lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <95ce8d16-17bf-490e-a084-ac16fe8ba813@gmail.com>
Date: Fri, 30 Aug 2024 16:59:35 -0700
From: Florian Fainelli <f.fainelli@...il.com>
To: ende.tan@...rfivetech.com, netdev@...r.kernel.org
Cc: andrew@...n.ch, alexandre.torgue@...s.st.com, joabreu@...opsys.com,
 davem@...emloft.net, edumazet@...gle.com, kuba@...nel.org,
 pabeni@...hat.com, mcoquelin.stm32@...il.com, linux-kernel@...r.kernel.org,
 linux-stm32@...md-mailman.stormreply.com,
 linux-arm-kernel@...ts.infradead.org, leyfoon.tan@...rfivetech.com,
 minda.chen@...rfivetech.com, endeneer@...il.com
Subject: Re: [net-next,v3,1/1] net: stmmac: Batch set RX OWN flag and other
 flags

On 8/29/24 06:40, ende.tan@...rfivetech.com wrote:
> From: Tan En De <ende.tan@...rfivetech.com>
> 
> Minimize access to the RX descriptor by collecting all the flags in a
> local variable and then updating the descriptor at once.
> 
> Signed-off-by: Tan En De <ende.tan@...rfivetech.com>
> ---
> v3:
> - Use local variable to batch set the descriptor flags.
> - This reduces the number of accesses to the descriptor.
> v2: https://patchwork.kernel.org/project/netdevbpf/patch/20240821060307.46350-1-ende.tan@starfivetech.com/
> - Avoid introducing a new function just to set the interrupt-on-completion
>    bit, as it is wasteful to do so.
> - Delegate the responsibility of calling dma_wmb() from main driver code
>    to set_rx_owner() callbacks (i.e. let callbacks to manage the low-level
>    ordering/barrier rather than cluttering up the main driver code).
> v1: https://patchwork.kernel.org/project/netdevbpf/patch/20240814092438.3129-1-ende.tan@starfivetech.com/
> ---
>   drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c   | 6 ++++--
>   drivers/net/ethernet/stmicro/stmmac/dwxgmac2_descs.c | 6 ++++--
>   2 files changed, 8 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c b/drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c
> index 1c5802e0d7f4..dfcbe7036988 100644
> --- a/drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c
> +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c
> @@ -186,10 +186,12 @@ static void dwmac4_set_tx_owner(struct dma_desc *p)
>   
>   static void dwmac4_set_rx_owner(struct dma_desc *p, int disable_rx_ic)
>   {
> -	p->des3 |= cpu_to_le32(RDES3_OWN | RDES3_BUFFER1_VALID_ADDR);
> +	u32 flags = cpu_to_le32(RDES3_OWN | RDES3_BUFFER1_VALID_ADDR);
>   
>   	if (!disable_rx_ic)
> -		p->des3 |= cpu_to_le32(RDES3_INT_ON_COMPLETION_EN);
> +		flags |= cpu_to_le32(RDES3_INT_ON_COMPLETION_EN);
> +
> +	p->des3 |= flags;

You could just batch the endian conversion too:

	u32 flags = DES3_OWN | RDES3_BUFFER1_VALID_ADDR;

	if (!disable_rx_ic)
		flags |= RDES3_INT_ON_COMPLETION_EN;

	p->desc3 |= cpu_to_le32(flags);

>   }
>   
>   static int dwmac4_get_tx_ls(struct dma_desc *p)
> diff --git a/drivers/net/ethernet/stmicro/stmmac/dwxgmac2_descs.c b/drivers/net/ethernet/stmicro/stmmac/dwxgmac2_descs.c
> index fc82862a612c..0c7ea939f787 100644
> --- a/drivers/net/ethernet/stmicro/stmmac/dwxgmac2_descs.c
> +++ b/drivers/net/ethernet/stmicro/stmmac/dwxgmac2_descs.c
> @@ -56,10 +56,12 @@ static void dwxgmac2_set_tx_owner(struct dma_desc *p)
>   
>   static void dwxgmac2_set_rx_owner(struct dma_desc *p, int disable_rx_ic)
>   {
> -	p->des3 |= cpu_to_le32(XGMAC_RDES3_OWN);
> +	u32 flags = cpu_to_le32(XGMAC_RDES3_OWN);
>   
>   	if (!disable_rx_ic)
> -		p->des3 |= cpu_to_le32(XGMAC_RDES3_IOC);
> +		 flags |= cpu_to_le32(XGMAC_RDES3_IOC);
> +
> +	p->des3 |= flags;

And likewise here, and that would likely squash the sparse warning, 
since you would not be assigning an 'u32' field to a __le32 field anymore.
-- 
Florian

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ