Commit 39786285 authored by Dave Jiang's avatar Dave Jiang Committed by Vinod Koul

dmaengine: idxd: fix dma device lifetime

The devm managed lifetime is incompatible with 'struct device' objects that
resides in idxd context. This is one of the series that clean up the idxd
driver 'struct device' lifetime. Remove embedding of dma_device and dma_chan
in idxd since it's not the only interface that idxd will use. The freeing of
the dma_device will be managed by the ->release() function.
Reported-by: default avatarJason Gunthorpe <jgg@nvidia.com>
Fixes: bfe1d560 ("dmaengine: idxd: Init and probe for Intel data accelerators")
Signed-off-by: default avatarDave Jiang <dave.jiang@intel.com>
Reviewed-by: default avatarDan Williams <dan.j.williams@intel.com>
Link: https://lore.kernel.org/r/161852983001.2203940.14817017492384561719.stgit@djiang5-desk3.ch.intel.comSigned-off-by: default avatarVinod Koul <vkoul@kernel.org>
parent 63606522
......@@ -186,8 +186,6 @@ int idxd_wq_alloc_resources(struct idxd_wq *wq)
desc->id = i;
desc->wq = wq;
desc->cpu = -1;
dma_async_tx_descriptor_init(&desc->txd, &wq->dma_chan);
desc->txd.tx_submit = idxd_dma_tx_submit;
}
return 0;
......
......@@ -14,7 +14,10 @@
static inline struct idxd_wq *to_idxd_wq(struct dma_chan *c)
{
return container_of(c, struct idxd_wq, dma_chan);
struct idxd_dma_chan *idxd_chan;
idxd_chan = container_of(c, struct idxd_dma_chan, chan);
return idxd_chan->wq;
}
void idxd_dma_complete_txd(struct idxd_desc *desc,
......@@ -135,7 +138,7 @@ static void idxd_dma_issue_pending(struct dma_chan *dma_chan)
{
}
dma_cookie_t idxd_dma_tx_submit(struct dma_async_tx_descriptor *tx)
static dma_cookie_t idxd_dma_tx_submit(struct dma_async_tx_descriptor *tx)
{
struct dma_chan *c = tx->chan;
struct idxd_wq *wq = to_idxd_wq(c);
......@@ -156,14 +159,25 @@ dma_cookie_t idxd_dma_tx_submit(struct dma_async_tx_descriptor *tx)
static void idxd_dma_release(struct dma_device *device)
{
struct idxd_dma_dev *idxd_dma = container_of(device, struct idxd_dma_dev, dma);
kfree(idxd_dma);
}
int idxd_register_dma_device(struct idxd_device *idxd)
{
struct dma_device *dma = &idxd->dma_dev;
struct idxd_dma_dev *idxd_dma;
struct dma_device *dma;
struct device *dev = &idxd->pdev->dev;
int rc;
idxd_dma = kzalloc_node(sizeof(*idxd_dma), GFP_KERNEL, dev_to_node(dev));
if (!idxd_dma)
return -ENOMEM;
dma = &idxd_dma->dma;
INIT_LIST_HEAD(&dma->channels);
dma->dev = &idxd->pdev->dev;
dma->dev = dev;
dma_cap_set(DMA_PRIVATE, dma->cap_mask);
dma_cap_set(DMA_COMPLETION_NO_ORDER, dma->cap_mask);
......@@ -179,35 +193,72 @@ int idxd_register_dma_device(struct idxd_device *idxd)
dma->device_alloc_chan_resources = idxd_dma_alloc_chan_resources;
dma->device_free_chan_resources = idxd_dma_free_chan_resources;
return dma_async_device_register(&idxd->dma_dev);
rc = dma_async_device_register(dma);
if (rc < 0) {
kfree(idxd_dma);
return rc;
}
idxd_dma->idxd = idxd;
/*
* This pointer is protected by the refs taken by the dma_chan. It will remain valid
* as long as there are outstanding channels.
*/
idxd->idxd_dma = idxd_dma;
return 0;
}
void idxd_unregister_dma_device(struct idxd_device *idxd)
{
dma_async_device_unregister(&idxd->dma_dev);
dma_async_device_unregister(&idxd->idxd_dma->dma);
}
int idxd_register_dma_channel(struct idxd_wq *wq)
{
struct idxd_device *idxd = wq->idxd;
struct dma_device *dma = &idxd->dma_dev;
struct dma_chan *chan = &wq->dma_chan;
int rc;
struct dma_device *dma = &idxd->idxd_dma->dma;
struct device *dev = &idxd->pdev->dev;
struct idxd_dma_chan *idxd_chan;
struct dma_chan *chan;
int rc, i;
idxd_chan = kzalloc_node(sizeof(*idxd_chan), GFP_KERNEL, dev_to_node(dev));
if (!idxd_chan)
return -ENOMEM;
memset(&wq->dma_chan, 0, sizeof(struct dma_chan));
chan = &idxd_chan->chan;
chan->device = dma;
list_add_tail(&chan->device_node, &dma->channels);
for (i = 0; i < wq->num_descs; i++) {
struct idxd_desc *desc = wq->descs[i];
dma_async_tx_descriptor_init(&desc->txd, chan);
desc->txd.tx_submit = idxd_dma_tx_submit;
}
rc = dma_async_device_channel_register(dma, chan);
if (rc < 0)
if (rc < 0) {
kfree(idxd_chan);
return rc;
}
wq->idxd_chan = idxd_chan;
idxd_chan->wq = wq;
get_device(&wq->conf_dev);
return 0;
}
void idxd_unregister_dma_channel(struct idxd_wq *wq)
{
struct dma_chan *chan = &wq->dma_chan;
struct idxd_dma_chan *idxd_chan = wq->idxd_chan;
struct dma_chan *chan = &idxd_chan->chan;
struct idxd_dma_dev *idxd_dma = wq->idxd->idxd_dma;
dma_async_device_channel_unregister(&wq->idxd->dma_dev, chan);
dma_async_device_channel_unregister(&idxd_dma->dma, chan);
list_del(&chan->device_node);
kfree(wq->idxd_chan);
wq->idxd_chan = NULL;
put_device(&wq->conf_dev);
}
......@@ -14,6 +14,9 @@
extern struct kmem_cache *idxd_desc_pool;
struct idxd_device;
struct idxd_wq;
#define IDXD_REG_TIMEOUT 50
#define IDXD_DRAIN_TIMEOUT 5000
......@@ -96,6 +99,11 @@ enum idxd_complete_type {
IDXD_COMPLETE_DEV_FAIL,
};
struct idxd_dma_chan {
struct dma_chan chan;
struct idxd_wq *wq;
};
struct idxd_wq {
void __iomem *portal;
struct device conf_dev;
......@@ -125,7 +133,7 @@ struct idxd_wq {
int compls_size;
struct idxd_desc **descs;
struct sbitmap_queue sbq;
struct dma_chan dma_chan;
struct idxd_dma_chan *idxd_chan;
char name[WQ_NAME_SIZE + 1];
u64 max_xfer_bytes;
u32 max_batch_size;
......@@ -162,6 +170,11 @@ enum idxd_device_flag {
IDXD_FLAG_PASID_ENABLED,
};
struct idxd_dma_dev {
struct idxd_device *idxd;
struct dma_device dma;
};
struct idxd_device {
enum idxd_type type;
struct device conf_dev;
......@@ -210,7 +223,7 @@ struct idxd_device {
int num_wq_irqs;
struct idxd_irq_entry *irq_entries;
struct dma_device dma_dev;
struct idxd_dma_dev *idxd_dma;
struct workqueue_struct *wq;
struct work_struct work;
};
......@@ -363,7 +376,6 @@ void idxd_unregister_dma_channel(struct idxd_wq *wq);
void idxd_parse_completion_status(u8 status, enum dmaengine_tx_result *res);
void idxd_dma_complete_txd(struct idxd_desc *desc,
enum idxd_complete_type comp_type);
dma_cookie_t idxd_dma_tx_submit(struct dma_async_tx_descriptor *tx);
/* cdev */
int idxd_cdev_register(void);
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment