linux-pci.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v2] PCI: mediatek: Fix wrong operator used
@ 2020-11-03 23:10 Ryder Lee
  2020-11-04 13:10 ` Bjorn Helgaas
  0 siblings, 1 reply; 2+ messages in thread
From: Ryder Lee @ 2020-11-03 23:10 UTC (permalink / raw)
  To: Bjorn Helgaas; +Cc: linux-pci, linux-mediatek, Ryder Lee

SoCs like MT2701 and MT7623 use mtk_pcie_startup_port() to compute "func",
but from the code, the result we get is incorrect.

	#define PCI_FUNC(devfn)         ((devfn) & 0x07)

	func = PCI_FUNC(port->slot << 3)

The "func" is always 0, which results in other functions may not have been
configured correctly. (i.e. FC credits and FTS)

Addresses-Coverity-ID: 1437218 ("Wrong operator used")
Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
---
V2 - revise commit log
---
 drivers/pci/controller/pcie-mediatek.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/pci/controller/pcie-mediatek.c b/drivers/pci/controller/pcie-mediatek.c
index cf4c18f0c25a..1980b01cee35 100644
--- a/drivers/pci/controller/pcie-mediatek.c
+++ b/drivers/pci/controller/pcie-mediatek.c
@@ -760,7 +760,7 @@ static struct pci_ops mtk_pcie_ops = {
 static int mtk_pcie_startup_port(struct mtk_pcie_port *port)
 {
 	struct mtk_pcie *pcie = port->pcie;
-	u32 func = PCI_FUNC(port->slot << 3);
+	u32 func = PCI_FUNC(port->slot);
 	u32 slot = PCI_SLOT(port->slot << 3);
 	u32 val;
 	int err;
-- 
2.18.0

^ permalink raw reply related	[flat|nested] 2+ messages in thread

* Re: [PATCH v2] PCI: mediatek: Fix wrong operator used
  2020-11-03 23:10 [PATCH v2] PCI: mediatek: Fix wrong operator used Ryder Lee
@ 2020-11-04 13:10 ` Bjorn Helgaas
  0 siblings, 0 replies; 2+ messages in thread
From: Bjorn Helgaas @ 2020-11-04 13:10 UTC (permalink / raw)
  To: Ryder Lee; +Cc: Bjorn Helgaas, linux-pci, linux-mediatek

The subject needs to say something about the user-visible problem
you're fixing, e.g., functions other than 0 aren't configured
correctly.

On Wed, Nov 04, 2020 at 07:10:11AM +0800, Ryder Lee wrote:
> SoCs like MT2701 and MT7623 use mtk_pcie_startup_port() to compute "func",
> but from the code, the result we get is incorrect.

This affects precisely "MT2701 and MT7623", not "SoCs *like* MT2701
and MT7623".  How about this:

  PCI: mediatek: Configure FC and FTS for functions other than 0

  "PCI_FUNC(port->slot << 3)" is always 0, so previously
  mtk_pcie_startup_port() only configured FC credits and FTs for
  function 0.

  Compute "func" correctly so we also configure functions other than
  0.  This affects MT2701 and MT7623.

> 	#define PCI_FUNC(devfn)         ((devfn) & 0x07)
> 
> 	func = PCI_FUNC(port->slot << 3)
> 
> The "func" is always 0, which results in other functions may not have been
> configured correctly. (i.e. FC credits and FTS)
> 
> Addresses-Coverity-ID: 1437218 ("Wrong operator used")
> Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
> ---
> V2 - revise commit log
> ---
>  drivers/pci/controller/pcie-mediatek.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/pci/controller/pcie-mediatek.c b/drivers/pci/controller/pcie-mediatek.c
> index cf4c18f0c25a..1980b01cee35 100644
> --- a/drivers/pci/controller/pcie-mediatek.c
> +++ b/drivers/pci/controller/pcie-mediatek.c
> @@ -760,7 +760,7 @@ static struct pci_ops mtk_pcie_ops = {
>  static int mtk_pcie_startup_port(struct mtk_pcie_port *port)
>  {
>  	struct mtk_pcie *pcie = port->pcie;
> -	u32 func = PCI_FUNC(port->slot << 3);
> +	u32 func = PCI_FUNC(port->slot);
>  	u32 slot = PCI_SLOT(port->slot << 3);
>  	u32 val;
>  	int err;
> -- 
> 2.18.0

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2020-11-04 13:10 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-11-03 23:10 [PATCH v2] PCI: mediatek: Fix wrong operator used Ryder Lee
2020-11-04 13:10 ` Bjorn Helgaas

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).