All of lore.kernel.org
 help / color / mirror / Atom feed
From: Ben Hutchings <ben.hutchings@codethink.co.uk>
To: Ben Dooks <ben.dooks@codethink.co.uk>, netdev@vger.kernel.org
Cc: linux-kernel@lists.codethink.co.uk, linux-usb@vger.kernel.org,
	oneukum@suse.com, linux-kernel@vger.kernel.org,
	davem@davemloft.net
Subject: [2/4] usbnet: smsc95xx: align tx-buffer to word
Date: Thu, 04 Oct 2018 17:53:16 +0100	[thread overview]
Message-ID: <1538671996.3024.269.camel@codethink.co.uk> (raw)

On Tue, 2018-10-02 at 10:26 +0100, Ben Dooks wrote:
> The tegra driver requires alignment of the buffer, so try and
> make this better by pushing the buffer start back to an word
> aligned address. At the worst this makes memcpy() easier as
> it is word aligned, at best it makes sure the usb can directly
> map the buffer.
> 
> Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
> [todo - make this configurable]
[...]

I don't think you need a separate kconfig symbol for this.  Aligning RX
buffers to words (or better, cache lines) is almost always a win, so
long as the CPU can handle misaligned fields in the network/transport
headers.  You can use #ifdef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS to
check that.

It seems like NET_IP_ALIGN should be defined to 0 or 2 depending on
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS, but NET_IP_ALIGN predates the
latter.

Ben.

WARNING: multiple messages have this Message-ID (diff)
From: Ben Hutchings <ben.hutchings@codethink.co.uk>
To: Ben Dooks <ben.dooks@codethink.co.uk>, netdev@vger.kernel.org
Cc: linux-kernel@lists.codethink.co.uk, linux-usb@vger.kernel.org,
	oneukum@suse.com, linux-kernel@vger.kernel.org,
	davem@davemloft.net
Subject: Re: [Linux-kernel] [PATCH 2/4] usbnet: smsc95xx: align tx-buffer to word
Date: Thu, 04 Oct 2018 17:53:16 +0100	[thread overview]
Message-ID: <1538671996.3024.269.camel@codethink.co.uk> (raw)
In-Reply-To: <20181002092645.1115-3-ben.dooks@codethink.co.uk>

On Tue, 2018-10-02 at 10:26 +0100, Ben Dooks wrote:
> The tegra driver requires alignment of the buffer, so try and
> make this better by pushing the buffer start back to an word
> aligned address. At the worst this makes memcpy() easier as
> it is word aligned, at best it makes sure the usb can directly
> map the buffer.
> 
> Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
> [todo - make this configurable]
[...]

I don't think you need a separate kconfig symbol for this.  Aligning RX
buffers to words (or better, cache lines) is almost always a win, so
long as the CPU can handle misaligned fields in the network/transport
headers.  You can use #ifdef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS to
check that.

It seems like NET_IP_ALIGN should be defined to 0 or 2 depending on
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS, but NET_IP_ALIGN predates the
latter.

Ben.

-- 
Ben Hutchings, Software Developer                         Codethink Ltd
https://www.codethink.co.uk/                 Dale House, 35 Dale Street
                                     Manchester, M1 2HF, United Kingdom

WARNING: multiple messages have this Message-ID (diff)
From: Ben Hutchings <ben.hutchings@codethink.co.uk>
To: Ben Dooks <ben.dooks@codethink.co.uk>, netdev@vger.kernel.org
Cc: linux-kernel@lists.codethink.co.uk, linux-usb@vger.kernel.org,
	oneukum@suse.com, linux-kernel@vger.kernel.org,
	davem@davemloft.net
Subject: Re: [PATCH 2/4] usbnet: smsc95xx: align tx-buffer to word
Date: Thu, 04 Oct 2018 17:53:16 +0100	[thread overview]
Message-ID: <1538671996.3024.269.camel@codethink.co.uk> (raw)
In-Reply-To: <20181002092645.1115-3-ben.dooks@codethink.co.uk>

On Tue, 2018-10-02 at 10:26 +0100, Ben Dooks wrote:
> The tegra driver requires alignment of the buffer, so try and
> make this better by pushing the buffer start back to an word
> aligned address. At the worst this makes memcpy() easier as
> it is word aligned, at best it makes sure the usb can directly
> map the buffer.
> 
> Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
> [todo - make this configurable]
[...]

I don't think you need a separate kconfig symbol for this.  Aligning RX
buffers to words (or better, cache lines) is almost always a win, so
long as the CPU can handle misaligned fields in the network/transport
headers.  You can use #ifdef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS to
check that.

It seems like NET_IP_ALIGN should be defined to 0 or 2 depending on
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS, but NET_IP_ALIGN predates the
latter.

Ben.

-- 
Ben Hutchings, Software Developer                         Codethink Ltd
https://www.codethink.co.uk/                 Dale House, 35 Dale Street
                                     Manchester, M1 2HF, United Kingdom

             reply	other threads:[~2018-10-04 16:53 UTC|newest]

Thread overview: 37+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-10-04 16:53 Ben Hutchings [this message]
2018-10-04 16:53 ` [PATCH 2/4] usbnet: smsc95xx: align tx-buffer to word Ben Hutchings
2018-10-04 16:53 ` [Linux-kernel] " Ben Hutchings
  -- strict thread matches above, loose matches on Subject: below --
2018-10-08  8:41 usbnet: smsc95xx: simplify tx_fixup code David Laight
2018-10-08  8:41 ` [PATCH] " David Laight
2018-10-06 17:28 David Miller
2018-10-06 17:28 ` [PATCH] " David Miller
2018-10-06 11:27 Ben Dooks
2018-10-06 11:27 ` [PATCH] " Ben Dooks
2018-10-05 21:24 David Miller
2018-10-05 21:24 ` [PATCH] " David Miller
2018-10-04 16:55 [3/4] usbnet: smsc95xx: check for csum being in last four bytes Ben Hutchings
2018-10-04 16:55 ` [PATCH 3/4] " Ben Hutchings
2018-10-04 16:55 ` [Linux-kernel] " Ben Hutchings
2018-10-03 16:25 usbnet: smsc95xx: simplify tx_fixup code Ben Dooks
2018-10-03 16:25 ` [PATCH] " Ben Dooks
2018-10-03 13:36 David Laight
2018-10-03 13:36 ` [PATCH] " David Laight
2018-10-02 16:56 Ben Dooks
2018-10-02 16:56 ` [PATCH] " Ben Dooks
2018-10-02 13:35 [2/4] usbnet: smsc95xx: align tx-buffer to word Ben Dooks
2018-10-02 13:35 ` [PATCH 2/4] " Ben Dooks
2018-10-02 13:19 [2/4] " David Laight
2018-10-02 13:19 ` [PATCH 2/4] " David Laight
2018-10-02 12:46 [1/4] usbnet: smsc95xx: add kconfig for turbo mode Andrew Lunn
2018-10-02 12:46 ` [PATCH 1/4] " Andrew Lunn
2018-10-02  9:45 [3/4] usbnet: smsc95xx: check for csum being in last four bytes Sergei Shtylyov
2018-10-02  9:45 ` [PATCH 3/4] " Sergei Shtylyov
2018-10-02  9:26 [4/4] usbnet: smsc95xx: fix rx packet alignment Ben Dooks
2018-10-02  9:26 ` [PATCH 4/4] " Ben Dooks
2018-10-02  9:26 [3/4] usbnet: smsc95xx: check for csum being in last four bytes Ben Dooks
2018-10-02  9:26 ` [PATCH 3/4] " Ben Dooks
2018-10-02  9:26 [2/4] usbnet: smsc95xx: align tx-buffer to word Ben Dooks
2018-10-02  9:26 ` [PATCH 2/4] " Ben Dooks
2018-10-02  9:26 [1/4] usbnet: smsc95xx: add kconfig for turbo mode Ben Dooks
2018-10-02  9:26 ` [PATCH 1/4] " Ben Dooks
2018-10-02  9:26 SMSC95XX driver updates Ben Dooks

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1538671996.3024.269.camel@codethink.co.uk \
    --to=ben.hutchings@codethink.co.uk \
    --cc=ben.dooks@codethink.co.uk \
    --cc=davem@davemloft.net \
    --cc=linux-kernel@lists.codethink.co.uk \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-usb@vger.kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=oneukum@suse.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body. 
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.