[RFC 2/2] drm/xe/FLR: Support PCIe FLR
Aravind Iddamsetty
aravind.iddamsetty at linux.intel.com
Thu Mar 21 11:38:28 UTC 2024
On 21/03/24 04:31, Lucas De Marchi wrote:
> On Wed, Mar 20, 2024 at 04:14:26PM +0530, Aravind Iddamsetty wrote:
>> PCI subsystem provides callbacks to inform the driver about a request to
>> do function level reset by user, initiated by writing to sysfs entry
>> /sys/bus/pci/devices/.../reset. This will allow the driver to handle FLR
>> without the need to do unbind and rebind as the driver needs to
>> reinitialize the device afresh post FLR.
>>
>> Cc: Rodrigo Vivi <rodrigo.vivi at intel.com>
>> Cc: Lucas De Marchi <lucas.demarchi at intel.com>
>> Signed-off-by: Aravind Iddamsetty <aravind.iddamsetty at linux.intel.com>
>> ---
>> drivers/gpu/drm/xe/Makefile | 1 +
>> drivers/gpu/drm/xe/xe_device_types.h | 3 +
>> drivers/gpu/drm/xe/xe_gt.c | 31 ++++++---
>> drivers/gpu/drm/xe/xe_gt.h | 1 +
>> drivers/gpu/drm/xe/xe_pci.c | 53 ++++++++++++++--
>> drivers/gpu/drm/xe/xe_pci.h | 6 +-
>> drivers/gpu/drm/xe/xe_pci_err.c | 94 ++++++++++++++++++++++++++++
>> 7 files changed, 174 insertions(+), 15 deletions(-)
>> create mode 100644 drivers/gpu/drm/xe/xe_pci_err.c
>>
>> diff --git a/drivers/gpu/drm/xe/Makefile b/drivers/gpu/drm/xe/Makefile
>> index 3c3e67885559..1447712fec65 100644
>> --- a/drivers/gpu/drm/xe/Makefile
>> +++ b/drivers/gpu/drm/xe/Makefile
>> @@ -114,6 +114,7 @@ xe-y += xe_bb.o \
>> xe_module.o \
>> xe_pat.o \
>> xe_pci.o \
>> + xe_pci_err.o \
>> xe_pcode.o \
>> xe_pm.o \
>> xe_preempt_fence.o \
>> diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h
>> index 9785eef2e5a4..e9b8c7cbb428 100644
>> --- a/drivers/gpu/drm/xe/xe_device_types.h
>> +++ b/drivers/gpu/drm/xe/xe_device_types.h
>> @@ -455,6 +455,9 @@ struct xe_device {
>> /** @needs_flr_on_fini: requests function-reset on fini */
>> bool needs_flr_on_fini;
>>
>> + /** @pci_state: PCI state of device */
>> + struct pci_saved_state *pci_state;
>> +
>> /* private: */
>>
>> #if IS_ENABLED(CONFIG_DRM_XE_DISPLAY)
>> diff --git a/drivers/gpu/drm/xe/xe_gt.c b/drivers/gpu/drm/xe/xe_gt.c
>> index 85408e7a932b..437874a9a5a0 100644
>> --- a/drivers/gpu/drm/xe/xe_gt.c
>> +++ b/drivers/gpu/drm/xe/xe_gt.c
>> @@ -629,6 +629,26 @@ static int do_gt_restart(struct xe_gt *gt)
>> return 0;
>> }
>>
>> +/* Idle the GT */
>> +int xe_idle_gt(struct xe_gt *gt)
>
> any non-static function should use xe_gt_ prefix.
Ok will change it.
>
>> +{
>> + int err;
>> +
>> + xe_gt_sanitize(gt);
>> +
>> + xe_uc_gucrc_disable(>->uc);
>> + xe_uc_stop_prepare(>->uc);
>> + xe_gt_pagefault_reset(gt);
>> +
>> + err = xe_uc_stop(>->uc);
>> + if (err)
>> + return err;
>> +
>> + xe_gt_tlb_invalidation_reset(gt);
>> +
>> + return err;
>> +}
>> +
>> static int gt_reset(struct xe_gt *gt)
>> {
>> int err;
>> @@ -645,21 +665,12 @@ static int gt_reset(struct xe_gt *gt)
>> }
>>
>> xe_pm_runtime_get(gt_to_xe(gt));
>> - xe_gt_sanitize(gt);
>>
>> err = xe_force_wake_get(gt_to_fw(gt), XE_FORCEWAKE_ALL);
>> if (err)
>> goto err_msg;
>>
>> - xe_uc_gucrc_disable(>->uc);
>> - xe_uc_stop_prepare(>->uc);
>> - xe_gt_pagefault_reset(gt);
>> -
>> - err = xe_uc_stop(>->uc);
>> - if (err)
>> - goto err_out;
>> -
>> - xe_gt_tlb_invalidation_reset(gt);
>> + xe_idle_gt(gt);
>
> this and the above should be in a commit alone
> "drm/xe: Extract xe_gt_idle() helper" with explanation it will be used
> in other places outside of gt_reset path.
sure will separate it out.
>
> but I'm a little bit confused here... why do you need to remove
> xe_gt_sanitize() inside the function to make gt idle?
if I understood right this controls how the invalidation is
sent via guc or mmio. hence moved into this function as we disable guc.
>
> Lucas De Marchi
>
>>
>> err = do_gt_reset(gt);
>> if (err)
>> diff --git a/drivers/gpu/drm/xe/xe_gt.h b/drivers/gpu/drm/xe/xe_gt.h
>> index ed6ea8057e35..77df919199cc 100644
>> --- a/drivers/gpu/drm/xe/xe_gt.h
>> +++ b/drivers/gpu/drm/xe/xe_gt.h
>> @@ -43,6 +43,7 @@ int xe_gt_resume(struct xe_gt *gt);
>> void xe_gt_reset_async(struct xe_gt *gt);
>> void xe_gt_sanitize(struct xe_gt *gt);
>> void xe_gt_remove(struct xe_gt *gt);
>> +int xe_idle_gt(struct xe_gt *gt);
>>
>> /**
>> * xe_gt_any_hw_engine_by_reset_domain - scan the list of engines and return the
>> diff --git a/drivers/gpu/drm/xe/xe_pci.c b/drivers/gpu/drm/xe/xe_pci.c
>> index c401d4890386..fcd2a7f66f7b 100644
>> --- a/drivers/gpu/drm/xe/xe_pci.c
>> +++ b/drivers/gpu/drm/xe/xe_pci.c
>> @@ -383,6 +383,41 @@ MODULE_DEVICE_TABLE(pci, pciidlist);
>>
>> #undef INTEL_VGA_DEVICE
>>
>> +static bool xe_save_pci_state(struct pci_dev *pdev)
>> +{
>> + struct xe_device *xe = pci_get_drvdata(pdev);
>> +
>> + if (pci_save_state(pdev))
>> + return false;
>> +
>> + kfree(xe->pci_state);
>> +
>> + xe->pci_state = pci_store_saved_state(pdev);
>> +
>> + if (!xe->pci_state) {
>> + drm_err(&xe->drm, "Failed to store PCI saved state\n");
>> + return false;
>> + }
>> +
>> + return true;
>> +}
>> +
>> +void xe_load_pci_state(struct pci_dev *pdev)
>> +{
>> + struct xe_device *xe = pci_get_drvdata(pdev);
>> + int ret;
>> +
>> + if (!xe->pci_state)
>> + return;
>> +
>> + ret = pci_load_saved_state(pdev, xe->pci_state);
>> + if (!ret) {
>> + pci_restore_state(pdev);
>> + } else {
>> + drm_warn(&xe->drm, "Failed to load PCI state err:%d\n", ret);
>> + }
>
> please check coding style here and invert condition so you keep the
> return-early pattern you started in the xe->pci_state check.
>
>
> ret = pci_load_saved_state(pdev, xe->pci_state);
> if (ret) {
> drm_warn(&xe->drm, "Failed to load PCI state err:%d\n", ret);
> return;
> }
Ok will correct it.
>
>> +}
>> +
>> /* is device_id present in comma separated list of ids */
>> static bool device_id_in_list(u16 device_id, const char *devices, bool negative)
>> {
>> @@ -688,10 +723,12 @@ static void xe_pci_remove(struct pci_dev *pdev)
>>
>> xe_device_remove(xe);
>> xe_pm_runtime_fini(xe);
>> +
>> + kfree(xe->pci_state);
>> pci_set_drvdata(pdev, NULL);
>> }
>>
>> -static int xe_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
>> +int xe_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
>> {
>> const struct xe_device_desc *desc = (const void *)ent->driver_data;
>> const struct xe_subplatform_desc *subplatform_desc;
>> @@ -786,6 +823,9 @@ static int xe_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
>> drm_dbg(&xe->drm, "d3cold: capable=%s\n",
>> str_yes_no(xe->d3cold.capable));
>>
>> + if (xe_save_pci_state(pdev))
>> + pci_restore_state(pdev);
>
> why do we save and restore in sequence when probing? If we only have to
> have a saved stated for possible errors later, why do we restore on
> success? /me confused
this is needed to avoid breaking the PCI core functionality of suspend/resume
to unset dev->state_saved which is checked in suspend
Documentation/power/pci.rst:
"Moreover, if the driver calls pci_save_state(),
the PCI subsystem will not execute either pci_prepare_to_sleep(), or
pci_set_power_state() for its device, so the driver is then responsible for
handling the device as appropriate."
>
>> +
>> return 0;
>> }
>>
>> @@ -833,7 +873,7 @@ static int xe_pci_suspend(struct device *dev)
>> */
>> d3cold_toggle(pdev, D3COLD_ENABLE);
>>
>> - pci_save_state(pdev);
>> + xe_save_pci_state(pdev);
>
> all these no change in behavior with xe_ functions wrapping the the pci
> ones could be delegated to a prep patch so it's easier to separate the
> noise from the actual changes.
Ok.
>
>> pci_disable_device(pdev);
>>
>> return 0;
>> @@ -857,6 +897,8 @@ static int xe_pci_resume(struct device *dev)
>>
>> pci_set_master(pdev);
>>
>> + xe_load_pci_state(pdev);
>> +
>> err = xe_pm_resume(pdev_to_xe_device(pdev));
>> if (err)
>> return err;
>> @@ -874,7 +916,7 @@ static int xe_pci_runtime_suspend(struct device *dev)
>> if (err)
>> return err;
>>
>> - pci_save_state(pdev);
>> + xe_save_pci_state(pdev);
>>
>> if (xe->d3cold.allowed) {
>> d3cold_toggle(pdev, D3COLD_ENABLE);
>> @@ -899,7 +941,7 @@ static int xe_pci_runtime_resume(struct device *dev)
>> if (err)
>> return err;
>>
>> - pci_restore_state(pdev);
>> + xe_load_pci_state(pdev);
>>
>> if (xe->d3cold.allowed) {
>> err = pci_enable_device(pdev);
>> @@ -928,6 +970,8 @@ static const struct dev_pm_ops xe_pm_ops = {
>> };
>> #endif
>>
>> +extern const struct pci_error_handlers xe_pci_err_handlers;
>
> not sure this really belongs into a separate compilation unit. Just
> leave it in xe_pci.c? Adding the extern here doesn't make it pretty.
>
I wanted to keep it separate so that it will be clean, will do as suggested below.
Thanks,
Aravind.
> if we are adding a separate .c, then the struct definition should
> probably be here, and you do it like
>
> static const struct pci_error_handlers xe_pci_err_handlers = {
> .reset_prepare = xe_pci_err_reset_prepare,
> .reset_done = xe_pci_err_reset_done,
> };
>
>
> See xe_pm_ops, that was left here rather than in xe_pm.c.
>
> Lucas De Marchi
>
>> +
>> static struct pci_driver xe_pci_driver = {
>> .name = DRIVER_NAME,
>> .id_table = pciidlist,
>> @@ -937,6 +981,7 @@ static struct pci_driver xe_pci_driver = {
>> #ifdef CONFIG_PM_SLEEP
>> .driver.pm = &xe_pm_ops,
>> #endif
>> + .err_handler = &xe_pci_err_handlers,
>> };
>>
>> int xe_register_pci_driver(void)
>> diff --git a/drivers/gpu/drm/xe/xe_pci.h b/drivers/gpu/drm/xe/xe_pci.h
>> index 611c1209b14c..9faf5380a09e 100644
>> --- a/drivers/gpu/drm/xe/xe_pci.h
>> +++ b/drivers/gpu/drm/xe/xe_pci.h
>> @@ -6,7 +6,11 @@
>> #ifndef _XE_PCI_H_
>> #define _XE_PCI_H_
>>
>> +struct pci_dev;
>> +struct pci_device_id;
>> +
>> int xe_register_pci_driver(void);
>> void xe_unregister_pci_driver(void);
>> -
>> +void xe_load_pci_state(struct pci_dev *pdev);
>> +int xe_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent);
>> #endif
>> diff --git a/drivers/gpu/drm/xe/xe_pci_err.c b/drivers/gpu/drm/xe/xe_pci_err.c
>> new file mode 100644
>> index 000000000000..ecc467a9c3af
>> --- /dev/null
>> +++ b/drivers/gpu/drm/xe/xe_pci_err.c
>> @@ -0,0 +1,94 @@
>> +// SPDX-License-Identifier: MIT
>> +/*
>> + * Copyright © 2024 Intel Corporation
>> + */
>> +
>> +#include <linux/pci.h>
>> +#include <drm/drm_drv.h>
>> +
>> +#include "xe_device.h"
>> +#include "xe_gt.h"
>> +#include "xe_gt_printk.h"
>> +#include "xe_pci.h"
>> +#include "xe_pm.h"
>> +
>> +/**
>> + * xe_pci_reset_prepare - Called when user issued a function level reset
>> + * via /sys/bus/pci/devices/.../reset.
>> + * @pdev: PCI device struct
>> + */
>> +static void xe_pci_reset_prepare(struct pci_dev *pdev)
>> +{
>> + struct xe_device *xe = pci_get_drvdata(pdev);
>> + struct xe_gt *gt;
>> + int id, err;
>> +
>> + pci_warn(pdev, "preparing for PCIe FLR reset\n");
>> +
>> + drm_warn(&xe->drm, "removing device access to userspace\n");
>> + drm_dev_unplug(&xe->drm);
>> +
>> + xe_pm_runtime_get(xe);
>> + /* idle the GTs */
>> + for_each_gt(gt, xe, id) {
>> + err = xe_force_wake_get(gt_to_fw(gt), XE_FORCEWAKE_ALL);
>> + if (err)
>> + goto reset;
>> + err = xe_idle_gt(gt);
>> + if (err) {
>> + xe_gt_err(gt, "failed to idle gt (%pe)\n", ERR_PTR(err));
>> + goto reset;
>> + }
>> +
>> + err = xe_force_wake_put(gt_to_fw(gt), XE_FORCEWAKE_ALL);
>> + XE_WARN_ON(err);
>> + }
>> + xe_pm_runtime_put(xe);
>> +
>> +reset:
>> + pci_disable_device(pdev);
>> +}
>> +
>> +/**
>> + * xe_pci_reset_done - Called when function level reset is done.
>> + * @pdev: PCI device struct
>> + */
>> +static void xe_pci_reset_done(struct pci_dev *pdev)
>> +{
>> + const struct pci_device_id *ent = pci_match_id(pdev->driver->id_table, pdev);
>> + struct xe_device *xe = pci_get_drvdata(pdev);
>> +
>> + dev_info(&pdev->dev,
>> + "PCI device went through FLR, reenabling the device\n");
>> +
>> + if (pci_enable_device(pdev)) {
>> + dev_err(&pdev->dev,
>> + "Cannot re-enable PCI device after reset\n");
>> + return;
>> + }
>> + pci_set_master(pdev);
>> + xe_load_pci_state(pdev);
>> +
>> + /*
>> + * We want to completely clean the driver and even destroy
>> + * the xe private data and reinitialize afresh similar to
>> + * probe
>> + */
>> + pdev->driver->remove(pdev);
>> + if (pci_dev_msi_enabled(pdev))
>> + pci_free_irq_vectors(pdev);
>> +
>> + devm_drm_release_action(&xe->drm);
>> + pci_disable_device(pdev);
>> +
>> + /*
>> + * if this fails the driver might be in a stale state, only option is
>> + * to unbind and rebind
>> + */
>> + xe_pci_probe(pdev, ent);
>> +}
>> +
>> +const struct pci_error_handlers xe_pci_err_handlers = {
>> + .reset_prepare = xe_pci_reset_prepare,
>> + .reset_done = xe_pci_reset_done,
>> +};
>> --
>> 2.25.1
>>
More information about the dri-devel
mailing list