From: Marek Vasut <marex@denx.de>
To: u-boot@lists.denx.de
Subject: [U-Boot] [PATCH v4] usb: align buffers at cacheline
Date: Fri, 2 Mar 2012 15:41:47 +0100 [thread overview]
Message-ID: <201203021541.48196.marex@denx.de> (raw)
In-Reply-To: <4F50D29A.4030905@nvidia.com>
> On Friday 02 March 2012 07:16 PM, Marek Vasut wrote:
> >> As DMA expects the buffers to be equal and larger then
> >> cache lines, This aligns buffers at cacheline.
> >>
> >> Signed-off-by: Puneet Saxena<puneets@nvidia.com>
> >> ---
> >>
> >> Changes for V3:
> >> - Removed local descriptor elements copy to global descriptor
> >> elements - Removed "Signed-off-by: Jim Lin<jilin@nvidia.com>" from
> >> commit
> >>
> >> message
> >>
> >> Changes for V4:
> >> - Added memcpy to copy local descriptor to global descriptor.
> >>
> >> Without that, USB version, class, vendor, product Id...etc is not
> >>
> >> configured. This information is useful for loading correct device driver
> >> and possible configuration.
> >>
> >> common/cmd_usb.c | 3 +-
> >> common/usb.c | 56
> >>
> >> ++++++++++++++++++++++------------------ common/usb_storage.c |
> >> 59 ++++++++++++++++++++---------------------- disk/part_dos.c
> >>
> >> | 2 +-
> >>
> >> drivers/usb/host/ehci-hcd.c | 8 ++++++
> >> include/scsi.h | 4 ++-
> >> 6 files changed, 73 insertions(+), 59 deletions(-)
> >>
> >> diff --git a/common/cmd_usb.c b/common/cmd_usb.c
> >> index 320667f..bca9d94 100644
> >> --- a/common/cmd_usb.c
> >> +++ b/common/cmd_usb.c
> >> @@ -150,7 +150,8 @@ void usb_display_class_sub(unsigned char dclass,
> >> unsigned char subclass,
> >>
> >> void usb_display_string(struct usb_device *dev, int index)
> >> {
> >>
> >> - char buffer[256];
> >> + ALLOC_CACHE_ALIGN_BUFFER(char, buffer, 256);
> >> +
> >>
> >> if (index != 0) {
> >>
> >> if (usb_string(dev, index,&buffer[0], 256)> 0)
> >>
> >> printf("String: \"%s\"", buffer);
> >>
> >> diff --git a/common/usb.c b/common/usb.c
> >> index 6e21ae2..42a44e2 100644
> >> --- a/common/usb.c
> >> +++ b/common/usb.c
> >> @@ -73,7 +73,6 @@ static struct usb_device usb_dev[USB_MAX_DEVICE];
> >>
> >> static int dev_index;
> >> static int running;
> >> static int asynch_allowed;
> >>
> >> -static struct devrequest setup_packet;
> >>
> >> char usb_started; /* flag for the started/stopped USB status */
> >>
> >> @@ -185,23 +184,25 @@ int usb_control_msg(struct usb_device *dev,
> >> unsigned int pipe, unsigned short value, unsigned short index,
> >>
> >> void *data, unsigned short size, int timeout)
> >>
> >> {
> >>
> >> + ALLOC_CACHE_ALIGN_BUFFER(struct devrequest, setup_packet,
> >> + sizeof(struct devrequest));
> >>
> >> if ((timeout == 0)&& (!asynch_allowed)) {
> >>
> >> /* request for a asynch control pipe is not allowed */
> >> return -1;
> >>
> >> }
> >>
> >> /* set setup command */
> >>
> >> - setup_packet.requesttype = requesttype;
> >> - setup_packet.request = request;
> >> - setup_packet.value = cpu_to_le16(value);
> >> - setup_packet.index = cpu_to_le16(index);
> >> - setup_packet.length = cpu_to_le16(size);
> >> + setup_packet->requesttype = requesttype;
> >> + setup_packet->request = request;
> >> + setup_packet->value = cpu_to_le16(value);
> >> + setup_packet->index = cpu_to_le16(index);
> >> + setup_packet->length = cpu_to_le16(size);
> >>
> >> USB_PRINTF("usb_control_msg: request: 0x%X, requesttype: 0x%X, " \
> >>
> >> "value 0x%X index 0x%X length 0x%X\n",
> >> request, requesttype, value, index, size);
> >>
> >> dev->status = USB_ST_NOT_PROC; /*not yet processed */
> >>
> >> - submit_control_msg(dev, pipe, data, size,&setup_packet);
> >> + submit_control_msg(dev, pipe, data, size, setup_packet);
> >>
> >> if (timeout == 0)
> >>
> >> return (int)size;
> >>
> >> @@ -698,7 +699,7 @@ static int usb_string_sub(struct usb_device *dev,
> >> unsigned int langid, */
> >>
> >> int usb_string(struct usb_device *dev, int index, char *buf, size_t
> >> size) {
> >>
> >> - unsigned char mybuf[USB_BUFSIZ];
> >> + ALLOC_CACHE_ALIGN_BUFFER(unsigned char, mybuf, USB_BUFSIZ);
> >>
> >> unsigned char *tbuf;
> >> int err;
> >> unsigned int u, idx;
> >>
> >> @@ -798,7 +799,7 @@ int usb_new_device(struct usb_device *dev)
> >>
> >> {
> >>
> >> int addr, err;
> >> int tmp;
> >>
> >> - unsigned char tmpbuf[USB_BUFSIZ];
> >> + ALLOC_CACHE_ALIGN_BUFFER(unsigned char, tmpbuf, USB_BUFSIZ);
> >>
> >> /* We still haven't set the Address yet */
> >> addr = dev->devnum;
> >>
> >> @@ -909,7 +910,7 @@ int usb_new_device(struct usb_device *dev)
> >>
> >> tmp = sizeof(dev->descriptor);
> >>
> >> err = usb_get_descriptor(dev, USB_DT_DEVICE, 0,
> >>
> >> - &dev->descriptor, sizeof(dev->descriptor));
> >> + desc, sizeof(dev->descriptor));
> >>
> >> if (err< tmp) {
> >>
> >> if (err< 0)
> >>
> >> printf("unable to get device descriptor (error=%d)\n",
> >>
> >> @@ -919,14 +920,18 @@ int usb_new_device(struct usb_device *dev)
> >>
> >> "(expected %i, got %i)\n", tmp, err);
> >>
> >> return 1;
> >>
> >> }
> >>
> >> +
> >> + /* Now copy the local device descriptor to global device descriptor*/
> >> + memcpy(&dev->descriptor, desc, sizeof(dev->descriptor));
> >
> > Hey, it's almost perfect!
> >
> > Just one last question -- why do you need to copy this stuff?
>
> We need to copy this stuff as I spoke in previous reply -
> "memcpy" is needed to configure Usb version, vendor id, prod Id ..etc
>
> > of the device. Its also useful to hook actual device driver and detect
> > no. of configuration supported. The effect can be verified
>
> with/without using memcpy
> in "usb tree" and "usb info" commands.
>
> > It's because dev-
> >
> >> descriptor is unaligned?
>
> No, As global dev - descriptor is unaligned we are passing aligned local
> dev-desc and memcpy in global.
Can't we just align the global one?
M
next prev parent reply other threads:[~2012-03-02 14:41 UTC|newest]
Thread overview: 83+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-02-17 10:50 [U-Boot] [PATCH] usb: align buffers at cache boundary Puneet Saxena
2012-02-17 17:23 ` Mike Frysinger
2012-02-23 14:25 ` [U-Boot] [PATCH 1/2] usb: align buffers at cacheline Puneet Saxena
2012-02-23 18:15 ` Mike Frysinger
2012-02-24 11:27 ` puneets
2012-02-24 12:42 ` Simon Glass
2012-02-27 15:36 ` [U-Boot] [PATCH v2 " Puneet Saxena
2012-02-27 16:49 ` Marek Vasut
2012-02-27 17:03 ` Simon Glass
2012-02-27 17:11 ` Marek Vasut
2012-02-27 17:27 ` Simon Glass
2012-02-29 14:21 ` [U-Boot] [PATCH v3 " Puneet Saxena
2012-02-29 21:35 ` Marek Vasut
2012-03-01 13:51 ` puneets
2012-03-01 18:38 ` Marek Vasut
2012-03-02 6:56 ` puneets
2012-03-02 10:43 ` Marek Vasut
2012-03-02 12:50 ` puneets
2012-03-02 12:58 ` Marek Vasut
2012-03-02 13:35 ` [U-Boot] [PATCH v4] " Puneet Saxena
2012-03-02 13:46 ` Marek Vasut
2012-03-02 14:00 ` puneets
2012-03-02 14:41 ` Marek Vasut [this message]
2012-03-02 15:21 ` puneets
2012-03-02 15:59 ` Marek Vasut
2012-03-02 16:45 ` Wolfgang Denk
2012-03-05 7:16 ` [U-Boot] [PATCH v5] " Puneet Saxena
2012-03-05 13:24 ` Eric Nelson
2012-03-05 13:35 ` Marek Vasut
2012-03-05 14:46 ` [U-Boot] [PATCH v8] " Puneet Saxena
2012-03-05 15:35 ` Marek Vasut
2012-03-05 18:18 ` Simon Glass
2012-03-06 0:36 ` Marek Vasut
2012-03-06 0:39 ` Marek Vasut
2012-03-06 7:00 ` puneets
2012-03-06 8:22 ` Marek Vasut
2012-03-06 3:07 ` Mike Frysinger
2012-03-07 7:12 ` puneets
2012-03-07 9:20 ` puneets
2012-03-07 22:06 ` Marek Vasut
2012-03-08 11:21 ` puneets
2012-03-08 14:12 ` Marek Vasut
2012-03-09 6:15 ` puneets
2012-03-09 12:03 ` Marek Vasut
2012-03-11 2:35 ` Mike Frysinger
2012-03-14 2:05 ` Marek Vasut
2012-03-16 4:39 ` Marek Vasut
2012-03-16 7:42 ` puneets
2012-03-16 8:52 ` Marek Vasut
2012-03-19 14:29 ` puneets
2012-03-19 14:43 ` Marek Vasut
2012-03-19 15:19 ` Tom Warren
2012-03-19 15:46 ` Marek Vasut
2012-04-02 15:59 ` Tom Warren
2012-04-02 16:11 ` Marek Vasut
2012-04-02 16:16 ` Tom Warren
2012-04-02 16:32 ` Marek Vasut
2012-04-03 6:05 ` puneets
2012-03-05 7:27 ` [U-Boot] [PATCH v6] " Puneet Saxena
2012-03-05 12:03 ` Marek Vasut
2012-03-05 12:21 ` [U-Boot] [PATCH v7] " Puneet Saxena
2012-03-05 12:41 ` Marek Vasut
2012-03-06 3:28 ` [U-Boot] [PATCH v4] " Mike Frysinger
2012-03-06 8:24 ` Marek Vasut
2012-03-06 16:42 ` Mike Frysinger
2012-02-29 14:21 ` [U-Boot] [PATCH v3 2/2] usb: Add CONFIG to fetch string descriptor Puneet Saxena
2012-02-29 21:29 ` Marek Vasut
2012-03-01 11:07 ` puneets
2012-03-01 11:45 ` Marek Vasut
2012-03-01 12:59 ` puneets
2012-03-01 13:13 ` Marek Vasut
2012-03-05 12:48 ` Marek Vasut
2012-03-05 13:14 ` puneets
2012-03-05 21:15 ` Marek Vasut
2012-02-28 9:34 ` [U-Boot] [PATCH v2 1/2] usb: align buffers at cacheline puneets
2012-02-29 21:38 ` Marek Vasut
2012-02-27 15:36 ` [U-Boot] [PATCH v2 2/2] usb: Add CONFIG to fetch string descriptor Puneet Saxena
2012-02-27 18:28 ` Mike Frysinger
2012-02-27 15:37 ` [U-Boot] [PATCH 1/2] usb: align buffers at cacheline puneets
2012-02-23 14:25 ` [U-Boot] [PATCH 2/2] usb: Add quirk "USB_QUIRK_STRING_FETCH_255" Puneet Saxena
2012-02-23 15:20 ` Tom Rini
2012-02-23 16:04 ` Tom Warren
2012-02-24 7:52 ` puneets
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=201203021541.48196.marex@denx.de \
--to=marex@denx.de \
--cc=u-boot@lists.denx.de \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox