On Mon, Oct 13, 2025 at 12:46:18PM +0200, Michal Wajdeczko wrote:
>
>
> On 10/11/2025 9:38 PM, Michał Winiarski wrote:
> > Add debugfs handlers for migration state and handle bitstream
> > .read()/.write() to convert from bitstream to/from migration data
> > packets.
> > As descriptor/trailer are handled at this layer - add handling for both
> > save and restore side.
> >
> > Signed-off-by: Michał Winiarski <[email protected]>
> > ---
> > drivers/gpu/drm/xe/xe_gt_sriov_pf_control.c | 18 +
> > drivers/gpu/drm/xe/xe_gt_sriov_pf_control.h | 1 +
> > drivers/gpu/drm/xe/xe_sriov_pf.c | 1 +
> > drivers/gpu/drm/xe/xe_sriov_pf_control.c | 5 +
> > drivers/gpu/drm/xe/xe_sriov_pf_debugfs.c | 45 +++
> > drivers/gpu/drm/xe/xe_sriov_pf_migration.c | 56 +++
> > .../gpu/drm/xe/xe_sriov_pf_migration_data.c | 353 ++++++++++++++++++
> > .../gpu/drm/xe/xe_sriov_pf_migration_data.h | 5 +
> > .../gpu/drm/xe/xe_sriov_pf_migration_types.h | 9 +
> > 9 files changed, 493 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_pf_control.c
> > b/drivers/gpu/drm/xe/xe_gt_sriov_pf_control.c
> > index 04a4e92133c2e..092d3d710bca1 100644
> > --- a/drivers/gpu/drm/xe/xe_gt_sriov_pf_control.c
> > +++ b/drivers/gpu/drm/xe/xe_gt_sriov_pf_control.c
> > @@ -814,6 +814,23 @@ bool xe_gt_sriov_pf_control_check_vf_data_wip(struct
> > xe_gt *gt, unsigned int vfi
> > return pf_check_vf_state(gt, vfid,
> > XE_GT_SRIOV_STATE_MIGRATION_DATA_WIP);
> > }
> >
> > +/**
> > + * xe_gt_sriov_pf_control_vf_data_eof() - indicate the end of SR-IOV VF
> > migration data production
> > + * @gt: the &struct xe_gt
> > + * @vfid: the VF identifier
> > + *
> > + * This function is for PF only.
> > + */
> > +void xe_gt_sriov_pf_control_vf_data_eof(struct xe_gt *gt, unsigned int
> > vfid)
> > +{
> > + struct wait_queue_head *wq =
> > xe_sriov_pf_migration_waitqueue(gt_to_xe(gt), vfid);
> > +
> > + if (!pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_MIGRATION_DATA_WIP))
> > + pf_enter_vf_state_machine_bug(gt, vfid);
> > +
> > + wake_up_all(wq);
> > +}
> > +
> > static void pf_exit_vf_save_wip(struct xe_gt *gt, unsigned int vfid)
> > {
> > pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_SAVE_WIP);
> > @@ -840,6 +857,7 @@ static bool pf_handle_vf_save_wip(struct xe_gt *gt,
> > unsigned int vfid)
> > if (!pf_check_vf_state(gt, vfid, XE_GT_SRIOV_STATE_SAVE_WIP))
> > return false;
> >
> > + xe_gt_sriov_pf_control_vf_data_eof(gt, vfid);
>
> above call can lead to state_machine_bug but here you just continue as
> nothing happen and moving to SAVED state
>
> maybe that logic of that function should moved to a helper that at least
> returns bool so you can make the right decision?
Won't be relevant after applying changes from previous patches.
>
> > pf_exit_vf_save_wip(gt, vfid);
> > pf_enter_vf_saved(gt, vfid);
> >
> > diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_pf_control.h
> > b/drivers/gpu/drm/xe/xe_gt_sriov_pf_control.h
> > index 2e121e8132dcf..caf20dd063b1b 100644
> > --- a/drivers/gpu/drm/xe/xe_gt_sriov_pf_control.h
> > +++ b/drivers/gpu/drm/xe/xe_gt_sriov_pf_control.h
> > @@ -15,6 +15,7 @@ int xe_gt_sriov_pf_control_init(struct xe_gt *gt);
> > void xe_gt_sriov_pf_control_restart(struct xe_gt *gt);
> >
> > bool xe_gt_sriov_pf_control_check_vf_data_wip(struct xe_gt *gt, unsigned
> > int vfid);
> > +void xe_gt_sriov_pf_control_vf_data_eof(struct xe_gt *gt, unsigned int
> > vfid);
> >
> > int xe_gt_sriov_pf_control_pause_vf(struct xe_gt *gt, unsigned int vfid);
> > int xe_gt_sriov_pf_control_resume_vf(struct xe_gt *gt, unsigned int vfid);
> > diff --git a/drivers/gpu/drm/xe/xe_sriov_pf.c
> > b/drivers/gpu/drm/xe/xe_sriov_pf.c
> > index 95743c7af8050..5d115627f3f2f 100644
> > --- a/drivers/gpu/drm/xe/xe_sriov_pf.c
> > +++ b/drivers/gpu/drm/xe/xe_sriov_pf.c
> > @@ -16,6 +16,7 @@
> > #include "xe_sriov_pf.h"
> > #include "xe_sriov_pf_helpers.h"
> > #include "xe_sriov_pf_migration.h"
> > +#include "xe_sriov_pf_migration_data.h"
> > #include "xe_sriov_pf_service.h"
> > #include "xe_sriov_printk.h"
> >
> > diff --git a/drivers/gpu/drm/xe/xe_sriov_pf_control.c
> > b/drivers/gpu/drm/xe/xe_sriov_pf_control.c
> > index e64c7b56172c6..10e1f18aa8b11 100644
> > --- a/drivers/gpu/drm/xe/xe_sriov_pf_control.c
> > +++ b/drivers/gpu/drm/xe/xe_sriov_pf_control.c
> > @@ -6,6 +6,7 @@
> > #include "xe_device.h"
> > #include "xe_gt_sriov_pf_control.h"
> > #include "xe_sriov_pf_control.h"
> > +#include "xe_sriov_pf_migration_data.h"
> > #include "xe_sriov_printk.h"
> >
> > /**
> > @@ -165,6 +166,10 @@ int xe_sriov_pf_control_save_vf(struct xe_device *xe,
> > unsigned int vfid)
> > unsigned int id;
> > int ret;
> >
> > + ret = xe_sriov_pf_migration_data_save_init(xe, vfid);
> > + if (ret)
> > + return ret;
> > +
> > for_each_gt(gt, xe, id) {
> > ret = xe_gt_sriov_pf_control_save_vf(gt, vfid);
> > if (ret)
> > diff --git a/drivers/gpu/drm/xe/xe_sriov_pf_debugfs.c
> > b/drivers/gpu/drm/xe/xe_sriov_pf_debugfs.c
> > index 74eeabef91c57..ce780719760a6 100644
> > --- a/drivers/gpu/drm/xe/xe_sriov_pf_debugfs.c
> > +++ b/drivers/gpu/drm/xe/xe_sriov_pf_debugfs.c
> > @@ -13,6 +13,7 @@
> > #include "xe_sriov_pf_control.h"
> > #include "xe_sriov_pf_debugfs.h"
> > #include "xe_sriov_pf_helpers.h"
> > +#include "xe_sriov_pf_migration_data.h"
> > #include "xe_sriov_pf_service.h"
> > #include "xe_sriov_printk.h"
> > #include "xe_tile_sriov_pf_debugfs.h"
> > @@ -71,6 +72,7 @@ static void pf_populate_pf(struct xe_device *xe, struct
> > dentry *pfdent)
> > * /sys/kernel/debug/dri/BDF/
> > * ├── sriov
> > * │ ├── vf1
> > + * │ │ ├── migration_data
> > * │ │ ├── pause
> > * │ │ ├── reset
> > * │ │ ├── resume
> > @@ -159,6 +161,48 @@ DEFINE_VF_CONTROL_ATTRIBUTE(reset_vf);
> > DEFINE_VF_RW_CONTROL_ATTRIBUTE(save_vf);
> > DEFINE_VF_RW_CONTROL_ATTRIBUTE(restore_vf);
> >
> > +static ssize_t data_write(struct file *file, const char __user *buf,
> > size_t count, loff_t *pos)
> > +{
> > + struct dentry *dent = file_dentry(file);
> > + struct dentry *vfdentry = dent->d_parent;
> > + struct dentry *migration_dentry = vfdentry->d_parent;
> > + unsigned int vfid = (uintptr_t)vfdentry->d_inode->i_private;
> > + struct xe_device *xe = migration_dentry->d_inode->i_private;
>
> we have extract_xe() / extract_vfid() helpers for that
Ok.
>
> > +
> > + xe_assert(xe, vfid);
> > + xe_sriov_pf_assert_vfid(xe, vfid);
> > +
> > + if (*pos)
> > + return -ESPIPE;
> > +
> > + return xe_sriov_pf_migration_data_write(xe, vfid, buf, count);
> > +}
> > +
> > +static ssize_t data_read(struct file *file, char __user *buf, size_t
> > count, loff_t *ppos)
> > +{
> > + struct dentry *dent = file_dentry(file);
> > + struct dentry *vfdentry = dent->d_parent;
> > + struct dentry *migration_dentry = vfdentry->d_parent;
> > + unsigned int vfid = (uintptr_t)vfdentry->d_inode->i_private;
> > + struct xe_device *xe = migration_dentry->d_inode->i_private;
> > +
> > + xe_assert(xe, vfid);
> > + xe_sriov_pf_assert_vfid(xe, vfid);
> > +
> > + if (*ppos)
> > + return -ESPIPE;
> > +
> > + return xe_sriov_pf_migration_data_read(xe, vfid, buf, count);
> > +}
> > +
> > +static const struct file_operations data_vf_fops = {
> > + .owner = THIS_MODULE,
> > + .open = simple_open,
> > + .write = data_write,
> > + .read = data_read,
> > + .llseek = default_llseek,
> > +};
> > +
> > static void pf_populate_vf(struct xe_device *xe, struct dentry *vfdent)
> > {
> > debugfs_create_file("pause", 0200, vfdent, xe, &pause_vf_fops);
> > @@ -167,6 +211,7 @@ static void pf_populate_vf(struct xe_device *xe, struct
> > dentry *vfdent)
> > debugfs_create_file("reset", 0200, vfdent, xe, &reset_vf_fops);
> > debugfs_create_file("save", 0600, vfdent, xe, &save_vf_fops);
> > debugfs_create_file("restore", 0600, vfdent, xe, &restore_vf_fops);
> > + debugfs_create_file("migration_data", 0600, vfdent, xe, &data_vf_fops);
> > }
> >
> > static void pf_populate_with_tiles(struct xe_device *xe, struct dentry
> > *dent, unsigned int vfid)
> > diff --git a/drivers/gpu/drm/xe/xe_sriov_pf_migration.c
> > b/drivers/gpu/drm/xe/xe_sriov_pf_migration.c
> > index d39cee66589b5..9cc178126cbdc 100644
> > --- a/drivers/gpu/drm/xe/xe_sriov_pf_migration.c
> > +++ b/drivers/gpu/drm/xe/xe_sriov_pf_migration.c
> > @@ -56,6 +56,18 @@ static bool pf_check_migration_support(struct xe_device
> > *xe)
> > return IS_ENABLED(CONFIG_DRM_XE_DEBUG);
> > }
> >
> > +static void pf_migration_cleanup(struct drm_device *dev, void *arg)
> > +{
> > + struct xe_sriov_pf_migration *migration = arg;
> > +
> > + if (!IS_ERR_OR_NULL(migration->pending))
> > + xe_sriov_pf_migration_data_free(migration->pending);
> > + if (!IS_ERR_OR_NULL(migration->trailer))
> > + xe_sriov_pf_migration_data_free(migration->trailer);
> > + if (!IS_ERR_OR_NULL(migration->descriptor))
> > + xe_sriov_pf_migration_data_free(migration->descriptor);
>
> maybe instead of checking IS_ERR_OR_NULL here, move the check to data_free() ?
Ok.
>
> > +}
> > +
> > /**
> > * xe_sriov_pf_migration_init() - Initialize support for SR-IOV VF
> > migration.
> > * @xe: the &struct xe_device
> > @@ -65,6 +77,7 @@ static bool pf_check_migration_support(struct xe_device
> > *xe)
> > int xe_sriov_pf_migration_init(struct xe_device *xe)
> > {
> > unsigned int n, totalvfs;
> > + int err;
> >
> > xe_assert(xe, IS_SRIOV_PF(xe));
> >
> > @@ -76,7 +89,15 @@ int xe_sriov_pf_migration_init(struct xe_device *xe)
> > for (n = 1; n <= totalvfs; n++) {
> > struct xe_sriov_pf_migration *migration = pf_pick_migration(xe,
> > n);
> >
> > + err = drmm_mutex_init(&xe->drm, &migration->lock);
> > + if (err)
> > + return err;
> > +
> > init_waitqueue_head(&migration->wq);
> > +
> > + err = drmm_add_action_or_reset(&xe->drm, pf_migration_cleanup,
> > migration);
> > + if (err)
> > + return err;
> > }
> >
> > return 0;
> > @@ -162,6 +183,36 @@ xe_sriov_pf_migration_consume(struct xe_device *xe,
> > unsigned int vfid)
> > return data;
> > }
> >
> > +static int pf_handle_descriptor(struct xe_device *xe, unsigned int vfid,
> > + struct xe_sriov_pf_migration_data *data)
> > +{
> > + if (data->tile != 0 || data->gt != 0)
> > + return -EINVAL;
> > +
> > + xe_sriov_pf_migration_data_free(data);
> > +
> > + return 0;
> > +}
> > +
> > +static int pf_handle_trailer(struct xe_device *xe, unsigned int vfid,
> > + struct xe_sriov_pf_migration_data *data)
> > +{
> > + struct xe_gt *gt;
> > + u8 gt_id;
> > +
> > + if (data->tile != 0 || data->gt != 0)
> > + return -EINVAL;
> > + if (data->offset != 0 || data->size != 0 || data->buff || data->bo)
> > + return -EINVAL;
>
> who will free the data packet if we return errors here?
The caller is responsible for releasing the packet in case of errors.
For GT-level packetes this goes to the ring and gets processed (and
freed) by the control worker. For descriptor / trailer it's processed
immediately, but we still follow the caller releasing upon failure to
produce pattern.
>
> > +
> > + xe_sriov_pf_migration_data_free(data);
> > +
> > + for_each_gt(gt, xe, gt_id)
> > + xe_gt_sriov_pf_control_vf_data_eof(gt, vfid);
> > +
> > + return 0;
> > +}
> > +
> > /**
> > * xe_sriov_pf_migration_produce() - Produce a SR-IOV VF migration data
> > packet for device to process
> > * @xe: the &struct xe_device
> > @@ -180,6 +231,11 @@ int xe_sriov_pf_migration_produce(struct xe_device
> > *xe, unsigned int vfid,
> > if (!IS_SRIOV_PF(xe))
> > return -ENODEV;
> >
> > + if (data->type == XE_SRIOV_MIG_DATA_DESCRIPTOR)
> > + return pf_handle_descriptor(xe, vfid, data);
> > + else if (data->type == XE_SRIOV_MIG_DATA_TRAILER)
> > + return pf_handle_trailer(xe, vfid, data);
> > +
> > gt = xe_device_get_gt(xe, data->gt);
> > if (!gt || data->tile != gt->tile->id) {
> > xe_sriov_err_ratelimited(xe, "VF%d Unknown GT - tile_id:%d,
> > gt_id:%d\n",
> > diff --git a/drivers/gpu/drm/xe/xe_sriov_pf_migration_data.c
> > b/drivers/gpu/drm/xe/xe_sriov_pf_migration_data.c
> > index cfc6b512c6674..9a2777dcf9a6b 100644
> > --- a/drivers/gpu/drm/xe/xe_sriov_pf_migration_data.c
> > +++ b/drivers/gpu/drm/xe/xe_sriov_pf_migration_data.c
> > @@ -5,7 +5,45 @@
> >
> > #include "xe_bo.h"
> > #include "xe_device.h"
> > +#include "xe_sriov_pf_helpers.h"
> > +#include "xe_sriov_pf_migration.h"
> > #include "xe_sriov_pf_migration_data.h"
> > +#include "xe_sriov_printk.h"
> > +
> > +static struct mutex *pf_migration_mutex(struct xe_device *xe, unsigned int
> > vfid)
> > +{
> > + xe_assert(xe, IS_SRIOV_PF(xe));
> > + xe_assert(xe, vfid <= xe_sriov_pf_get_totalvfs(xe));
> > + return &xe->sriov.pf.vfs[vfid].migration.lock;
> > +}
> > +
> > +static struct xe_sriov_pf_migration_data **pf_pick_pending(struct
> > xe_device *xe, unsigned int vfid)
> > +{
> > + xe_assert(xe, IS_SRIOV_PF(xe));
> > + xe_assert(xe, vfid <= xe_sriov_pf_get_totalvfs(xe));
> > + lockdep_assert_held(pf_migration_mutex(xe, vfid));
> > +
> > + return &xe->sriov.pf.vfs[vfid].migration.pending;
> > +}
> > +
> > +static struct xe_sriov_pf_migration_data **
> > +pf_pick_descriptor(struct xe_device *xe, unsigned int vfid)
> > +{
> > + xe_assert(xe, IS_SRIOV_PF(xe));
> > + xe_assert(xe, vfid <= xe_sriov_pf_get_totalvfs(xe));
> > + lockdep_assert_held(pf_migration_mutex(xe, vfid));
> > +
> > + return &xe->sriov.pf.vfs[vfid].migration.descriptor;
> > +}
> > +
> > +static struct xe_sriov_pf_migration_data **pf_pick_trailer(struct
> > xe_device *xe, unsigned int vfid)
> > +{
> > + xe_assert(xe, IS_SRIOV_PF(xe));
> > + xe_assert(xe, vfid <= xe_sriov_pf_get_totalvfs(xe));
> > + lockdep_assert_held(pf_migration_mutex(xe, vfid));
> > +
> > + return &xe->sriov.pf.vfs[vfid].migration.trailer;
> > +}
> >
> > static bool data_needs_bo(struct xe_sriov_pf_migration_data *data)
> > {
> > @@ -133,3 +171,318 @@ int xe_sriov_pf_migration_data_init_from_hdr(struct
> > xe_sriov_pf_migration_data *
> >
> > return mig_data_init(data);
> > }
> > +
> > +static ssize_t vf_mig_data_hdr_read(struct xe_sriov_pf_migration_data
> > *data,
> > + char __user *buf, size_t len)
> > +{
> > + loff_t offset = sizeof(data->hdr) - data->hdr_remaining;
> > +
> > + if (!data->hdr_remaining)
> > + return -EINVAL;
> > +
> > + if (len > data->hdr_remaining)
> > + len = data->hdr_remaining;
> > +
> > + if (copy_to_user(buf, (void *)&data->hdr + offset, len))
> > + return -EFAULT;
> > +
> > + data->hdr_remaining -= len;
> > +
> > + return len;
> > +}
> > +
> > +static ssize_t vf_mig_data_read(struct xe_sriov_pf_migration_data *data,
> > + char __user *buf, size_t len)
> > +{
> > + if (len > data->remaining)
> > + len = data->remaining;
> > +
> > + if (copy_to_user(buf, data->vaddr + (data->size - data->remaining),
> > len))
> > + return -EFAULT;
> > +
> > + data->remaining -= len;
> > +
> > + return len;
> > +}
> > +
> > +static ssize_t __vf_mig_data_read_single(struct xe_sriov_pf_migration_data
> > **data,
> > + unsigned int vfid, char __user *buf,
> > size_t len)
> > +{
> > + ssize_t copied = 0;
> > +
> > + if ((*data)->hdr_remaining)
> > + copied = vf_mig_data_hdr_read(*data, buf, len);
> > + else
> > + copied = vf_mig_data_read(*data, buf, len);
> > +
> > + if ((*data)->remaining == 0 && (*data)->hdr_remaining == 0) {
> > + xe_sriov_pf_migration_data_free(*data);
> > + *data = NULL;
> > + }
> > +
> > + return copied;
> > +}
> > +
> > +static struct xe_sriov_pf_migration_data **vf_mig_pick_data(struct
> > xe_device *xe, unsigned int vfid)
> > +{
> > + struct xe_sriov_pf_migration_data **data;
> > +
> > + data = pf_pick_descriptor(xe, vfid);
> > + if (*data)
> > + return data;
> > +
> > + data = pf_pick_pending(xe, vfid);
> > + if (*data == NULL)
> > + *data = xe_sriov_pf_migration_consume(xe, vfid);
> > + if (!IS_ERR_OR_NULL(*data))
> > + return data;
> > + else if (IS_ERR(*data) && PTR_ERR(*data) != -ENODATA)
> > + return data;
> > +
> > + data = pf_pick_trailer(xe, vfid);
> > + if (*data)
> > + return data;
> > +
> > + return ERR_PTR(-ENODATA);
> > +}
> > +
> > +static ssize_t vf_mig_data_read_single(struct xe_device *xe, unsigned int
> > vfid,
> > + char __user *buf, size_t len)
> > +{
> > + struct xe_sriov_pf_migration_data **data = vf_mig_pick_data(xe, vfid);
> > +
> > + if (IS_ERR_OR_NULL(data))
> > + return PTR_ERR(data);
> > +
> > + return __vf_mig_data_read_single(data, vfid, buf, len);
> > +}
> > +
> > +/**
> > + * xe_sriov_pf_migration_data_read() - Read migration data from the device
> > + * @gt: the &struct xe_device
>
> @xe
Ok.
>
> > + * @vfid: the VF identifier
> > + * @buf: start address of userspace buffer
> > + * @len: requested read size from userspace
> > + *
> > + * Return: number of bytes that has been successfully read
> > + * 0 if no more migration data is available
> > + * -errno on failure
>
> you likely need to add some punctuation here to properly render the doc
Ok.
>
> > + */
> > +ssize_t xe_sriov_pf_migration_data_read(struct xe_device *xe, unsigned int
> > vfid,
> > + char __user *buf, size_t len)
> > +{
> > + ssize_t ret, consumed = 0;
> > +
> > + xe_assert(xe, IS_SRIOV_PF(xe));
> > +
> > + ret = mutex_lock_interruptible(pf_migration_mutex(xe, vfid));
> > + if (ret)
> > + return ret;
> > +
> > + while (consumed < len) {
> > + ret = vf_mig_data_read_single(xe, vfid, buf, len - consumed);
> > + if (ret == -ENODATA)
> > + goto out;
> > + if (ret < 0) {
> > + mutex_unlock(pf_migration_mutex(xe, vfid));
> > + return ret;
> > + }
> > +
> > + consumed += ret;
> > + buf += ret;
> > + }
> > +
> > +out:
> > + mutex_unlock(pf_migration_mutex(xe, vfid));
> > + return consumed;
> > +}
> > +
> > +static ssize_t vf_mig_hdr_write(struct xe_sriov_pf_migration_data *data,
> > + const char __user *buf, size_t len)
> > +{
> > + loff_t offset = sizeof(data->hdr) - data->hdr_remaining;
> > + int ret;
> > +
> > + if (WARN_ON(!data->hdr_remaining))
>
> xe_WARN_ON(xe, ... ) if having full WARN is really important
Removed.
>
> > + return -EINVAL;
> > +
> > + if (len > data->hdr_remaining)
> > + len = data->hdr_remaining;
> > +
> > + if (copy_from_user((void *)&data->hdr + offset, buf, len))
> > + return -EFAULT;
> > +
> > + data->hdr_remaining -= len;
> > +
> > + if (!data->hdr_remaining) {
> > + ret = xe_sriov_pf_migration_data_init_from_hdr(data);
> > + if (ret)
> > + return ret;
> > + }
> > +
> > + return len;
> > +}
> > +
> > +static ssize_t vf_mig_data_write(struct xe_sriov_pf_migration_data *data,
> > + const char __user *buf, size_t len)
> > +{
> > + if (len > data->remaining)
> > + len = data->remaining;
> > +
> > + if (copy_from_user(data->vaddr + (data->size - data->remaining), buf,
> > len))
> > + return -EFAULT;
> > +
> > + data->remaining -= len;
> > +
> > + return len;
> > +}
> > +
> > +static ssize_t vf_mig_data_write_single(struct xe_device *xe, unsigned int
> > vfid,
> > + const char __user *buf, size_t len)
> > +{
> > + struct xe_sriov_pf_migration_data **data = pf_pick_pending(xe, vfid);
> > + int ret;
> > + ssize_t copied;
> > +
> > + if (IS_ERR_OR_NULL(*data)) {
> > + *data = xe_sriov_pf_migration_data_alloc(xe);
> > + if (*data == NULL)
> > + return -ENOMEM;
> > + }
> > +
> > + if ((*data)->hdr_remaining)
> > + copied = vf_mig_hdr_write(*data, buf, len);
> > + else
> > + copied = vf_mig_data_write(*data, buf, len);
> > +
> > + if ((*data)->hdr_remaining == 0 && (*data)->remaining == 0) {
> > + ret = xe_sriov_pf_migration_produce(xe, vfid, *data);
> > + if (ret) {
> > + xe_sriov_pf_migration_data_free(*data);
> > + return ret;
> > + }
> > +
> > + *data = NULL;
> > + }
> > +
> > + return copied;
> > +}
> > +
> > +/**
> > + * xe_sriov_pf_migration_data_write() - Write migration data to the device
> > + * @gt: the &struct xe_device
>
> @xe
Ok.
>
> > + * @vfid: the VF identifier
> > + * @buf: start address of userspace buffer
> > + * @len: requested write size from userspace
> > + *
> > + * Return: number of bytes that has been successfully written
> > + * -errno on failure
> > + */
> > +ssize_t xe_sriov_pf_migration_data_write(struct xe_device *xe, unsigned
> > int vfid,
> > + const char __user *buf, size_t len)
> > +{
> > + ssize_t ret, produced = 0;
> > +
> > + xe_assert(xe, IS_SRIOV_PF(xe));
> > +
> > + ret = mutex_lock_interruptible(pf_migration_mutex(xe, vfid));
> > + if (ret)
> > + return ret;
>
> scoped_cond_guard(mutex_intr, return -EINTR, pf_migration_mutex(xe, vfid)) ?
Ok.
Thanks,
-Michał
>
> > +
> > + while (produced < len) {
> > + ret = vf_mig_data_write_single(xe, vfid, buf, len - produced);
> > + if (ret < 0) {
> > + mutex_unlock(pf_migration_mutex(xe, vfid));
> > + return ret;
> > + }
> > +
> > + produced += ret;
> > + buf += ret;
> > + }
> > +
> > + mutex_unlock(pf_migration_mutex(xe, vfid));
> > + return produced;
> > +}
> > +
> > +#define MIGRATION_DESC_SIZE 4
> > +static size_t pf_desc_init(struct xe_device *xe, unsigned int vfid)
> > +{
> > + struct xe_sriov_pf_migration_data **desc = pf_pick_descriptor(xe, vfid);
> > + struct xe_sriov_pf_migration_data *data;
> > + int ret;
> > +
> > + data = xe_sriov_pf_migration_data_alloc(xe);
> > + if (!data)
> > + return -ENOMEM;
> > +
> > + ret = xe_sriov_pf_migration_data_init(data, 0, 0,
> > XE_SRIOV_MIG_DATA_DESCRIPTOR,
> > + 0, MIGRATION_DESC_SIZE);
> > + if (ret) {
> > + xe_sriov_pf_migration_data_free(data);
> > + return ret;
> > + }
> > +
> > + *desc = data;
> > +
> > + return 0;
> > +}
> > +
> > +static void pf_pending_init(struct xe_device *xe, unsigned int vfid)
> > +{
> > + struct xe_sriov_pf_migration_data **data = pf_pick_pending(xe, vfid);
> > +
> > + *data = NULL;
> > +}
> > +
> > +#define MIGRATION_TRAILER_SIZE 0
> > +static int pf_trailer_init(struct xe_device *xe, unsigned int vfid)
> > +{
> > + struct xe_sriov_pf_migration_data **trailer = pf_pick_trailer(xe, vfid);
> > + struct xe_sriov_pf_migration_data *data;
> > + int ret;
> > +
> > + data = xe_sriov_pf_migration_data_alloc(xe);
> > + if (!data)
> > + return -ENOMEM;
> > +
> > + ret = xe_sriov_pf_migration_data_init(data, 0, 0,
> > XE_SRIOV_MIG_DATA_TRAILER,
> > + 0, MIGRATION_TRAILER_SIZE);
> > + if (ret) {
> > + xe_sriov_pf_migration_data_free(data);
> > + return ret;
> > + }
> > +
> > + *trailer = data;
> > +
> > + return 0;
> > +}
> > +
> > +/**
> > + * xe_sriov_pf_migration_data_save_init() - Initialize the pending save
> > migration data.
> > + * @gt: the &struct xe_device
> > + * @vfid: the VF identifier
> > + *
> > + * Return: 0 on success, -errno on failure
> > + */
> > +int xe_sriov_pf_migration_data_save_init(struct xe_device *xe, unsigned
> > int vfid)
> > +{
> > + int ret;
> > +
> > + ret = mutex_lock_interruptible(pf_migration_mutex(xe, vfid));
> > + if (ret)
> > + return ret;
> > +
> > + ret = pf_desc_init(xe, vfid);
> > + if (ret)
> > + goto out;
> > +
> > + ret = pf_trailer_init(xe, vfid);
> > + if (ret)
> > + goto out;
> > +
> > + pf_pending_init(xe, vfid);
> > +
> > +out:
> > + mutex_unlock(pf_migration_mutex(xe, vfid));
> > + return ret;
> > +}
> > diff --git a/drivers/gpu/drm/xe/xe_sriov_pf_migration_data.h
> > b/drivers/gpu/drm/xe/xe_sriov_pf_migration_data.h
> > index 1dde4cfcdbc47..5b96c7f224002 100644
> > --- a/drivers/gpu/drm/xe/xe_sriov_pf_migration_data.h
> > +++ b/drivers/gpu/drm/xe/xe_sriov_pf_migration_data.h
> > @@ -28,5 +28,10 @@ void xe_sriov_pf_migration_data_free(struct
> > xe_sriov_pf_migration_data *snapshot
> > int xe_sriov_pf_migration_data_init(struct xe_sriov_pf_migration_data
> > *data, u8 tile_id, u8 gt_id,
> > unsigned int type, loff_t offset, size_t
> > size);
> > int xe_sriov_pf_migration_data_init_from_hdr(struct
> > xe_sriov_pf_migration_data *snapshot);
> > +ssize_t xe_sriov_pf_migration_data_read(struct xe_device *xe, unsigned int
> > vfid,
> > + char __user *buf, size_t len);
> > +ssize_t xe_sriov_pf_migration_data_write(struct xe_device *xe, unsigned
> > int vfid,
> > + const char __user *buf, size_t len);
> > +int xe_sriov_pf_migration_data_save_init(struct xe_device *xe, unsigned
> > int vfid);
> >
> > #endif
> > diff --git a/drivers/gpu/drm/xe/xe_sriov_pf_migration_types.h
> > b/drivers/gpu/drm/xe/xe_sriov_pf_migration_types.h
> > index 80fdea32b884a..c5d75bb7f39c0 100644
> > --- a/drivers/gpu/drm/xe/xe_sriov_pf_migration_types.h
> > +++ b/drivers/gpu/drm/xe/xe_sriov_pf_migration_types.h
> > @@ -7,6 +7,7 @@
> > #define _XE_SRIOV_PF_MIGRATION_TYPES_H_
> >
> > #include <linux/types.h>
> > +#include <linux/mutex_types.h>
> > #include <linux/wait.h>
> >
> > struct xe_sriov_pf_migration_data {
> > @@ -32,6 +33,14 @@ struct xe_sriov_pf_migration_data {
> > struct xe_sriov_pf_migration {
> > /** @wq: waitqueue used to avoid busy-waiting for snapshot
> > production/consumption */
> > wait_queue_head_t wq;
> > + /** @lock: Mutex protecting the migration data */
> > + struct mutex lock;
> > + /** @pending: currently processed data packet of VF resource */
> > + struct xe_sriov_pf_migration_data *pending;
> > + /** @trailer: data packet used to indicate the end of stream */
> > + struct xe_sriov_pf_migration_data *trailer;
> > + /** @descriptor: data packet containing the metadata describing the
> > device */
> > + struct xe_sriov_pf_migration_data *descriptor;
> > };
> >
> > #endif
>
