diff mbox series

[v2] ath9k: use unmanaged PCI functions in ath9k_pci_owl_loader

Message ID 3b46f6c7-4372-4cc9-9a7c-2c1c06d29324@gmail.com
State New
Headers show
Series [v2] ath9k: use unmanaged PCI functions in ath9k_pci_owl_loader | expand

Commit Message

Heiner Kallweit April 9, 2024, 8:28 p.m. UTC
Only managed PCI resource in the driver is the iomapped bar. However the bar
is unmapped in the same function. Therefore using the device-managed
versions just causes overhead, w/o any benefit. Once this is switched to the
non-managed versions, there's nothing left to be managed for
pcim_enable_device(). Therefore we can reduce overhead here too and switch to
the non-managed version as well. This includes removing the no longer needed
call to pcim_pin_device().

Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
---
v2:
- extended commit message
---
 drivers/net/wireless/ath/ath9k/ath9k_pci_owl_loader.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

Comments

Toke Høiland-Jørgensen July 30, 2024, 9:59 a.m. UTC | #1
Heiner Kallweit <hkallweit1@gmail.com> writes:

> Only managed PCI resource in the driver is the iomapped bar. However the bar
> is unmapped in the same function. Therefore using the device-managed
> versions just causes overhead, w/o any benefit. Once this is switched to the
> non-managed versions, there's nothing left to be managed for
> pcim_enable_device(). Therefore we can reduce overhead here too and switch to
> the non-managed version as well. This includes removing the no longer needed
> call to pcim_pin_device().
>
> Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>

Acked-by: Toke Høiland-Jørgensen <toke@toke.dk>
Kalle Valo Aug. 1, 2024, 2:58 p.m. UTC | #2
Heiner Kallweit <hkallweit1@gmail.com> wrote:

> Only managed PCI resource in the driver is the iomapped bar. However the bar
> is unmapped in the same function. Therefore using the device-managed
> versions just causes overhead, w/o any benefit. Once this is switched to the
> non-managed versions, there's nothing left to be managed for
> pcim_enable_device(). Therefore we can reduce overhead here too and switch to
> the non-managed version as well. This includes removing the no longer needed
> call to pcim_pin_device().
> 
> Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
> Acked-by: Toke Høiland-Jørgensen <toke@toke.dk>
> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>

Patch applied to ath-next branch of ath.git, thanks.

aa0d7643c8dd wifi: ath9k: use unmanaged PCI functions in ath9k_pci_owl_loader
diff mbox series

Patch

diff --git a/drivers/net/wireless/ath/ath9k/ath9k_pci_owl_loader.c b/drivers/net/wireless/ath/ath9k/ath9k_pci_owl_loader.c
index a5eb43f30..bc93ca075 100644
--- a/drivers/net/wireless/ath/ath9k/ath9k_pci_owl_loader.c
+++ b/drivers/net/wireless/ath/ath9k/ath9k_pci_owl_loader.c
@@ -65,7 +65,7 @@  static int ath9k_pci_fixup(struct pci_dev *pdev, const u16 *cal_data,
 
 	dev_info(&pdev->dev, "fixup device configuration\n");
 
-	mem = pcim_iomap(pdev, 0, 0);
+	mem = pci_iomap(pdev, 0, 0);
 	if (!mem) {
 		dev_err(&pdev->dev, "ioremap error\n");
 		return -EINVAL;
@@ -103,7 +103,7 @@  static int ath9k_pci_fixup(struct pci_dev *pdev, const u16 *cal_data,
 	pci_write_config_word(pdev, PCI_COMMAND, cmd);
 
 	pci_write_config_dword(pdev, PCI_BASE_ADDRESS_0, bar0);
-	pcim_iounmap(pdev, mem);
+	pci_iounmap(pdev, mem);
 
 	pci_disable_device(pdev);
 
@@ -200,11 +200,9 @@  static int owl_probe(struct pci_dev *pdev,
 	const char *eeprom_name;
 	int err = 0;
 
-	if (pcim_enable_device(pdev))
+	if (pci_enable_device(pdev))
 		return -EIO;
 
-	pcim_pin_device(pdev);
-
 	ctx = devm_kzalloc(&pdev->dev, sizeof(*ctx), GFP_KERNEL);
 	if (!ctx)
 		return -ENOMEM;