Re: [PATCH] PCI: rockchip: fix uninitialized variable use

[Bjorn Helgaas]

On Thu, Sep 22, 2016 at 11:39:59AM +0200, Arnd Bergmann wrote:
> The newly added pcie-rockchip driver fails to initialize the
> io_size variable if the DT doesn't provide ranges for the PCI
> I/O space, as found by building it with -Wmaybe-uninitialized:
> 
> drivers/pci/host/pcie-rockchip.c: In function 'rockchip_pcie_probe':
> drivers/pci/host/pcie-rockchip.c:1007:6: warning: 'io_size' may be used 
> uninitialized in this function [-Wmaybe-uninitialized]
> 
> This adds an appropriate initialization immediately in front of
> the loop, so the io_size is zero as expected afterwards for that
> case.
> 
> Fixes: abe17181b16f ("PCI: rockchip: Add Rockchip PCIe controller support")
> Signed-off-by: Arnd Bergmann 

Thanks, I folded this into the original commit and added a note that it's
from you.

> ---
>  drivers/pci/host/pcie-rockchip.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/pci/host/pcie-rockchip.c 
> b/drivers/pci/host/pcie-rockchip.c
> index c3593e633ccd..8bedc1e1ef80 100644
> --- a/drivers/pci/host/pcie-rockchip.c
> +++ b/drivers/pci/host/pcie-rockchip.c
> @@ -1078,6 +1078,7 @@ static int rockchip_pcie_probe(struct platform_device 
> *pdev)
>   goto err_vpcie;
>  
>   /* Get the I/O and memory ranges from DT */
> + io_size = 0;
>   resource_list_for_each_entry(win, ) {
>   switch (resource_type(win->res)) {
>   case IORESOURCE_IO:
> -- 
> 2.9.0
> 
> 
> ___
> linux-arm-kernel mailing list
> linux-arm-ker...@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH] PCI: rockchip: fix uninitialized variable use

[Bjorn Helgaas]

On Thu, Sep 22, 2016 at 11:39:59AM +0200, Arnd Bergmann wrote:
> The newly added pcie-rockchip driver fails to initialize the
> io_size variable if the DT doesn't provide ranges for the PCI
> I/O space, as found by building it with -Wmaybe-uninitialized:
> 
> drivers/pci/host/pcie-rockchip.c: In function 'rockchip_pcie_probe':
> drivers/pci/host/pcie-rockchip.c:1007:6: warning: 'io_size' may be used 
> uninitialized in this function [-Wmaybe-uninitialized]
> 
> This adds an appropriate initialization immediately in front of
> the loop, so the io_size is zero as expected afterwards for that
> case.
> 
> Fixes: abe17181b16f ("PCI: rockchip: Add Rockchip PCIe controller support")
> Signed-off-by: Arnd Bergmann 

Thanks, I folded this into the original commit and added a note that it's
from you.

> ---
>  drivers/pci/host/pcie-rockchip.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/pci/host/pcie-rockchip.c 
> b/drivers/pci/host/pcie-rockchip.c
> index c3593e633ccd..8bedc1e1ef80 100644
> --- a/drivers/pci/host/pcie-rockchip.c
> +++ b/drivers/pci/host/pcie-rockchip.c
> @@ -1078,6 +1078,7 @@ static int rockchip_pcie_probe(struct platform_device 
> *pdev)
>   goto err_vpcie;
>  
>   /* Get the I/O and memory ranges from DT */
> + io_size = 0;
>   resource_list_for_each_entry(win, ) {
>   switch (resource_type(win->res)) {
>   case IORESOURCE_IO:
> -- 
> 2.9.0
> 
> 
> ___
> linux-arm-kernel mailing list
> linux-arm-ker...@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH] PCI: rockchip: fix uninitialized variable use

[Shawn Lin]

Hi Arnd,

在 2016/9/22 17:39, Arnd Bergmann 写道:

The newly added pcie-rockchip driver fails to initialize the
io_size variable if the DT doesn't provide ranges for the PCI
I/O space, as found by building it with -Wmaybe-uninitialized:

drivers/pci/host/pcie-rockchip.c: In function 'rockchip_pcie_probe':
drivers/pci/host/pcie-rockchip.c:1007:6: warning: 'io_size' may be used 
uninitialized in this function [-Wmaybe-uninitialized]



Seems like we miss this when refactoring the code a bit.
Thanks for fixing it.

Acked-by: Shawn Lin 


This adds an appropriate initialization immediately in front of
the loop, so the io_size is zero as expected afterwards for that
case.

Fixes: abe17181b16f ("PCI: rockchip: Add Rockchip PCIe controller support")
Signed-off-by: Arnd Bergmann 
---
 drivers/pci/host/pcie-rockchip.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/pci/host/pcie-rockchip.c b/drivers/pci/host/pcie-rockchip.c
index c3593e633ccd..8bedc1e1ef80 100644
--- a/drivers/pci/host/pcie-rockchip.c
+++ b/drivers/pci/host/pcie-rockchip.c
@@ -1078,6 +1078,7 @@ static int rockchip_pcie_probe(struct platform_device 
*pdev)
goto err_vpcie;

/* Get the I/O and memory ranges from DT */
+   io_size = 0;
resource_list_for_each_entry(win, ) {
switch (resource_type(win->res)) {
case IORESOURCE_IO:




--
Best Regards
Shawn Lin

Re: [PATCH] PCI: rockchip: fix uninitialized variable use

[Shawn Lin]

Hi Arnd,

在 2016/9/22 17:39, Arnd Bergmann 写道:

The newly added pcie-rockchip driver fails to initialize the
io_size variable if the DT doesn't provide ranges for the PCI
I/O space, as found by building it with -Wmaybe-uninitialized:

drivers/pci/host/pcie-rockchip.c: In function 'rockchip_pcie_probe':
drivers/pci/host/pcie-rockchip.c:1007:6: warning: 'io_size' may be used 
uninitialized in this function [-Wmaybe-uninitialized]



Seems like we miss this when refactoring the code a bit.
Thanks for fixing it.

Acked-by: Shawn Lin 


This adds an appropriate initialization immediately in front of
the loop, so the io_size is zero as expected afterwards for that
case.

Fixes: abe17181b16f ("PCI: rockchip: Add Rockchip PCIe controller support")
Signed-off-by: Arnd Bergmann 
---
 drivers/pci/host/pcie-rockchip.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/pci/host/pcie-rockchip.c b/drivers/pci/host/pcie-rockchip.c
index c3593e633ccd..8bedc1e1ef80 100644
--- a/drivers/pci/host/pcie-rockchip.c
+++ b/drivers/pci/host/pcie-rockchip.c
@@ -1078,6 +1078,7 @@ static int rockchip_pcie_probe(struct platform_device 
*pdev)
goto err_vpcie;

/* Get the I/O and memory ranges from DT */
+   io_size = 0;
resource_list_for_each_entry(win, ) {
switch (resource_type(win->res)) {
case IORESOURCE_IO:




--
Best Regards
Shawn Lin

Re: [PATCH] PCI: rockchip: fix uninitialized variable use

[Brian Norris]

On Thu, Sep 22, 2016 at 11:39:59AM +0200, Arnd Bergmann wrote:
> The newly added pcie-rockchip driver fails to initialize the
> io_size variable if the DT doesn't provide ranges for the PCI
> I/O space, as found by building it with -Wmaybe-uninitialized:
> 
> drivers/pci/host/pcie-rockchip.c: In function 'rockchip_pcie_probe':
> drivers/pci/host/pcie-rockchip.c:1007:6: warning: 'io_size' may be used 
> uninitialized in this function [-Wmaybe-uninitialized]
> 
> This adds an appropriate initialization immediately in front of
> the loop, so the io_size is zero as expected afterwards for that
> case.
> 
> Fixes: abe17181b16f ("PCI: rockchip: Add Rockchip PCIe controller support")
> Signed-off-by: Arnd Bergmann 
> ---
>  drivers/pci/host/pcie-rockchip.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/pci/host/pcie-rockchip.c 
> b/drivers/pci/host/pcie-rockchip.c
> index c3593e633ccd..8bedc1e1ef80 100644
> --- a/drivers/pci/host/pcie-rockchip.c
> +++ b/drivers/pci/host/pcie-rockchip.c
> @@ -1078,6 +1078,7 @@ static int rockchip_pcie_probe(struct platform_device 
> *pdev)
>   goto err_vpcie;
>  
>   /* Get the I/O and memory ranges from DT */
> + io_size = 0;
>   resource_list_for_each_entry(win, ) {
>   switch (resource_type(win->res)) {
>   case IORESOURCE_IO:

I feel like we've fixed this one before, but then it's been refactored
many times along the way...

Reviewed-by: Brian Norris 

Re: [PATCH] PCI: rockchip: fix uninitialized variable use

[Brian Norris]

On Thu, Sep 22, 2016 at 11:39:59AM +0200, Arnd Bergmann wrote:
> The newly added pcie-rockchip driver fails to initialize the
> io_size variable if the DT doesn't provide ranges for the PCI
> I/O space, as found by building it with -Wmaybe-uninitialized:
> 
> drivers/pci/host/pcie-rockchip.c: In function 'rockchip_pcie_probe':
> drivers/pci/host/pcie-rockchip.c:1007:6: warning: 'io_size' may be used 
> uninitialized in this function [-Wmaybe-uninitialized]
> 
> This adds an appropriate initialization immediately in front of
> the loop, so the io_size is zero as expected afterwards for that
> case.
> 
> Fixes: abe17181b16f ("PCI: rockchip: Add Rockchip PCIe controller support")
> Signed-off-by: Arnd Bergmann 
> ---
>  drivers/pci/host/pcie-rockchip.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/pci/host/pcie-rockchip.c 
> b/drivers/pci/host/pcie-rockchip.c
> index c3593e633ccd..8bedc1e1ef80 100644
> --- a/drivers/pci/host/pcie-rockchip.c
> +++ b/drivers/pci/host/pcie-rockchip.c
> @@ -1078,6 +1078,7 @@ static int rockchip_pcie_probe(struct platform_device 
> *pdev)
>   goto err_vpcie;
>  
>   /* Get the I/O and memory ranges from DT */
> + io_size = 0;
>   resource_list_for_each_entry(win, ) {
>   switch (resource_type(win->res)) {
>   case IORESOURCE_IO:

I feel like we've fixed this one before, but then it's been refactored
many times along the way...

Reviewed-by: Brian Norris