From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Date: Mon, 20 Jan 2014 19:16:14 +0000 Subject: Re: [PATCH v5] can: add Renesas R-Car CAN driver Message-Id: <20140120.111614.394228612374217980.davem@davemloft.net> List-Id: References: <52DD0CC1.70205@pengutronix.de> <52DD0F72.1000005@pengutronix.de> In-Reply-To: <52DD0F72.1000005@pengutronix.de> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: mkl@pengutronix.de Cc: geert@linux-m68k.org, sergei.shtylyov@cogentembedded.com, netdev@vger.kernel.org, wg@grandegger.com, linux-can@vger.kernel.org, linux-sh@vger.kernel.org, vksavl@gmail.com From: Marc Kleine-Budde Date: Mon, 20 Jan 2014 12:58:42 +0100 > On 01/20/2014 12:52 PM, Geert Uytterhoeven wrote: >> On Mon, Jan 20, 2014 at 12:47 PM, Marc Kleine-Budde wrote: >>> On 01/20/2014 12:43 PM, Geert Uytterhoeven wrote: >>>> On Thu, Dec 26, 2013 at 10:37 PM, Sergei Shtylyov >>>> wrote: >>>>> Changes in version 3: >>>> >>>>> - added '__packed' to 'struct rcar_can_mbox_regs' and 'struct rcar_can_regs'; >>>>> - removed unneeded type cast in the probe() method. >>>> >>>>> +/* Mailbox registers structure */ >>>>> +struct rcar_can_mbox_regs { >>>>> + u32 id; /* IDE and RTR bits, SID and EID */ >>>>> + u8 stub; /* Not used */ >>>>> + u8 dlc; /* Data Length Code - bits [0..3] */ >>>>> + u8 data[8]; /* Data Bytes */ >>>>> + u8 tsh; /* Time Stamp Higher Byte */ >>>>> + u8 tsl; /* Time Stamp Lower Byte */ >>>>> +} __packed; >>>> >>>> Sorry, I missed the earlier discussion, but why the _packed? >>>> One u32 and 12 bytes makes a nice multiple of 4. >>> >>> Better safe than sorry, it's the layout of the registers in hardware, >>> don't let the compiler optimize here. >> >> Actually __packed makes it less safe, as the compiler now assumes >> the u32 id member is unaligned, and thus may turn 32-bit accesses into 4 >> byte accesses. >> >> Fortunately it won't happen in this case as the code uses writel/readl to >> acces the id member. > > Yes, as this are registers they must not be accessed directly. However > we can use "__attribute__ ((packed, aligned(4)))" to tell the compiler > that the base address of this struct is always aligned to 4 bytes. I truly think using packed here is rediculous, please remove it unless you can prove that things won't work without it. Thanks. From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [PATCH v5] can: add Renesas R-Car CAN driver Date: Mon, 20 Jan 2014 11:16:14 -0800 (PST) Message-ID: <20140120.111614.394228612374217980.davem@davemloft.net> References: <52DD0CC1.70205@pengutronix.de> <52DD0F72.1000005@pengutronix.de> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <52DD0F72.1000005@pengutronix.de> Sender: linux-sh-owner@vger.kernel.org To: mkl@pengutronix.de Cc: geert@linux-m68k.org, sergei.shtylyov@cogentembedded.com, netdev@vger.kernel.org, wg@grandegger.com, linux-can@vger.kernel.org, linux-sh@vger.kernel.org, vksavl@gmail.com List-Id: linux-can.vger.kernel.org From: Marc Kleine-Budde Date: Mon, 20 Jan 2014 12:58:42 +0100 > On 01/20/2014 12:52 PM, Geert Uytterhoeven wrote: >> On Mon, Jan 20, 2014 at 12:47 PM, Marc Kleine-Budde wrote: >>> On 01/20/2014 12:43 PM, Geert Uytterhoeven wrote: >>>> On Thu, Dec 26, 2013 at 10:37 PM, Sergei Shtylyov >>>> wrote: >>>>> Changes in version 3: >>>> >>>>> - added '__packed' to 'struct rcar_can_mbox_regs' and 'struct rcar_can_regs'; >>>>> - removed unneeded type cast in the probe() method. >>>> >>>>> +/* Mailbox registers structure */ >>>>> +struct rcar_can_mbox_regs { >>>>> + u32 id; /* IDE and RTR bits, SID and EID */ >>>>> + u8 stub; /* Not used */ >>>>> + u8 dlc; /* Data Length Code - bits [0..3] */ >>>>> + u8 data[8]; /* Data Bytes */ >>>>> + u8 tsh; /* Time Stamp Higher Byte */ >>>>> + u8 tsl; /* Time Stamp Lower Byte */ >>>>> +} __packed; >>>> >>>> Sorry, I missed the earlier discussion, but why the _packed? >>>> One u32 and 12 bytes makes a nice multiple of 4. >>> >>> Better safe than sorry, it's the layout of the registers in hardware, >>> don't let the compiler optimize here. >> >> Actually __packed makes it less safe, as the compiler now assumes >> the u32 id member is unaligned, and thus may turn 32-bit accesses into 4 >> byte accesses. >> >> Fortunately it won't happen in this case as the code uses writel/readl to >> acces the id member. > > Yes, as this are registers they must not be accessed directly. However > we can use "__attribute__ ((packed, aligned(4)))" to tell the compiler > that the base address of this struct is always aligned to 4 bytes. I truly think using packed here is rediculous, please remove it unless you can prove that things won't work without it. Thanks.