[PATCH v2 0/8] Phase out hybrid PCI devres API

[PATCH v2 0/8] Phase out hybrid PCI devres API

[Philipp Stanner]

Changes in v2:
  - Rename error path. (Jakub)
  - Apply some RBs

Fixes a number of minor issues with the usage of the PCI API in net.
Notbaly, it replaces calls to the sometimes-managed
pci_request_regions() to the always-managed pcim_request_all_regions(),
enabling us to remove that hybrid functionality from PCI.
Philipp Stanner (8):
  net: prestera: Use pure PCI devres API
  net: octeontx2: Use pure PCI devres API
  net: tulip: Use pure PCI devres API
  net: ethernet: natsemi: Use pure PCI devres API
  net: ethernet: sis900: Use pure PCI devres API
  net: mdio: thunder: Use pure PCI devres API
  net: thunder_bgx: Use pure PCI devres API
  net: thunder_bgx: Don't disable PCI device manually

 drivers/net/ethernet/cavium/thunder/thunder_bgx.c  | 13 ++++---------
 drivers/net/ethernet/dec/tulip/tulip_core.c        |  2 +-
 drivers/net/ethernet/dec/tulip/winbond-840.c       |  2 +-
 .../net/ethernet/marvell/octeontx2/nic/otx2_pf.c   | 14 ++++----------
 .../net/ethernet/marvell/octeontx2/nic/otx2_vf.c   | 14 ++++----------
 drivers/net/ethernet/marvell/octeontx2/nic/rep.c   | 12 +++++-------
 .../net/ethernet/marvell/prestera/prestera_pci.c   |  6 ++----
 drivers/net/ethernet/natsemi/natsemi.c             |  2 +-
 drivers/net/ethernet/sis/sis900.c                  |  2 +-
 drivers/net/mdio/mdio-thunder.c                    | 10 +++-------
 10 files changed, 26 insertions(+), 51 deletions(-)

-- 
2.48.1

[PATCH v2 1/8] net: prestera: Use pure PCI devres API

[Philipp Stanner]

The currently used function pci_request_regions() is one of the
problematic "hybrid devres" PCI functions, which are sometimes managed
through devres, and sometimes not (depending on whether
pci_enable_device() or pcim_enable_device() has been called before).

The PCI subsystem wants to remove this behavior and, therefore, needs to
port all users to functions that don't have this problem.

Furthermore, the PCI function being managed implies that it's not
necessary to call pci_release_regions() manually.

Remove the calls to pci_release_regions().

Replace pci_request_regions() with pcim_request_all_regions().

Signed-off-by: Philipp Stanner <phasta@kernel.org>
Acked-by: Elad Nachman <enachman@marvell.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
---
 drivers/net/ethernet/marvell/prestera/prestera_pci.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/marvell/prestera/prestera_pci.c b/drivers/net/ethernet/marvell/prestera/prestera_pci.c
index 35857dc19542..c45d108b2f6d 100644
--- a/drivers/net/ethernet/marvell/prestera/prestera_pci.c
+++ b/drivers/net/ethernet/marvell/prestera/prestera_pci.c
@@ -845,9 +845,9 @@ static int prestera_pci_probe(struct pci_dev *pdev,
 		goto err_pci_enable_device;
 	}
 
-	err = pci_request_regions(pdev, driver_name);
+	err = pcim_request_all_regions(pdev, driver_name);
 	if (err) {
-		dev_err(&pdev->dev, "pci_request_regions failed\n");
+		dev_err(&pdev->dev, "pcim_request_all_regions failed\n");
 		goto err_pci_request_regions;
 	}
 
@@ -938,7 +938,6 @@ static int prestera_pci_probe(struct pci_dev *pdev,
 err_pp_ioremap:
 err_mem_ioremap:
 err_dma_mask:
-	pci_release_regions(pdev);
 err_pci_request_regions:
 err_pci_enable_device:
 	return err;
@@ -953,7 +952,6 @@ static void prestera_pci_remove(struct pci_dev *pdev)
 	pci_free_irq_vectors(pdev);
 	destroy_workqueue(fw->wq);
 	prestera_fw_uninit(fw);
-	pci_release_regions(pdev);
 }
 
 static const struct pci_device_id prestera_pci_devices[] = {
-- 
2.48.1

[PATCH v2 2/8] net: octeontx2: Use pure PCI devres API

[Philipp Stanner]

The currently used function pci_request_regions() is one of the
problematic "hybrid devres" PCI functions, which are sometimes managed
through devres, and sometimes not (depending on whether
pci_enable_device() or pcim_enable_device() has been called before).

The PCI subsystem wants to remove this behavior and, therefore, needs to
port all users to functions that don't have this problem.

Furthermore, the PCI function being managed implies that it's not
necessary to call pci_release_regions() manually.

Remove the calls to pci_release_regions().

Replace pci_request_regions() with pcim_request_all_regions().

Signed-off-by: Philipp Stanner <phasta@kernel.org>
---
 .../net/ethernet/marvell/octeontx2/nic/otx2_pf.c   | 14 ++++----------
 .../net/ethernet/marvell/octeontx2/nic/otx2_vf.c   | 14 ++++----------
 drivers/net/ethernet/marvell/octeontx2/nic/rep.c   | 12 +++++-------
 3 files changed, 13 insertions(+), 27 deletions(-)

diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/otx2_pf.c b/drivers/net/ethernet/marvell/octeontx2/nic/otx2_pf.c
index cfed9ec5b157..0aee8e3861f3 100644
--- a/drivers/net/ethernet/marvell/octeontx2/nic/otx2_pf.c
+++ b/drivers/net/ethernet/marvell/octeontx2/nic/otx2_pf.c
@@ -3048,7 +3048,7 @@ static int otx2_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 		return err;
 	}
 
-	err = pci_request_regions(pdev, DRV_NAME);
+	err = pcim_request_all_regions(pdev, DRV_NAME);
 	if (err) {
 		dev_err(dev, "PCI request regions failed 0x%x\n", err);
 		return err;
@@ -3057,7 +3057,7 @@ static int otx2_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	err = dma_set_mask_and_coherent(dev, DMA_BIT_MASK(48));
 	if (err) {
 		dev_err(dev, "DMA mask config failed, abort\n");
-		goto err_release_regions;
+		return err;
 	}
 
 	pci_set_master(pdev);
@@ -3067,10 +3067,8 @@ static int otx2_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	qos_txqs = min_t(int, qcount, OTX2_QOS_MAX_LEAF_NODES);
 
 	netdev = alloc_etherdev_mqs(sizeof(*pf), qcount + qos_txqs, qcount);
-	if (!netdev) {
-		err = -ENOMEM;
-		goto err_release_regions;
-	}
+	if (!netdev)
+		return -ENOMEM;
 
 	pci_set_drvdata(pdev, netdev);
 	SET_NETDEV_DEV(netdev, &pdev->dev);
@@ -3246,8 +3244,6 @@ static int otx2_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 err_free_netdev:
 	pci_set_drvdata(pdev, NULL);
 	free_netdev(netdev);
-err_release_regions:
-	pci_release_regions(pdev);
 	return err;
 }
 
@@ -3447,8 +3443,6 @@ static void otx2_remove(struct pci_dev *pdev)
 	pci_free_irq_vectors(pf->pdev);
 	pci_set_drvdata(pdev, NULL);
 	free_netdev(netdev);
-
-	pci_release_regions(pdev);
 }
 
 static struct pci_driver otx2_pf_driver = {
diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/otx2_vf.c b/drivers/net/ethernet/marvell/octeontx2/nic/otx2_vf.c
index 7ef3ba477d49..fb4da816d218 100644
--- a/drivers/net/ethernet/marvell/octeontx2/nic/otx2_vf.c
+++ b/drivers/net/ethernet/marvell/octeontx2/nic/otx2_vf.c
@@ -548,7 +548,7 @@ static int otx2vf_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 		return err;
 	}
 
-	err = pci_request_regions(pdev, DRV_NAME);
+	err = pcim_request_all_regions(pdev, DRV_NAME);
 	if (err) {
 		dev_err(dev, "PCI request regions failed 0x%x\n", err);
 		return err;
@@ -557,7 +557,7 @@ static int otx2vf_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	err = dma_set_mask_and_coherent(dev, DMA_BIT_MASK(48));
 	if (err) {
 		dev_err(dev, "DMA mask config failed, abort\n");
-		goto err_release_regions;
+		return err;
 	}
 
 	pci_set_master(pdev);
@@ -565,10 +565,8 @@ static int otx2vf_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	qcount = num_online_cpus();
 	qos_txqs = min_t(int, qcount, OTX2_QOS_MAX_LEAF_NODES);
 	netdev = alloc_etherdev_mqs(sizeof(*vf), qcount + qos_txqs, qcount);
-	if (!netdev) {
-		err = -ENOMEM;
-		goto err_release_regions;
-	}
+	if (!netdev)
+		return -ENOMEM;
 
 	pci_set_drvdata(pdev, netdev);
 	SET_NETDEV_DEV(netdev, &pdev->dev);
@@ -765,8 +763,6 @@ static int otx2vf_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 err_free_netdev:
 	pci_set_drvdata(pdev, NULL);
 	free_netdev(netdev);
-err_release_regions:
-	pci_release_regions(pdev);
 	return err;
 }
 
@@ -815,8 +811,6 @@ static void otx2vf_remove(struct pci_dev *pdev)
 	pci_free_irq_vectors(vf->pdev);
 	pci_set_drvdata(pdev, NULL);
 	free_netdev(netdev);
-
-	pci_release_regions(pdev);
 }
 
 static struct pci_driver otx2vf_driver = {
diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/rep.c b/drivers/net/ethernet/marvell/octeontx2/nic/rep.c
index 04e08e06f30f..0bb8a971c4ff 100644
--- a/drivers/net/ethernet/marvell/octeontx2/nic/rep.c
+++ b/drivers/net/ethernet/marvell/octeontx2/nic/rep.c
@@ -763,7 +763,7 @@ static int rvu_rep_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 		return err;
 	}
 
-	err = pci_request_regions(pdev, DRV_NAME);
+	err = pcim_request_all_regions(pdev, DRV_NAME);
 	if (err) {
 		dev_err(dev, "PCI request regions failed 0x%x\n", err);
 		return err;
@@ -772,7 +772,7 @@ static int rvu_rep_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	err = dma_set_mask_and_coherent(dev, DMA_BIT_MASK(48));
 	if (err) {
 		dev_err(dev, "DMA mask config failed, abort\n");
-		goto err_release_regions;
+		goto err_set_drv_data;
 	}
 
 	pci_set_master(pdev);
@@ -780,7 +780,7 @@ static int rvu_rep_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
 	if (!priv) {
 		err = -ENOMEM;
-		goto err_release_regions;
+		goto err_set_drv_data;
 	}
 
 	pci_set_drvdata(pdev, priv);
@@ -797,7 +797,7 @@ static int rvu_rep_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 
 	err = otx2_init_rsrc(pdev, priv);
 	if (err)
-		goto err_release_regions;
+		goto err_set_drv_data;
 
 	priv->iommu_domain = iommu_get_domain_for_dev(dev);
 
@@ -820,9 +820,8 @@ static int rvu_rep_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	otx2_disable_mbox_intr(priv);
 	otx2_pfaf_mbox_destroy(priv);
 	pci_free_irq_vectors(pdev);
-err_release_regions:
+err_set_drv_data:
 	pci_set_drvdata(pdev, NULL);
-	pci_release_regions(pdev);
 	return err;
 }
 
@@ -842,7 +841,6 @@ static void rvu_rep_remove(struct pci_dev *pdev)
 	otx2_pfaf_mbox_destroy(priv);
 	pci_free_irq_vectors(priv->pdev);
 	pci_set_drvdata(pdev, NULL);
-	pci_release_regions(pdev);
 }
 
 static struct pci_driver rvu_rep_driver = {
-- 
2.48.1

[PATCH v2 3/8] net: tulip: Use pure PCI devres API

[Philipp Stanner]

The currently used function pci_request_regions() is one of the
problematic "hybrid devres" PCI functions, which are sometimes managed
through devres, and sometimes not (depending on whether
pci_enable_device() or pcim_enable_device() has been called before).

The PCI subsystem wants to remove this behavior and, therefore, needs to
port all users to functions that don't have this problem.

Replace pci_request_regions() with pcim_request_all_regions().

Signed-off-by: Philipp Stanner <phasta@kernel.org>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
---
 drivers/net/ethernet/dec/tulip/tulip_core.c  | 2 +-
 drivers/net/ethernet/dec/tulip/winbond-840.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/dec/tulip/tulip_core.c b/drivers/net/ethernet/dec/tulip/tulip_core.c
index c8c53121557f..bec76e7bf5dd 100644
--- a/drivers/net/ethernet/dec/tulip/tulip_core.c
+++ b/drivers/net/ethernet/dec/tulip/tulip_core.c
@@ -1411,7 +1411,7 @@ static int tulip_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 
 	/* grab all resources from both PIO and MMIO regions, as we
 	 * don't want anyone else messing around with our hardware */
-	if (pci_request_regions(pdev, DRV_NAME))
+	if (pcim_request_all_regions(pdev, DRV_NAME))
 		return -ENODEV;
 
 	ioaddr = pcim_iomap(pdev, TULIP_BAR, tulip_tbl[chip_idx].io_size);
diff --git a/drivers/net/ethernet/dec/tulip/winbond-840.c b/drivers/net/ethernet/dec/tulip/winbond-840.c
index 5930cdec6f2f..e593273b2867 100644
--- a/drivers/net/ethernet/dec/tulip/winbond-840.c
+++ b/drivers/net/ethernet/dec/tulip/winbond-840.c
@@ -375,7 +375,7 @@ static int w840_probe1(struct pci_dev *pdev, const struct pci_device_id *ent)
 		return -ENOMEM;
 	SET_NETDEV_DEV(dev, &pdev->dev);
 
-	if (pci_request_regions(pdev, DRV_NAME))
+	if (pcim_request_all_regions(pdev, DRV_NAME))
 		goto err_out_netdev;
 
 	ioaddr = pci_iomap(pdev, TULIP_BAR, netdev_res_size);
-- 
2.48.1

[PATCH v2 4/8] net: ethernet: natsemi: Use pure PCI devres API

[Philipp Stanner]

The currently used function pci_request_regions() is one of the
problematic "hybrid devres" PCI functions, which are sometimes managed
through devres, and sometimes not (depending on whether
pci_enable_device() or pcim_enable_device() has been called before).

The PCI subsystem wants to remove this behavior and, therefore, needs to
port all users to functions that don't have this problem.

Replace pci_request_regions() with pcim_request_all_regions().

Signed-off-by: Philipp Stanner <phasta@kernel.org>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
---
 drivers/net/ethernet/natsemi/natsemi.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/natsemi/natsemi.c b/drivers/net/ethernet/natsemi/natsemi.c
index 05606692e631..dd279788cf9e 100644
--- a/drivers/net/ethernet/natsemi/natsemi.c
+++ b/drivers/net/ethernet/natsemi/natsemi.c
@@ -846,7 +846,7 @@ static int natsemi_probe1(struct pci_dev *pdev, const struct pci_device_id *ent)
 		return -ENOMEM;
 	SET_NETDEV_DEV(dev, &pdev->dev);
 
-	i = pci_request_regions(pdev, DRV_NAME);
+	i = pcim_request_all_regions(pdev, DRV_NAME);
 	if (i)
 		goto err_pci_request_regions;
 
-- 
2.48.1

[PATCH v2 5/8] net: ethernet: sis900: Use pure PCI devres API

[Philipp Stanner]

The currently used function pci_request_regions() is one of the
problematic "hybrid devres" PCI functions, which are sometimes managed
through devres, and sometimes not (depending on whether
pci_enable_device() or pcim_enable_device() has been called before).

The PCI subsystem wants to remove this behavior and, therefore, needs to
port all users to functions that don't have this problem.

Replace pci_request_regions() with pcim_request_all_regions().

Signed-off-by: Philipp Stanner <phasta@kernel.org>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
---
 drivers/net/ethernet/sis/sis900.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/sis/sis900.c b/drivers/net/ethernet/sis/sis900.c
index 332cbd725900..df869f82cae8 100644
--- a/drivers/net/ethernet/sis/sis900.c
+++ b/drivers/net/ethernet/sis/sis900.c
@@ -468,7 +468,7 @@ static int sis900_probe(struct pci_dev *pci_dev,
 	SET_NETDEV_DEV(net_dev, &pci_dev->dev);
 
 	/* We do a request_region() to register /proc/ioports info. */
-	ret = pci_request_regions(pci_dev, "sis900");
+	ret = pcim_request_all_regions(pci_dev, "sis900");
 	if (ret)
 		goto err_out;
 
-- 
2.48.1

Re: [PATCH v2 5/8] net: ethernet: sis900: Use pure PCI devres API

[Daniele Venzano]

On 25/04/2025 10:57, Philipp Stanner wrote:
> The currently used function pci_request_regions() is one of the
> problematic "hybrid devres" PCI functions, which are sometimes managed
> through devres, and sometimes not (depending on whether
> pci_enable_device() or pcim_enable_device() has been called before).
>
> The PCI subsystem wants to remove this behavior and, therefore, needs to
> port all users to functions that don't have this problem.
>
> Replace pci_request_regions() with pcim_request_all_regions().
>
> Signed-off-by: Philipp Stanner <phasta@kernel.org>
> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Acked-by: Daniele Venzano <venza@brownhat.org>
> ---
>   drivers/net/ethernet/sis/sis900.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/net/ethernet/sis/sis900.c b/drivers/net/ethernet/sis/sis900.c
> index 332cbd725900..df869f82cae8 100644
> --- a/drivers/net/ethernet/sis/sis900.c
> +++ b/drivers/net/ethernet/sis/sis900.c
> @@ -468,7 +468,7 @@ static int sis900_probe(struct pci_dev *pci_dev,
>   	SET_NETDEV_DEV(net_dev, &pci_dev->dev);
>   
>   	/* We do a request_region() to register /proc/ioports info. */
> -	ret = pci_request_regions(pci_dev, "sis900");
> +	ret = pcim_request_all_regions(pci_dev, "sis900");
>   	if (ret)
>   		goto err_out;
>   

[PATCH v2 6/8] net: mdio: thunder: Use pure PCI devres API

[Philipp Stanner]

The currently used function pci_request_regions() is one of the
problematic "hybrid devres" PCI functions, which are sometimes managed
through devres, and sometimes not (depending on whether
pci_enable_device() or pcim_enable_device() has been called before).

The PCI subsystem wants to remove this behavior and, therefore, needs to
port all users to functions that don't have this problem.

Furthermore, the PCI function being managed implies that it's not
necessary to call pci_release_regions() manually.

Remove the calls to pci_release_regions().

Replace pci_request_regions() with pcim_request_all_regions().

Signed-off-by: Philipp Stanner <phasta@kernel.org>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
---
 drivers/net/mdio/mdio-thunder.c | 10 +++-------
 1 file changed, 3 insertions(+), 7 deletions(-)

diff --git a/drivers/net/mdio/mdio-thunder.c b/drivers/net/mdio/mdio-thunder.c
index 1e1aa72b1eff..a3047f7258a7 100644
--- a/drivers/net/mdio/mdio-thunder.c
+++ b/drivers/net/mdio/mdio-thunder.c
@@ -40,16 +40,16 @@ static int thunder_mdiobus_pci_probe(struct pci_dev *pdev,
 		return err;
 	}
 
-	err = pci_request_regions(pdev, KBUILD_MODNAME);
+	err = pcim_request_all_regions(pdev, KBUILD_MODNAME);
 	if (err) {
-		dev_err(&pdev->dev, "pci_request_regions failed\n");
+		dev_err(&pdev->dev, "pcim_request_all_regions failed\n");
 		goto err_disable_device;
 	}
 
 	nexus->bar0 = pcim_iomap(pdev, 0, pci_resource_len(pdev, 0));
 	if (!nexus->bar0) {
 		err = -ENOMEM;
-		goto err_release_regions;
+		goto err_disable_device;
 	}
 
 	i = 0;
@@ -107,9 +107,6 @@ static int thunder_mdiobus_pci_probe(struct pci_dev *pdev,
 	}
 	return 0;
 
-err_release_regions:
-	pci_release_regions(pdev);
-
 err_disable_device:
 	pci_set_drvdata(pdev, NULL);
 	return err;
@@ -129,7 +126,6 @@ static void thunder_mdiobus_pci_remove(struct pci_dev *pdev)
 		mdiobus_unregister(bus->mii_bus);
 		oct_mdio_writeq(0, bus->register_base + SMI_EN);
 	}
-	pci_release_regions(pdev);
 	pci_set_drvdata(pdev, NULL);
 }
 
-- 
2.48.1

[PATCH v2 7/8] net: thunder_bgx: Use pure PCI devres API

[Philipp Stanner]

The currently used function pci_request_regions() is one of the
problematic "hybrid devres" PCI functions, which are sometimes managed
through devres, and sometimes not (depending on whether
pci_enable_device() or pcim_enable_device() has been called before).

The PCI subsystem wants to remove this behavior and, therefore, needs to
port all users to functions that don't have this problem.

Furthermore, the PCI function being managed implies that it's not
necessary to call pci_release_regions() manually.

Remove the calls to pci_release_regions().

Replace pci_request_regions() with pcim_request_all_regions().

Signed-off-by: Philipp Stanner <phasta@kernel.org>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
---
 drivers/net/ethernet/cavium/thunder/thunder_bgx.c | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/cavium/thunder/thunder_bgx.c b/drivers/net/ethernet/cavium/thunder/thunder_bgx.c
index 608cc6af5af1..c9369bdd04e0 100644
--- a/drivers/net/ethernet/cavium/thunder/thunder_bgx.c
+++ b/drivers/net/ethernet/cavium/thunder/thunder_bgx.c
@@ -1605,7 +1605,7 @@ static int bgx_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 		return dev_err_probe(dev, err, "Failed to enable PCI device\n");
 	}
 
-	err = pci_request_regions(pdev, DRV_NAME);
+	err = pcim_request_all_regions(pdev, DRV_NAME);
 	if (err) {
 		dev_err(dev, "PCI request regions failed 0x%x\n", err);
 		goto err_disable_device;
@@ -1616,7 +1616,7 @@ static int bgx_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	if (!bgx->reg_base) {
 		dev_err(dev, "BGX: Cannot map CSR memory space, aborting\n");
 		err = -ENOMEM;
-		goto err_release_regions;
+		goto err_disable_device;
 	}
 
 	set_max_bgx_per_node(pdev);
@@ -1688,8 +1688,6 @@ static int bgx_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 err_enable:
 	bgx_vnic[bgx->bgx_id] = NULL;
 	pci_free_irq(pdev, GMPX_GMI_TX_INT, bgx);
-err_release_regions:
-	pci_release_regions(pdev);
 err_disable_device:
 	pci_disable_device(pdev);
 	pci_set_drvdata(pdev, NULL);
@@ -1710,7 +1708,6 @@ static void bgx_remove(struct pci_dev *pdev)
 	pci_free_irq(pdev, GMPX_GMI_TX_INT, bgx);
 
 	bgx_vnic[bgx->bgx_id] = NULL;
-	pci_release_regions(pdev);
 	pci_disable_device(pdev);
 	pci_set_drvdata(pdev, NULL);
 }
-- 
2.48.1

[PATCH v2 8/8] net: thunder_bgx: Don't disable PCI device manually

[Philipp Stanner]

thunder_bgx's PCI device is enabled with pcim_enable_device(), a managed
devres function which ensures that the device gets enabled on driver
detach automatically.

Remove the calls to pci_disable_device().

Signed-off-by: Philipp Stanner <phasta@kernel.org>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
---
 drivers/net/ethernet/cavium/thunder/thunder_bgx.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/cavium/thunder/thunder_bgx.c b/drivers/net/ethernet/cavium/thunder/thunder_bgx.c
index c9369bdd04e0..3b7ad744b2dd 100644
--- a/drivers/net/ethernet/cavium/thunder/thunder_bgx.c
+++ b/drivers/net/ethernet/cavium/thunder/thunder_bgx.c
@@ -1608,7 +1608,7 @@ static int bgx_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	err = pcim_request_all_regions(pdev, DRV_NAME);
 	if (err) {
 		dev_err(dev, "PCI request regions failed 0x%x\n", err);
-		goto err_disable_device;
+		goto err_zero_drv_data;
 	}
 
 	/* MAP configuration registers */
@@ -1616,7 +1616,7 @@ static int bgx_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	if (!bgx->reg_base) {
 		dev_err(dev, "BGX: Cannot map CSR memory space, aborting\n");
 		err = -ENOMEM;
-		goto err_disable_device;
+		goto err_zero_drv_data;
 	}
 
 	set_max_bgx_per_node(pdev);
@@ -1688,8 +1688,7 @@ static int bgx_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 err_enable:
 	bgx_vnic[bgx->bgx_id] = NULL;
 	pci_free_irq(pdev, GMPX_GMI_TX_INT, bgx);
-err_disable_device:
-	pci_disable_device(pdev);
+err_zero_drv_data:
 	pci_set_drvdata(pdev, NULL);
 	return err;
 }
@@ -1708,7 +1707,6 @@ static void bgx_remove(struct pci_dev *pdev)
 	pci_free_irq(pdev, GMPX_GMI_TX_INT, bgx);
 
 	bgx_vnic[bgx->bgx_id] = NULL;
-	pci_disable_device(pdev);
 	pci_set_drvdata(pdev, NULL);
 }
 
-- 
2.48.1

Re: [PATCH v2 0/8] Phase out hybrid PCI devres API

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net-next.git (main)
by Jakub Kicinski <kuba@kernel.org>:

On Fri, 25 Apr 2025 10:57:33 +0200 you wrote:
> Changes in v2:
>   - Rename error path. (Jakub)
>   - Apply some RBs
> 
> Fixes a number of minor issues with the usage of the PCI API in net.
> Notbaly, it replaces calls to the sometimes-managed
> pci_request_regions() to the always-managed pcim_request_all_regions(),
> enabling us to remove that hybrid functionality from PCI.
> Philipp Stanner (8):
>   net: prestera: Use pure PCI devres API
>   net: octeontx2: Use pure PCI devres API
>   net: tulip: Use pure PCI devres API
>   net: ethernet: natsemi: Use pure PCI devres API
>   net: ethernet: sis900: Use pure PCI devres API
>   net: mdio: thunder: Use pure PCI devres API
>   net: thunder_bgx: Use pure PCI devres API
>   net: thunder_bgx: Don't disable PCI device manually
> 
> [...]

Here is the summary with links:
  - [v2,1/8] net: prestera: Use pure PCI devres API
    https://git.kernel.org/netdev/net-next/c/66ada7471155
  - [v2,2/8] net: octeontx2: Use pure PCI devres API
    https://git.kernel.org/netdev/net-next/c/48217b834529
  - [v2,3/8] net: tulip: Use pure PCI devres API
    https://git.kernel.org/netdev/net-next/c/adc36d0914f6
  - [v2,4/8] net: ethernet: natsemi: Use pure PCI devres API
    https://git.kernel.org/netdev/net-next/c/2a5a74947a2b
  - [v2,5/8] net: ethernet: sis900: Use pure PCI devres API
    https://git.kernel.org/netdev/net-next/c/6e5f7a5b5e0c
  - [v2,6/8] net: mdio: thunder: Use pure PCI devres API
    https://git.kernel.org/netdev/net-next/c/fad4d94d9ae5
  - [v2,7/8] net: thunder_bgx: Use pure PCI devres API
    https://git.kernel.org/netdev/net-next/c/06133ddc3590
  - [v2,8/8] net: thunder_bgx: Don't disable PCI device manually
    https://git.kernel.org/netdev/net-next/c/1549bd06e340

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html