
+ dev_dbg(&dev->sbd.core,
+ "%s:%u: bio %u: %u segs %u sectors from %lu\n",
+ __func__, __LINE__, i, bio_segments(bio),
+ bio_sectors(bio), sector);
+ bio_for_each_segment(bvec, bio, j) {
+ size = bio_cur_sectors(bio)*KERNEL_SECTOR_SIZE;
+ buf = __bio_kmap_atomic(bio, j, KM_USER0);
+ if (gather)
+ memcpy(dev->bounce_buf+offset, buf, size);
+ else
+ memcpy(buf, dev->bounce_buf+offset, size);
+ offset += size;
+ __bio_kunmap_atomic(bio, KM_USER0);
+ }
+ sectors += bio_sectors(bio);
+ i++;
+ }
+}
+
+static void ps3disk_handle_request_sg(struct ps3_storage_device *dev,
+ struct request *req)
+{
+ struct ps3disk_private *priv = ps3disk_priv(dev);
+ int uptodate = 1;
+ int write = rq_data_dir(req);
+ const char *op = write ? "write" : "read";
+ u64 res;
+
+#ifdef DEBUG
+ unsigned int n = 0;
+ struct bio *bio;
+ rq_for_each_bio(bio, req)
+ n++;
+ dev_dbg(&dev->sbd.core,
+ "%s:%u: %s req has %u bios for %lu sectors %lu hard sectors\n",
+ __func__, __LINE__, op, n, req->nr_sectors,
+ req->hard_nr_sectors);
+#endif
+
+ if (write)
+ ps3disk_scatter_gather(dev, req, 1);
+
+ res = ps3stor_read_write_sectors(dev, dev->bounce_lpar,
+ req->sector*priv->blocking_factor,
+ req->nr_sectors*priv->blocking_factor,
+ write);
+ if (res) {
+ dev_err(&dev->sbd.core, "%s:%u: %s failed 0x%lx\n", __func__,
+ __LINE__, op, res);
+ uptodate = 0;
+ } else if (!write)
+ ps3disk_scatter_gather(dev, req, 0);
+
+ spin_lock_irq(&priv->lock);
+ if (!end_that_request_first(req, uptodate, req->nr_sectors)) {
+ blkdev_dequeue_request(req);
+ end_that_request_last(req, uptodate);
+ }
+ spin_unlock_irq(&priv->lock);
+}
+
+static int ps3disk_thread(void *data)
+{
+ struct ps3_storage_device *dev = data;
+ struct ps3disk_private *priv = ps3disk_priv(dev);
+ request_queue_t *q = priv->queue;
+ struct request *req;
+
+ dev_dbg(&dev->sbd.core, "%s thread init\n", __func__);
+
+ current->flags |= PF_NOFREEZE;
+
+ while (!kthread_should_stop()) {
+ spin_lock_irq(&priv->lock);
+ set_current_state(TASK_INTERRUPTIBLE);
+ req = elv_next_request(q);
+ if (!req) {
+ spin_unlock_irq(&priv->lock);
+ schedule();
+ continue;
+ }
+ if (!blk_fs_request(req)) {
+ blk_dump_rq_flags(req, DEVICE_NAME " bad request");
+ end_request(req, 0);
+ spin_unlock_irq(&priv->lock);
+ continue;
+ }
+ spin_unlock_irq(&priv->lock);
+ ps3disk_handle_request_sg(dev, req);
+ }
+
+ dev_dbg(&dev->sbd.core, "%s thread exit\n", __func__);
+ return 0;
+}
+
+static int ps3disk_sync_cache(struct ps3_storage_device *dev)
+{
+ int res;
+
+ dev_dbg(&dev->sbd.core, "%s:%u: sync cache\n", __func__, __LINE__);
+
+ res = ps3stor_send_command(dev, LV1_STORAGE_ATA_HDDOUT, 0, 0, 0, 0);
+ if (res) {
+ dev_err(&dev->sbd.core, "%s:%u: sync cache failed 0x%lx\n",
+ __func__, __LINE__, dev->lv1_status);
+ return -EIO;
+ }
+ return 0;
+}
+
+static int ps3disk_issue_flush(request_queue_t *q, struct gendisk *gendisk,
+ sector_t *sector)
+{
+ struct ps3_storage_device *dev = q->queuedata;
+ return ps3disk_sync_cache(dev);
+}
+
+static void ps3disk_prepare_flush(request_queue_t *q, struct request *req)
+{
+ // FIXME Is this the correct thing to do?
+ struct ps3_storage_device *dev = q->queuedata;
+ ps3disk_sync_cache(dev);
+ memset(req->cmd, 0, sizeof(req->cmd));
+ req->cmd_type = REQ_TYPE_FLUSH;
+}
+
+static void ps3disk_request(request_queue_t *q)
+{
+ struct ps3_storage_device *dev = q->queuedata;
+ struct ps3disk_private *priv = ps3disk_priv(dev);
+ wake_up_process(priv->thread);
+}
+
+static unsigned long ps3disk_mask;
+
+static int __devinit ps3disk_probe(struct ps3_system_bus_device *_dev)
+{
+ struct ps3_storage_device *dev = to_ps3_storage_device(&_dev->core);
+ struct ps3disk_private *priv;
+ int error;
+ unsigned int devidx;
+ struct request_queue *queue;
+ struct gendisk *gendisk;
+ struct task_struct *task;
+
+ if (dev->blk_size < KERNEL_SECTOR_SIZE) {
+ dev_err(&dev->sbd.core,
+ "%s:%u: cannot handle block size %lu\n", __func__,
+ __LINE__, dev->blk_size);
+ return -EINVAL;
+ }
+
+ BUILD_BUG_ON(PS3DISK_MAX_DISKS > BITS_PER_LONG);
+ devidx = find_first_zero_bit(&ps3disk_mask, PS3DISK_MAX_DISKS);
+ if (devidx >= PS3DISK_MAX_DISKS) {
+ dev_err(&dev->sbd.core, "%s:%u: Too many disks\n", __func__,
+ __LINE__);
+ return -ENOSPC;
+ }
+ __set_bit(devidx, &ps3disk_mask);
+
+ priv = kzalloc(sizeof(*priv), GFP_KERNEL);
+ if (!priv) {
+ error = -ENOMEM;
+ goto fail;
+ }
+
+ ps3disk_priv(dev) = priv;
+ spin_lock_init(&priv->lock);
+
+ dev->bounce_size = BOUNCE_SIZE;
+ dev->bounce_buf = kmalloc(BOUNCE_SIZE, GFP_DMA);
+ if (!dev->bounce_buf) {
+ error = -ENOMEM;
+ goto fail_free_priv;
+ }
+
+ error = ps3stor_setup(dev, DEVICE_NAME);
+ if (error)
+ goto fail_free_bounce;
+
+ queue = blk_init_queue(ps3disk_request, &priv->lock);
+ if (!queue) {
+ dev_err(&dev->sbd.core, "%s:%u: blk_init_queue failed\n",
+ __func__, __LINE__);
+ error = -ENOMEM;
+ goto fail_teardown;
+ }
+
+ priv->queue = queue;
+ queue->queuedata = dev;
+
+ blk_queue_bounce_limit(queue, BLK_BOUNCE_HIGH);
+
+ blk_queue_max_sectors(queue, dev->bounce_size/KERNEL_SECTOR_SIZE);
+ blk_queue_segment_boundary(queue, -1UL);
+ blk_queue_dma_alignment(queue, dev->blk_size-1);
+ blk_queue_hardsect_size(queue, dev->blk_size);
+
+ blk_queue_issue_flush_fn(queue, ps3disk_issue_flush);
+ blk_queue_ordered(queue, QUEUE_ORDERED_DRAIN_FLUSH,
+ ps3disk_prepare_flush);
+
+ blk_queue_max_phys_segments(queue, -1);
+ blk_queue_max_hw_segments(queue, -1);
+ blk_queue_max_segment_size(queue, dev->bounce_size);
+
+ gendisk = alloc_disk(PS3DISK_MINORS);
+ if (!gendisk) {
+ dev_err(&dev->sbd.core, "%s:%u: alloc_disk failed\n", __func__,
+ __LINE__);
+ error = -ENOMEM;
+ goto fail_cleanup_queue;
+ }
+
+ priv->gendisk = gendisk;
+ gendisk->major = ps3disk_major;
+ gendisk->first_minor = devidx * PS3DISK_MINORS;
+ gendisk->fops = &ps3disk_fops;
+ gendisk->queue = queue;
+ gendisk->private_data = dev;
+ snprintf(gendisk->disk_name, sizeof(gendisk->disk_name), PS3DISK_NAME,
+ devidx+'a');
+ priv->blocking_factor = dev->blk_size/KERNEL_SECTOR_SIZE;
+ set_capacity(gendisk,
+ dev->regions[dev->region_idx].size*priv->blocking_factor);
+
+ task = kthread_run(ps3disk_thread, dev, DEVICE_NAME);
+ if (IS_ERR(task)) {
+ error = PTR_ERR(task);
+ goto fail_free_disk;
+ }
+ priv->thread = task;
+
+ add_disk(gendisk);
+ return 0;
+
+fail_free_disk:
+ put_disk(priv->gendisk);
+fail_cleanup_queue:
+ blk_cleanup_queue(queue);
+fail_teardown:
+ ps3stor_teardown(dev);
+fail_free_bounce:
+ kfree(dev->bounce_buf);
+fail_free_priv:
+ kfree(priv);
+fail:
+ __clear_bit(devidx, &ps3disk_mask);
+ return error;
+}
+
+static int ps3disk_remove(struct ps3_system_bus_device *_dev)
+{
+ struct ps3_storage_device *dev = to_ps3_storage_device(&_dev->core);
+ struct ps3disk_private *priv = ps3disk_priv(dev);
+
+ kthread_stop(priv->thread);
+ __clear_bit(priv->gendisk->first_minor / PS3DISK_MINORS,
+ &ps3disk_mask);
+ del_gendisk(priv->gendisk);
+ put_disk(priv->gendisk);
+ blk_cleanup_queue(priv->queue);
+ dev_notice(&dev->sbd.core, "Synchronizing disk cache\n");
+ ps3disk_sync_cache(dev);
+ ps3stor_teardown(dev);
+ kfree(dev->bounce_buf);
+ kfree(priv);
+ return 0;
+}
+
+
+static struct ps3_system_bus_driver ps3disk = {
+ .match_id = PS3_MATCH_ID_STOR_DISK,
+ .core.name = DEVICE_NAME,
+ .core.owner = THIS_MODULE,
+ .probe = ps3disk_probe,
+ .remove = ps3disk_remove,
+ .shutdown = ps3disk_remove,
+};
+
+
+static int __init ps3disk_init(void)
+{
+ int error;
+
+ error = register_blkdev(ps3disk_major, DEVICE_NAME);
+ if (error <= 0) {
+ printk(KERN_ERR "%s:%u: register_blkdev failed %d\n", __func__,
+ __LINE__, error);
+ return error;
+ }
+ if (!ps3disk_major)
+ ps3disk_major = error;
+
+ pr_info("%s:%u: registered block device major %d\n", __func__,
+ __LINE__, ps3disk_major);
+
+ return ps3_system_bus_driver_register(&ps3disk, PS3_IOBUS_SB);
+}
+
+static void __exit ps3disk_exit(void)
+{
+ unregister_blkdev(ps3disk_major, DEVICE_NAME);
+
+ ps3_system_bus_driver_unregister(&ps3disk);
+}
+
+module_init(ps3disk_init);
+module_exit(ps3disk_exit);
+
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("PS3 Disk Storage Driver");
+MODULE_AUTHOR("Sony Corporation");
+

--
Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- Sony Network and Software Technology Center Europe (NSCE)
[email protected] ------- The Corporate Village, Da Vincilaan 7-D1
Voice +32-2-7008453 Fax +32-2-7008622 ---------------- B-1935 Zaventem, Belgium


2007-05-25 11:45:39

by Olaf Hering

[permalink] [raw]
Subject: Re: [patch 5/7] ps3: Disk Storage Driver

On Fri, May 25, [email protected] wrote:

> Add a Disk Storage Driver for the PS3:

There is no device symlink in /sys/block/ps3da/

2007-05-25 16:26:51

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [patch 5/7] ps3: Disk Storage Driver

On Friday 25 May 2007, [email protected] wrote:
> +static void ps3disk_scatter_gather(struct ps3_storage_device *dev,
> + struct request *req, int gather)
> +{
> + unsigned int sectors = 0, offset = 0;
> + struct bio *bio;
> + sector_t sector;
> + struct bio_vec *bvec;
> + unsigned int i = 0, j;
> + size_t size;
> + void *buf;
> +
> + rq_for_each_bio(bio, req) {
> + sector = bio->bi_sector;
> + dev_dbg(&dev->sbd.core,
> + "%s:%u: bio %u: %u segs %u sectors from %lu\n",
> + __func__, __LINE__, i, bio_segments(bio),
> + bio_sectors(bio), sector);
> + bio_for_each_segment(bvec, bio, j) {
> + size = bio_cur_sectors(bio)*KERNEL_SECTOR_SIZE;
> + buf = __bio_kmap_atomic(bio, j, KM_USER0);
> + if (gather)
> + memcpy(dev->bounce_buf+offset, buf, size);
> + else
> + memcpy(buf, dev->bounce_buf+offset, size);
> + offset += size;
> + __bio_kunmap_atomic(bio, KM_USER0);
> + }
> + sectors += bio_sectors(bio);
> + i++;
> + }
> +}

So the hypervison uses guest-real addresses here? I would have expected
it to use the kernel page tables, which lets you use vmap() to do
scatter-gather.

> +static int ps3disk_thread(void *data)
> +{
> + struct ps3_storage_device *dev = data;
> + struct ps3disk_private *priv = ps3disk_priv(dev);
> + request_queue_t *q = priv->queue;
> + struct request *req;
> +
> + dev_dbg(&dev->sbd.core, "%s thread init\n", __func__);
> +
> + current->flags |= PF_NOFREEZE;
> +
> + while (!kthread_should_stop()) {
> + spin_lock_irq(&priv->lock);
> + set_current_state(TASK_INTERRUPTIBLE);
> + req = elv_next_request(q);
> + if (!req) {
> + spin_unlock_irq(&priv->lock);
> + schedule();
> + continue;
> + }
> + if (!blk_fs_request(req)) {
> + blk_dump_rq_flags(req, DEVICE_NAME " bad request");
> + end_request(req, 0);
> + spin_unlock_irq(&priv->lock);
> + continue;
> + }
> + spin_unlock_irq(&priv->lock);
> + ps3disk_handle_request_sg(dev, req);
> + }
> +
> + dev_dbg(&dev->sbd.core, "%s thread exit\n", __func__);
> + return 0;
> +}

I don't really understand what the kthread is needed for. You probably
thought about multiple options and ended up with this, but having
a comment in front of it might be helpful.

Arnd <><

2007-05-25 19:40:53

by Geert Uytterhoeven

[permalink] [raw]
Subject: Re: [patch 5/7] ps3: Disk Storage Driver

On Fri, 25 May 2007, Arnd Bergmann wrote:
> On Friday 25 May 2007, [email protected] wrote:
> > +static void ps3disk_scatter_gather(struct ps3_storage_device *dev,
> > + struct request *req, int gather)
> > +{
> > + unsigned int sectors = 0, offset = 0;
> > + struct bio *bio;
> > + sector_t sector;
> > + struct bio_vec *bvec;
> > + unsigned int i = 0, j;
> > + size_t size;
> > + void *buf;
> > +
> > + rq_for_each_bio(bio, req) {
> > + sector = bio->bi_sector;
> > + dev_dbg(&dev->sbd.core,
> > + "%s:%u: bio %u: %u segs %u sectors from %lu\n",
> > + __func__, __LINE__, i, bio_segments(bio),
> > + bio_sectors(bio), sector);
> > + bio_for_each_segment(bvec, bio, j) {
> > + size = bio_cur_sectors(bio)*KERNEL_SECTOR_SIZE;
> > + buf = __bio_kmap_atomic(bio, j, KM_USER0);
> > + if (gather)
> > + memcpy(dev->bounce_buf+offset, buf, size);
> > + else
> > + memcpy(buf, dev->bounce_buf+offset, size);
> > + offset += size;
> > + __bio_kunmap_atomic(bio, KM_USER0);
> > + }
> > + sectors += bio_sectors(bio);
> > + i++;
> > + }
> > +}
>
> So the hypervison uses guest-real addresses here? I would have expected
> it to use the kernel page tables, which lets you use vmap() to do
> scatter-gather.

Yes, it uses logical partion addresses, so we cannot create a virtually
contiguous mapping.

> > +static int ps3disk_thread(void *data)
> > +{
> > + struct ps3_storage_device *dev = data;
> > + struct ps3disk_private *priv = ps3disk_priv(dev);
> > + request_queue_t *q = priv->queue;
> > + struct request *req;
> > +
> > + dev_dbg(&dev->sbd.core, "%s thread init\n", __func__);
> > +
> > + current->flags |= PF_NOFREEZE;
> > +
> > + while (!kthread_should_stop()) {
> > + spin_lock_irq(&priv->lock);
> > + set_current_state(TASK_INTERRUPTIBLE);
> > + req = elv_next_request(q);
> > + if (!req) {
> > + spin_unlock_irq(&priv->lock);
> > + schedule();
> > + continue;
> > + }
> > + if (!blk_fs_request(req)) {
> > + blk_dump_rq_flags(req, DEVICE_NAME " bad request");
> > + end_request(req, 0);
> > + spin_unlock_irq(&priv->lock);
> > + continue;
> > + }
> > + spin_unlock_irq(&priv->lock);
> > + ps3disk_handle_request_sg(dev, req);
> > + }
> > +
> > + dev_dbg(&dev->sbd.core, "%s thread exit\n", __func__);
> > + return 0;
> > +}
>
> I don't really understand what the kthread is needed for. You probably
> thought about multiple options and ended up with this, but having
> a comment in front of it might be helpful.

I used a kthread because the request function of a block device driver must be
non-blocking, and ps3stor_read_write_sectors() calls wait_for_completion().

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- Sony Network and Software Technology Center Europe (NSCE)
[email protected] ------- The Corporate Village, Da Vincilaan 7-D1
Voice +32-2-7008453 Fax +32-2-7008622 ---------------- B-1935 Zaventem, Belgium

2007-05-25 19:43:29

by Geert Uytterhoeven

[permalink] [raw]
Subject: Re: [patch 5/7] ps3: Disk Storage Driver

On Fri, 25 May 2007, Olaf Hering wrote:
> On Fri, May 25, [email protected] wrote:
> > Add a Disk Storage Driver for the PS3:
>
> There is no device symlink in /sys/block/ps3da/

Interesting... Do you know how to create it?

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- Sony Network and Software Technology Center Europe (NSCE)
[email protected] ------- The Corporate Village, Da Vincilaan 7-D1
Voice +32-2-7008453 Fax +32-2-7008622 ---------------- B-1935 Zaventem, Belgium

2007-05-25 20:44:00

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [patch 5/7] ps3: Disk Storage Driver

On Friday 25 May 2007, Geert Uytterhoeven wrote:
>
> > I don't really understand what the kthread is needed for. You probably
> > thought about multiple options and ended up with this, but having
> > a comment in front of it might be helpful.
>
> I used a kthread because the request function of a block device driver must be
> non-blocking, and ps3stor_read_write_sectors() calls wait_for_completion().

Ok, but why does it call wait_for_completion() then?
I thought you could end_that_request_* from the interrupt handler instead.

Arnd <><

2007-05-25 20:47:33

by Olaf Hering

[permalink] [raw]
Subject: Re: [patch 5/7] ps3: Disk Storage Driver

On Fri, May 25, Geert Uytterhoeven wrote:

> On Fri, 25 May 2007, Olaf Hering wrote:
> > On Fri, May 25, [email protected] wrote:
> > > Add a Disk Storage Driver for the PS3:
> >
> > There is no device symlink in /sys/block/ps3da/
>
> Interesting... Do you know how to create it?

No, that was not obvious.
But for gelic it goes like this:

+++ b/drivers/net/gelic_net.c
@@ -1364,6 +1364,7 @@ static int gelic_net_setup_netdev(struct
u64 v1, v2;

SET_MODULE_OWNER(netdev);
+ SET_NETDEV_DEV(netdev, &card->dev->core);
spin_lock_init(&card->tx_dma_lock);

card->rx_csum = GELIC_NET_RX_CSUM_DEFAULT;

2007-05-25 21:23:14

by Geert Uytterhoeven

[permalink] [raw]
Subject: Re: [patch 5/7] ps3: Disk Storage Driver

On Fri, 25 May 2007, Arnd Bergmann wrote:
> On Friday 25 May 2007, Geert Uytterhoeven wrote:
> > > I don't really understand what the kthread is needed for. You probably
> > > thought about multiple options and ended up with this, but having
> > > a comment in front of it might be helpful.
> >
> > I used a kthread because the request function of a block device driver must be
> > non-blocking, and ps3stor_read_write_sectors() calls wait_for_completion().
>
> Ok, but why does it call wait_for_completion() then?
> I thought you could end_that_request_* from the interrupt handler instead.

Actually I tried that first, but I ran into other problems, like my request
handler being called continuously and requests gotten stuck. But maybe it was
just a locking bug on my side.

I can retry, but a disadvantage will be that there will be less code shared
with ps3flash and ps3rom.

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- Sony Network and Software Technology Center Europe (NSCE)
[email protected] ------- The Corporate Village, Da Vincilaan 7-D1
Voice +32-2-7008453 Fax +32-2-7008622 ---------------- B-1935 Zaventem, Belgium

2007-05-25 22:45:27

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [patch 5/7] ps3: Disk Storage Driver

On Friday 25 May 2007, Geert Uytterhoeven wrote:
> On Fri, 25 May 2007, Arnd Bergmann wrote:
>
> > Ok, but why does it call wait_for_completion() then?
> > I thought you could end_that_request_* from the interrupt handler instead.
>
> Actually I tried that first, but I ran into other problems, like my request
> handler being called continuously and requests gotten stuck. But maybe it was
> just a locking bug on my side.
>
> I can retry, but a disadvantage will be that there will be less code shared
> with ps3flash and ps3rom.

Not sure how much difference it will make performance-wise, but the context
switch for each bio adds some extra cost at least. Changing it
means you no longer share the ps3stor_read_write_sectors, but can at
the same time simplify the disk driver, so that won't hurt in total.

I don't care much, but I think it's worth trying.

Arnd <><

2007-05-25 22:49:34

by Benjamin Herrenschmidt

[permalink] [raw]
Subject: Re: [patch 5/7] ps3: Disk Storage Driver


> So the hypervison uses guest-real addresses here? I would have expected
> it to use the kernel page tables, which lets you use vmap() to do
> scatter-gather.

Ugh ? Maybe s390 can do that but no other hypervisor that I know
about :-) It would be nice, sure, but heh.

> I don't really understand what the kthread is needed for. You probably
> thought about multiple options and ended up with this, but having
> a comment in front of it might be helpful.

Yeah, me neither... the driver looks very very very unefficient to me. I
though the kthread was useful for hotplug detection becasue the
hypervisor don't signal us, but from the patch, it looks like it's also
used for actual request processing which is very yucky.

Ben.



2007-05-25 22:53:46

by Benjamin Herrenschmidt

[permalink] [raw]
Subject: Re: [patch 5/7] ps3: Disk Storage Driver

On Fri, 2007-05-25 at 21:40 +0200, Geert Uytterhoeven wrote:
>
> I used a kthread because the request function of a block device driver
> must be
> non-blocking, and ps3stor_read_write_sectors() calls
> wait_for_completion().

Which as I said before looks terribly sad... Why the heck would it have
to do that ?

Ben.