[PATCH] ntb: reduce stack usage in idt_scan_mws

Arnd Bergmann posted 1 patch 9 months, 3 weeks ago
drivers/ntb/hw/idt/ntb_hw_idt.c | 18 +++++++-----------
1 file changed, 7 insertions(+), 11 deletions(-)
[PATCH] ntb: reduce stack usage in idt_scan_mws
Posted by Arnd Bergmann 9 months, 3 weeks ago
From: Arnd Bergmann <arnd@arndb.de>

idt_scan_mws() puts a large fixed-size array on the stack and copies
it into a smaller dynamically allocated array at the end. On 32-bit
targets, the fixed size can easily exceed the warning limit for
possible stack overflow:

drivers/ntb/hw/idt/ntb_hw_idt.c:1041:27: error: stack frame size (1032) exceeds limit (1024) in 'idt_scan_mws' [-Werror,-Wframe-larger-than]

Change it to instead just always use dynamic allocation for the
array from the start. It's too big for the stack, but not actually
all that much for a permanent allocation.

Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/all/202205111109.PiKTruEj-lkp@intel.com/
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
A number of users have reported this in the past, but I couldn't
find any other patch for it so far.
---
 drivers/ntb/hw/idt/ntb_hw_idt.c | 18 +++++++-----------
 1 file changed, 7 insertions(+), 11 deletions(-)

diff --git a/drivers/ntb/hw/idt/ntb_hw_idt.c b/drivers/ntb/hw/idt/ntb_hw_idt.c
index 544d8a4d2af5..f27df8d7f3b9 100644
--- a/drivers/ntb/hw/idt/ntb_hw_idt.c
+++ b/drivers/ntb/hw/idt/ntb_hw_idt.c
@@ -1041,7 +1041,7 @@ static inline char *idt_get_mw_name(enum idt_mw_type mw_type)
 static struct idt_mw_cfg *idt_scan_mws(struct idt_ntb_dev *ndev, int port,
 				       unsigned char *mw_cnt)
 {
-	struct idt_mw_cfg mws[IDT_MAX_NR_MWS], *ret_mws;
+	struct idt_mw_cfg *mws;
 	const struct idt_ntb_bar *bars;
 	enum idt_mw_type mw_type;
 	unsigned char widx, bidx, en_cnt;
@@ -1049,6 +1049,11 @@ static struct idt_mw_cfg *idt_scan_mws(struct idt_ntb_dev *ndev, int port,
 	int aprt_size;
 	u32 data;
 
+	mws = devm_kcalloc(&ndev->ntb.pdev->dev, IDT_MAX_NR_MWS,
+			   sizeof(*mws), GFP_KERNEL);
+	if (!mws)
+		return ERR_PTR(-ENOMEM);
+
 	/* Retrieve the array of the BARs registers */
 	bars = portdata_tbl[port].bars;
 
@@ -1103,16 +1108,7 @@ static struct idt_mw_cfg *idt_scan_mws(struct idt_ntb_dev *ndev, int port,
 		}
 	}
 
-	/* Allocate memory for memory window descriptors */
-	ret_mws = devm_kcalloc(&ndev->ntb.pdev->dev, *mw_cnt, sizeof(*ret_mws),
-			       GFP_KERNEL);
-	if (!ret_mws)
-		return ERR_PTR(-ENOMEM);
-
-	/* Copy the info of detected memory windows */
-	memcpy(ret_mws, mws, (*mw_cnt)*sizeof(*ret_mws));
-
-	return ret_mws;
+	return mws;
 }
 
 /*
-- 
2.39.5
Re: [PATCH] ntb: reduce stack usage in idt_scan_mws
Posted by Dave Jiang 9 months, 3 weeks ago

On 2/21/25 1:57 AM, Arnd Bergmann wrote:
> From: Arnd Bergmann <arnd@arndb.de>
> 
> idt_scan_mws() puts a large fixed-size array on the stack and copies
> it into a smaller dynamically allocated array at the end. On 32-bit
> targets, the fixed size can easily exceed the warning limit for
> possible stack overflow:
> 
> drivers/ntb/hw/idt/ntb_hw_idt.c:1041:27: error: stack frame size (1032) exceeds limit (1024) in 'idt_scan_mws' [-Werror,-Wframe-larger-than]
> 
> Change it to instead just always use dynamic allocation for the
> array from the start. It's too big for the stack, but not actually
> all that much for a permanent allocation.
> 
> Reported-by: kernel test robot <lkp@intel.com>
> Closes: https://lore.kernel.org/all/202205111109.PiKTruEj-lkp@intel.com/
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>

LGTM. Old code didn't make sense to declare on stack, allocate later and memcpy.
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
> ---
> A number of users have reported this in the past, but I couldn't
> find any other patch for it so far.
> ---
>  drivers/ntb/hw/idt/ntb_hw_idt.c | 18 +++++++-----------
>  1 file changed, 7 insertions(+), 11 deletions(-)
> 
> diff --git a/drivers/ntb/hw/idt/ntb_hw_idt.c b/drivers/ntb/hw/idt/ntb_hw_idt.c
> index 544d8a4d2af5..f27df8d7f3b9 100644
> --- a/drivers/ntb/hw/idt/ntb_hw_idt.c
> +++ b/drivers/ntb/hw/idt/ntb_hw_idt.c
> @@ -1041,7 +1041,7 @@ static inline char *idt_get_mw_name(enum idt_mw_type mw_type)
>  static struct idt_mw_cfg *idt_scan_mws(struct idt_ntb_dev *ndev, int port,
>  				       unsigned char *mw_cnt)
>  {
> -	struct idt_mw_cfg mws[IDT_MAX_NR_MWS], *ret_mws;
> +	struct idt_mw_cfg *mws;
>  	const struct idt_ntb_bar *bars;
>  	enum idt_mw_type mw_type;
>  	unsigned char widx, bidx, en_cnt;
> @@ -1049,6 +1049,11 @@ static struct idt_mw_cfg *idt_scan_mws(struct idt_ntb_dev *ndev, int port,
>  	int aprt_size;
>  	u32 data;
>  
> +	mws = devm_kcalloc(&ndev->ntb.pdev->dev, IDT_MAX_NR_MWS,
> +			   sizeof(*mws), GFP_KERNEL);
> +	if (!mws)
> +		return ERR_PTR(-ENOMEM);
> +
>  	/* Retrieve the array of the BARs registers */
>  	bars = portdata_tbl[port].bars;
>  
> @@ -1103,16 +1108,7 @@ static struct idt_mw_cfg *idt_scan_mws(struct idt_ntb_dev *ndev, int port,
>  		}
>  	}
>  
> -	/* Allocate memory for memory window descriptors */
> -	ret_mws = devm_kcalloc(&ndev->ntb.pdev->dev, *mw_cnt, sizeof(*ret_mws),
> -			       GFP_KERNEL);
> -	if (!ret_mws)
> -		return ERR_PTR(-ENOMEM);
> -
> -	/* Copy the info of detected memory windows */
> -	memcpy(ret_mws, mws, (*mw_cnt)*sizeof(*ret_mws));
> -
> -	return ret_mws;
> +	return mws;
>  }
>  
>  /*
Re: [PATCH] ntb: reduce stack usage in idt_scan_mws
Posted by Damien Le Moal 9 months, 3 weeks ago
On 2/21/25 17:57, Arnd Bergmann wrote:
> From: Arnd Bergmann <arnd@arndb.de>
> 
> idt_scan_mws() puts a large fixed-size array on the stack and copies
> it into a smaller dynamically allocated array at the end. On 32-bit
> targets, the fixed size can easily exceed the warning limit for
> possible stack overflow:
> 
> drivers/ntb/hw/idt/ntb_hw_idt.c:1041:27: error: stack frame size (1032) exceeds limit (1024) in 'idt_scan_mws' [-Werror,-Wframe-larger-than]
> 
> Change it to instead just always use dynamic allocation for the
> array from the start. It's too big for the stack, but not actually
> all that much for a permanent allocation.
> 
> Reported-by: kernel test robot <lkp@intel.com>
> Closes: https://lore.kernel.org/all/202205111109.PiKTruEj-lkp@intel.com/
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>

Looks good to me.

Reviewed-by: Damien Le Moal <dlemoal@kernel.org>

-- 
Damien Le Moal
Western Digital Research