coreboot-kgpe-d16/payloads/libpayload/drivers/usb/usbmsc.c

718 lines
20 KiB
C
Raw Normal View History

/*
*
* Copyright (C) 2008 coresystems GmbH
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
//#define USB_DEBUG
#include <endian.h>
#include <usb/usb.h>
#include <usb/usbmsc.h>
#include <usb/usbdisk.h>
enum {
msc_subclass_rbc = 0x1,
msc_subclass_mmc2 = 0x2,
msc_subclass_qic157 = 0x3,
msc_subclass_ufi = 0x4,
msc_subclass_sff8070i = 0x5,
msc_subclass_scsitrans = 0x6
};
static const char *msc_subclass_strings[7] = {
"(none)",
"RBC",
"MMC-2",
"QIC-157",
"UFI",
"SFF-8070i",
"SCSI transparent"
};
enum {
msc_proto_cbi_wcomp = 0x0,
msc_proto_cbi_wocomp = 0x1,
msc_proto_bulk_only = 0x50
};
static const char *msc_protocol_strings[0x51] = {
"Control/Bulk/Interrupt protocol (with command completion interrupt)",
"Control/Bulk/Interrupt protocol (with no command completion interrupt)",
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
"Bulk-Only Transport"
};
static void
usb_msc_create_disk (usbdev_t *dev)
{
if (usbdisk_create) {
usbdisk_create (dev);
MSC_INST (dev)->usbdisk_created = 1;
}
}
static void
usb_msc_remove_disk (usbdev_t *dev)
{
if (MSC_INST (dev)->usbdisk_created && usbdisk_remove) {
usbdisk_remove (dev);
MSC_INST (dev)->usbdisk_created = 0;
}
}
static void
usb_msc_destroy (usbdev_t *dev)
{
if (dev->data) {
usb_msc_remove_disk (dev);
free (dev->data);
}
dev->data = 0;
}
const int DEV_RESET = 0xff;
const int GET_MAX_LUN = 0xfe;
/* Many USB3 devices do not work with large transfer requests.
* Limit the request size to 64KB chunks to ensure maximum compatibility. */
const int MAX_CHUNK_BYTES = 1024 * 64;
const unsigned int cbw_signature = 0x43425355;
const unsigned int csw_signature = 0x53425355;
typedef struct {
unsigned int dCBWSignature;
unsigned int dCBWTag;
unsigned int dCBWDataTransferLength;
unsigned char bmCBWFlags;
unsigned long bCBWLUN:4;
unsigned long:4;
unsigned long bCBWCBLength:5;
unsigned long:3;
unsigned char CBWCB[31 - 15];
} __packed cbw_t;
typedef struct {
unsigned int dCSWSignature;
unsigned int dCSWTag;
unsigned int dCSWDataResidue;
unsigned char bCSWStatus;
} __packed csw_t;
enum {
/*
* MSC commands can be
* successful,
* fail with proper response or
* fail totally, which results in detaching of the USB device
* and immediate cleanup of the usbdev_t structure.
* In the latter case the caller has to make sure, that he won't
* use the device any more.
*/
MSC_COMMAND_OK = 0, MSC_COMMAND_FAIL, MSC_COMMAND_DETACHED
};
static int
request_sense (usbdev_t *dev);
static int
request_sense_no_media (usbdev_t *dev);
static void
usb_msc_poll (usbdev_t *dev);
static int
reset_transport (usbdev_t *dev)
{
dev_req_t dr;
memset (&dr, 0, sizeof (dr));
dr.bmRequestType = 0;
dr.data_dir = host_to_device;
#ifndef QEMU
dr.req_type = class_type;
dr.req_recp = iface_recp;
#endif
dr.bRequest = DEV_RESET;
dr.wValue = 0;
dr.wIndex = 0;
dr.wLength = 0;
if (MSC_INST (dev)->quirks & USB_MSC_QUIRK_NO_RESET)
return MSC_COMMAND_FAIL;
/* if any of these fails, detach device, as we are lost */
libpayload: Make USB transfer functions return amount of bytes The USB bulk and control transfer functions in libpayload currently always return 0 for success and 1 for all errors. This is sufficient for current use cases (essentially just mass storage), but other classes (like certain Ethernet adapters) need to be able to tell if a transfer reached the intended amount of bytes, or if it fell short. This patch slightly changes that USB API to return -1 on errors, and the amount of transferred bytes on successes. All drivers in the current libpayload mainline are modified to conform to the new error detection model. Any third party users of this API will need to adapt their if (...<controller>->bulk/control(...)) checks to if (...<controller>->bulk/control(...) < 0) as well. The host controller drivers for OHCI and EHCI correctly implement the new behavior. UHCI and the XHCI stub just comply with the new API by returning 0 or -1, but do not actually count the returned bytes. Signed-off-by: Julius Werner <jwerner@chromium.org> Reviewed-on: https://gerrit.chromium.org/gerrit/48308 Reviewed-by: Gabe Black <gabeblack@chromium.org> Reviewed-by: Stefan Reinauer <reinauer@google.com> Tested-by: Gabe Black <gabeblack@chromium.org> Commit-Queue: Gabe Black <gabeblack@chromium.org> Updated the patch to support XHCI as well. Change-Id: Ic2ea2810c5edb992cbe185bc9711d2f8f557cae6 (cherry picked from commit e39e2d84762a3804653d950a228ed2269c651458) Signed-off-by: Isaac Christensen <isaac.christensen@se-eng.com> Reviewed-on: http://review.coreboot.org/6390 Tested-by: build bot (Jenkins) Reviewed-by: Ronald G. Minnich <rminnich@gmail.com> Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
2013-02-21 22:41:40 +01:00
if (dev->controller->control (dev, OUT, sizeof (dr), &dr, 0, 0) < 0 ||
clear_stall (MSC_INST (dev)->bulk_in) ||
clear_stall (MSC_INST (dev)->bulk_out)) {
usb_debug ("Detaching unresponsive device.\n");
usb_detach_device (dev->controller, dev->address);
return MSC_COMMAND_DETACHED;
}
/* return fail as we are only called in case of failure */
return MSC_COMMAND_FAIL;
}
/* device may stall this command, so beware! */
libpayload: usbmsc: Implement limited LUN support I always thought the support for multiple logical SCSI units in the USB mass storage class was a dead feature. Turns out that it's actually used by SD card readers that provide multiple slots (e.g. one regular sized and one micro-SD). Implementing perfect support for that would require a major redesign of the whole MSC stack, since the one device -> one disk assumption is deeply embedded in our data structures. Instead, this patch implements a poor man's LUN support that will just cycle through all available LUNs (in multiple calls to usb_msc_poll()) until it finds a connected device. This should be reasonable enough to allow these card readers to be usable while only requiring superficial changes. Also removes the unused 'protocol' attribute of usb_msc_inst_t. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=Alternatively plug an SD or micro-SD card (or both) into my card reader, confirm that one of them is correctly detected at all times. Original-Change-Id: I3df4ca88afe2dcf7928b823aa2a73c2b0f599cf2 Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198101 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 960534a20e4334772c29355bb0d310b3f41b31ee) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I39909fc96e32c9a5d76651d91c2b5c16c89ace9e Reviewed-on: http://review.coreboot.org/7904 Tested-by: build bot (Jenkins) Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net> Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com>
2014-05-03 01:35:50 +02:00
static void
initialize_luns (usbdev_t *dev)
{
libpayload: usbmsc: Implement limited LUN support I always thought the support for multiple logical SCSI units in the USB mass storage class was a dead feature. Turns out that it's actually used by SD card readers that provide multiple slots (e.g. one regular sized and one micro-SD). Implementing perfect support for that would require a major redesign of the whole MSC stack, since the one device -> one disk assumption is deeply embedded in our data structures. Instead, this patch implements a poor man's LUN support that will just cycle through all available LUNs (in multiple calls to usb_msc_poll()) until it finds a connected device. This should be reasonable enough to allow these card readers to be usable while only requiring superficial changes. Also removes the unused 'protocol' attribute of usb_msc_inst_t. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=Alternatively plug an SD or micro-SD card (or both) into my card reader, confirm that one of them is correctly detected at all times. Original-Change-Id: I3df4ca88afe2dcf7928b823aa2a73c2b0f599cf2 Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198101 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 960534a20e4334772c29355bb0d310b3f41b31ee) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I39909fc96e32c9a5d76651d91c2b5c16c89ace9e Reviewed-on: http://review.coreboot.org/7904 Tested-by: build bot (Jenkins) Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net> Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com>
2014-05-03 01:35:50 +02:00
usbmsc_inst_t *msc = MSC_INST (dev);
dev_req_t dr;
dr.bmRequestType = 0;
dr.data_dir = device_to_host;
#ifndef QEMU
dr.req_type = class_type;
dr.req_recp = iface_recp;
#endif
dr.bRequest = GET_MAX_LUN;
dr.wValue = 0;
dr.wIndex = 0;
dr.wLength = 1;
if (MSC_INST (dev)->quirks & USB_MSC_QUIRK_NO_LUNS ||
dev->controller->control (dev, IN, sizeof (dr), &dr,
libpayload: usbmsc: Implement limited LUN support I always thought the support for multiple logical SCSI units in the USB mass storage class was a dead feature. Turns out that it's actually used by SD card readers that provide multiple slots (e.g. one regular sized and one micro-SD). Implementing perfect support for that would require a major redesign of the whole MSC stack, since the one device -> one disk assumption is deeply embedded in our data structures. Instead, this patch implements a poor man's LUN support that will just cycle through all available LUNs (in multiple calls to usb_msc_poll()) until it finds a connected device. This should be reasonable enough to allow these card readers to be usable while only requiring superficial changes. Also removes the unused 'protocol' attribute of usb_msc_inst_t. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=Alternatively plug an SD or micro-SD card (or both) into my card reader, confirm that one of them is correctly detected at all times. Original-Change-Id: I3df4ca88afe2dcf7928b823aa2a73c2b0f599cf2 Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198101 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 960534a20e4334772c29355bb0d310b3f41b31ee) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I39909fc96e32c9a5d76651d91c2b5c16c89ace9e Reviewed-on: http://review.coreboot.org/7904 Tested-by: build bot (Jenkins) Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net> Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com>
2014-05-03 01:35:50 +02:00
sizeof (msc->num_luns), &msc->num_luns) < 0)
msc->num_luns = 0; /* assume only 1 lun if req fails */
msc->num_luns++; /* Get Max LUN returns number of last LUN */
msc->lun = 0;
}
unsigned int tag;
static void
wrap_cbw (cbw_t *cbw, int datalen, cbw_direction dir, const u8 *cmd,
libpayload: usbmsc: Implement limited LUN support I always thought the support for multiple logical SCSI units in the USB mass storage class was a dead feature. Turns out that it's actually used by SD card readers that provide multiple slots (e.g. one regular sized and one micro-SD). Implementing perfect support for that would require a major redesign of the whole MSC stack, since the one device -> one disk assumption is deeply embedded in our data structures. Instead, this patch implements a poor man's LUN support that will just cycle through all available LUNs (in multiple calls to usb_msc_poll()) until it finds a connected device. This should be reasonable enough to allow these card readers to be usable while only requiring superficial changes. Also removes the unused 'protocol' attribute of usb_msc_inst_t. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=Alternatively plug an SD or micro-SD card (or both) into my card reader, confirm that one of them is correctly detected at all times. Original-Change-Id: I3df4ca88afe2dcf7928b823aa2a73c2b0f599cf2 Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198101 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 960534a20e4334772c29355bb0d310b3f41b31ee) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I39909fc96e32c9a5d76651d91c2b5c16c89ace9e Reviewed-on: http://review.coreboot.org/7904 Tested-by: build bot (Jenkins) Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net> Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com>
2014-05-03 01:35:50 +02:00
int cmdlen, u8 lun)
{
memset (cbw, 0, sizeof (cbw_t));
/* commands are typically shorter, but we don't want overflows */
if (cmdlen > sizeof(cbw->CBWCB)) {
cmdlen = sizeof(cbw->CBWCB);
}
cbw->dCBWSignature = cbw_signature;
cbw->dCBWTag = ++tag;
libpayload: usbmsc: Implement limited LUN support I always thought the support for multiple logical SCSI units in the USB mass storage class was a dead feature. Turns out that it's actually used by SD card readers that provide multiple slots (e.g. one regular sized and one micro-SD). Implementing perfect support for that would require a major redesign of the whole MSC stack, since the one device -> one disk assumption is deeply embedded in our data structures. Instead, this patch implements a poor man's LUN support that will just cycle through all available LUNs (in multiple calls to usb_msc_poll()) until it finds a connected device. This should be reasonable enough to allow these card readers to be usable while only requiring superficial changes. Also removes the unused 'protocol' attribute of usb_msc_inst_t. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=Alternatively plug an SD or micro-SD card (or both) into my card reader, confirm that one of them is correctly detected at all times. Original-Change-Id: I3df4ca88afe2dcf7928b823aa2a73c2b0f599cf2 Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198101 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 960534a20e4334772c29355bb0d310b3f41b31ee) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I39909fc96e32c9a5d76651d91c2b5c16c89ace9e Reviewed-on: http://review.coreboot.org/7904 Tested-by: build bot (Jenkins) Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net> Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com>
2014-05-03 01:35:50 +02:00
cbw->bCBWLUN = lun;
cbw->dCBWDataTransferLength = datalen;
cbw->bmCBWFlags = dir;
libpayload: usbmsc: Set correct allocation length for REQUEST SENSE So I was debugging this faulty USB SD card reader that would just fail it's REQUEST SENSE response for some reason (sending the CSW immediately without the data), cursing those damn device vendors for building non-compliant crap like I always do... when I noticed that we do not actually set the Allocation Length field in our REQUEST SENSE command block at all! We set a length in the CBW, but the SCSI command still has its own length field and the SCSI spec specifically says that the device has to return the exact amount of bytes listed there (even if it's 0). I don't know what's more suprising: that we had such a blatant bug in this stack for so long, or that this card reader is really the first device to actually be spec compliant in that regard. This patch fixes the bug and changes the command block structures to be a little easier to read (why that field was called 'lun' before is beyond me... LUN is a transport level thing and should never appear in the command block at all, for any command). It also fixes a memcpy() in wrap_cbw() to avoid a read buffer overflow that might expose stack frame data to the device. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=The card reader works now (for it's first LUN at least). Original-Change-Id: I86fdcae2ea4d2e2939e3676d31d8b6a4e797873b Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198100 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 88943d9715994a14c50e74170f2453cceca0983b) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I3097c223248c07c866a33d4ab8f3db1a7082a815 Reviewed-on: http://review.coreboot.org/7903 Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com> Tested-by: build bot (Jenkins) Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com> Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
2014-05-02 23:51:03 +02:00
memcpy (cbw->CBWCB, cmd, cmdlen);
cbw->bCBWCBLength = cmdlen;
}
static int
get_csw (endpoint_t *ep, csw_t *csw)
{
hci_t *ctrlr = ep->dev->controller;
int ret = ctrlr->bulk (ep, sizeof (csw_t), (u8 *) csw, 1);
/* Some broken sticks send a zero-length packet at the end of their data
transfer which would show up here. Skip it to get the actual CSW. */
if (ret == 0)
ret = ctrlr->bulk (ep, sizeof (csw_t), (u8 *) csw, 1);
if (ret < 0) {
clear_stall (ep);
ret = ctrlr->bulk (ep, sizeof (csw_t), (u8 *) csw, 1);
if (ret < 0)
return reset_transport (ep->dev);
}
if (ret != sizeof(csw_t) || csw->dCSWTag != tag ||
csw->dCSWSignature != csw_signature) {
usb_debug ("MSC: received malformed CSW\n");
return reset_transport (ep->dev);
}
return MSC_COMMAND_OK;
}
static int
execute_command (usbdev_t *dev, cbw_direction dir, const u8 *cb, int cblen,
u8 *buf, int buflen, int residue_ok)
{
cbw_t cbw;
csw_t csw;
int always_succeed = 0;
if ((cb[0] == 0x1b) && (cb[4] == 1)) { //start command, always succeed
always_succeed = 1;
}
libpayload: usbmsc: Implement limited LUN support I always thought the support for multiple logical SCSI units in the USB mass storage class was a dead feature. Turns out that it's actually used by SD card readers that provide multiple slots (e.g. one regular sized and one micro-SD). Implementing perfect support for that would require a major redesign of the whole MSC stack, since the one device -> one disk assumption is deeply embedded in our data structures. Instead, this patch implements a poor man's LUN support that will just cycle through all available LUNs (in multiple calls to usb_msc_poll()) until it finds a connected device. This should be reasonable enough to allow these card readers to be usable while only requiring superficial changes. Also removes the unused 'protocol' attribute of usb_msc_inst_t. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=Alternatively plug an SD or micro-SD card (or both) into my card reader, confirm that one of them is correctly detected at all times. Original-Change-Id: I3df4ca88afe2dcf7928b823aa2a73c2b0f599cf2 Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198101 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 960534a20e4334772c29355bb0d310b3f41b31ee) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I39909fc96e32c9a5d76651d91c2b5c16c89ace9e Reviewed-on: http://review.coreboot.org/7904 Tested-by: build bot (Jenkins) Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net> Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com>
2014-05-03 01:35:50 +02:00
wrap_cbw (&cbw, buflen, dir, cb, cblen, MSC_INST (dev)->lun);
if (dev->controller->
libpayload: Make USB transfer functions return amount of bytes The USB bulk and control transfer functions in libpayload currently always return 0 for success and 1 for all errors. This is sufficient for current use cases (essentially just mass storage), but other classes (like certain Ethernet adapters) need to be able to tell if a transfer reached the intended amount of bytes, or if it fell short. This patch slightly changes that USB API to return -1 on errors, and the amount of transferred bytes on successes. All drivers in the current libpayload mainline are modified to conform to the new error detection model. Any third party users of this API will need to adapt their if (...<controller>->bulk/control(...)) checks to if (...<controller>->bulk/control(...) < 0) as well. The host controller drivers for OHCI and EHCI correctly implement the new behavior. UHCI and the XHCI stub just comply with the new API by returning 0 or -1, but do not actually count the returned bytes. Signed-off-by: Julius Werner <jwerner@chromium.org> Reviewed-on: https://gerrit.chromium.org/gerrit/48308 Reviewed-by: Gabe Black <gabeblack@chromium.org> Reviewed-by: Stefan Reinauer <reinauer@google.com> Tested-by: Gabe Black <gabeblack@chromium.org> Commit-Queue: Gabe Black <gabeblack@chromium.org> Updated the patch to support XHCI as well. Change-Id: Ic2ea2810c5edb992cbe185bc9711d2f8f557cae6 (cherry picked from commit e39e2d84762a3804653d950a228ed2269c651458) Signed-off-by: Isaac Christensen <isaac.christensen@se-eng.com> Reviewed-on: http://review.coreboot.org/6390 Tested-by: build bot (Jenkins) Reviewed-by: Ronald G. Minnich <rminnich@gmail.com> Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
2013-02-21 22:41:40 +01:00
bulk (MSC_INST (dev)->bulk_out, sizeof (cbw), (u8 *) &cbw, 0) < 0) {
return reset_transport (dev);
}
if (buflen > 0) {
if (dir == cbw_direction_data_in) {
if (dev->controller->
libpayload: Make USB transfer functions return amount of bytes The USB bulk and control transfer functions in libpayload currently always return 0 for success and 1 for all errors. This is sufficient for current use cases (essentially just mass storage), but other classes (like certain Ethernet adapters) need to be able to tell if a transfer reached the intended amount of bytes, or if it fell short. This patch slightly changes that USB API to return -1 on errors, and the amount of transferred bytes on successes. All drivers in the current libpayload mainline are modified to conform to the new error detection model. Any third party users of this API will need to adapt their if (...<controller>->bulk/control(...)) checks to if (...<controller>->bulk/control(...) < 0) as well. The host controller drivers for OHCI and EHCI correctly implement the new behavior. UHCI and the XHCI stub just comply with the new API by returning 0 or -1, but do not actually count the returned bytes. Signed-off-by: Julius Werner <jwerner@chromium.org> Reviewed-on: https://gerrit.chromium.org/gerrit/48308 Reviewed-by: Gabe Black <gabeblack@chromium.org> Reviewed-by: Stefan Reinauer <reinauer@google.com> Tested-by: Gabe Black <gabeblack@chromium.org> Commit-Queue: Gabe Black <gabeblack@chromium.org> Updated the patch to support XHCI as well. Change-Id: Ic2ea2810c5edb992cbe185bc9711d2f8f557cae6 (cherry picked from commit e39e2d84762a3804653d950a228ed2269c651458) Signed-off-by: Isaac Christensen <isaac.christensen@se-eng.com> Reviewed-on: http://review.coreboot.org/6390 Tested-by: build bot (Jenkins) Reviewed-by: Ronald G. Minnich <rminnich@gmail.com> Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
2013-02-21 22:41:40 +01:00
bulk (MSC_INST (dev)->bulk_in, buflen, buf, 0) < 0)
clear_stall (MSC_INST (dev)->bulk_in);
} else {
if (dev->controller->
libpayload: Make USB transfer functions return amount of bytes The USB bulk and control transfer functions in libpayload currently always return 0 for success and 1 for all errors. This is sufficient for current use cases (essentially just mass storage), but other classes (like certain Ethernet adapters) need to be able to tell if a transfer reached the intended amount of bytes, or if it fell short. This patch slightly changes that USB API to return -1 on errors, and the amount of transferred bytes on successes. All drivers in the current libpayload mainline are modified to conform to the new error detection model. Any third party users of this API will need to adapt their if (...<controller>->bulk/control(...)) checks to if (...<controller>->bulk/control(...) < 0) as well. The host controller drivers for OHCI and EHCI correctly implement the new behavior. UHCI and the XHCI stub just comply with the new API by returning 0 or -1, but do not actually count the returned bytes. Signed-off-by: Julius Werner <jwerner@chromium.org> Reviewed-on: https://gerrit.chromium.org/gerrit/48308 Reviewed-by: Gabe Black <gabeblack@chromium.org> Reviewed-by: Stefan Reinauer <reinauer@google.com> Tested-by: Gabe Black <gabeblack@chromium.org> Commit-Queue: Gabe Black <gabeblack@chromium.org> Updated the patch to support XHCI as well. Change-Id: Ic2ea2810c5edb992cbe185bc9711d2f8f557cae6 (cherry picked from commit e39e2d84762a3804653d950a228ed2269c651458) Signed-off-by: Isaac Christensen <isaac.christensen@se-eng.com> Reviewed-on: http://review.coreboot.org/6390 Tested-by: build bot (Jenkins) Reviewed-by: Ronald G. Minnich <rminnich@gmail.com> Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
2013-02-21 22:41:40 +01:00
bulk (MSC_INST (dev)->bulk_out, buflen, buf, 0) < 0)
clear_stall (MSC_INST (dev)->bulk_out);
}
}
int ret = get_csw (MSC_INST (dev)->bulk_in, &csw);
if (ret) {
return ret;
} else if (always_succeed == 1) {
/* return success, regardless of message */
return MSC_COMMAND_OK;
} else if (csw.bCSWStatus == 2) {
/* phase error, reset transport */
return reset_transport (dev);
} else if (csw.bCSWStatus == 0) {
if ((csw.dCSWDataResidue == 0) || residue_ok)
/* no error, exit */
return MSC_COMMAND_OK;
else
/* missed some bytes */
return MSC_COMMAND_FAIL;
} else {
if (cb[0] == 0x03)
/* requesting sense failed, that's bad */
return MSC_COMMAND_FAIL;
else if (cb[0] == 0)
/* If command was TEST UNIT READY determine if the
* device is of removable type indicating no media
* found. */
return request_sense_no_media (dev);
/* error "check condition" or reserved error */
ret = request_sense (dev);
/* return fail or the status of request_sense if it's worse */
return ret ? ret : MSC_COMMAND_FAIL;
}
}
typedef struct {
unsigned char command; //0
unsigned char res1; //1
unsigned int block; //2-5
unsigned char res2; //6
unsigned short numblocks; //7-8
libpayload: usbmsc: Set correct allocation length for REQUEST SENSE So I was debugging this faulty USB SD card reader that would just fail it's REQUEST SENSE response for some reason (sending the CSW immediately without the data), cursing those damn device vendors for building non-compliant crap like I always do... when I noticed that we do not actually set the Allocation Length field in our REQUEST SENSE command block at all! We set a length in the CBW, but the SCSI command still has its own length field and the SCSI spec specifically says that the device has to return the exact amount of bytes listed there (even if it's 0). I don't know what's more suprising: that we had such a blatant bug in this stack for so long, or that this card reader is really the first device to actually be spec compliant in that regard. This patch fixes the bug and changes the command block structures to be a little easier to read (why that field was called 'lun' before is beyond me... LUN is a transport level thing and should never appear in the command block at all, for any command). It also fixes a memcpy() in wrap_cbw() to avoid a read buffer overflow that might expose stack frame data to the device. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=The card reader works now (for it's first LUN at least). Original-Change-Id: I86fdcae2ea4d2e2939e3676d31d8b6a4e797873b Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198100 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 88943d9715994a14c50e74170f2453cceca0983b) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I3097c223248c07c866a33d4ab8f3db1a7082a815 Reviewed-on: http://review.coreboot.org/7903 Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com> Tested-by: build bot (Jenkins) Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com> Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
2014-05-02 23:51:03 +02:00
unsigned char control; //9 - the block is 10 bytes long
} __packed cmdblock_t;
typedef struct {
unsigned char command; //0
unsigned char res1; //1
unsigned char res2; //2
unsigned char res3; //3
libpayload: usbmsc: Set correct allocation length for REQUEST SENSE So I was debugging this faulty USB SD card reader that would just fail it's REQUEST SENSE response for some reason (sending the CSW immediately without the data), cursing those damn device vendors for building non-compliant crap like I always do... when I noticed that we do not actually set the Allocation Length field in our REQUEST SENSE command block at all! We set a length in the CBW, but the SCSI command still has its own length field and the SCSI spec specifically says that the device has to return the exact amount of bytes listed there (even if it's 0). I don't know what's more suprising: that we had such a blatant bug in this stack for so long, or that this card reader is really the first device to actually be spec compliant in that regard. This patch fixes the bug and changes the command block structures to be a little easier to read (why that field was called 'lun' before is beyond me... LUN is a transport level thing and should never appear in the command block at all, for any command). It also fixes a memcpy() in wrap_cbw() to avoid a read buffer overflow that might expose stack frame data to the device. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=The card reader works now (for it's first LUN at least). Original-Change-Id: I86fdcae2ea4d2e2939e3676d31d8b6a4e797873b Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198100 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 88943d9715994a14c50e74170f2453cceca0983b) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I3097c223248c07c866a33d4ab8f3db1a7082a815 Reviewed-on: http://review.coreboot.org/7903 Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com> Tested-by: build bot (Jenkins) Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com> Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
2014-05-02 23:51:03 +02:00
union { //4
struct {
unsigned long start:1; // for START STOP UNIT
unsigned long:7;
};
unsigned char length; // for REQUEST SENSE
};
unsigned char control; //5
} __packed cmdblock6_t;
/**
* Like readwrite_blocks, but for soft-sectors of 512b size. Converts the
* start and count from 512b units.
* Start and count must be aligned so that they match the native
* sector size.
*
* @param dev device to access
* @param start first sector to access
* @param n number of sectors to access
* @param dir direction of access: cbw_direction_data_in == read, cbw_direction_data_out == write
* @param buf buffer to read into or write from. Must be at least n*512 bytes
* @return 0 on success, 1 on failure
*/
int
readwrite_blocks_512 (usbdev_t *dev, int start, int n,
cbw_direction dir, u8 *buf)
{
int blocksize_divider = MSC_INST(dev)->blocksize / 512;
return readwrite_blocks (dev, start / blocksize_divider,
n / blocksize_divider, dir, buf);
}
/**
* Reads or writes a number of sequential blocks on a USB storage device.
* As it uses the READ(10) SCSI-2 command, it's limited to storage devices
* of at most 2TB. It assumes sectors of 512 bytes.
*
* @param dev device to access
* @param start first sector to access
* @param n number of sectors to access
* @param dir direction of access: cbw_direction_data_in == read, cbw_direction_data_out == write
* @param buf buffer to read into or write from. Must be at least n*sectorsize bytes
* @return 0 on success, 1 on failure
*/
static int
readwrite_chunk (usbdev_t *dev, int start, int n, cbw_direction dir, u8 *buf)
{
cmdblock_t cb;
memset (&cb, 0, sizeof (cb));
if (dir == cbw_direction_data_in) {
// read
cb.command = 0x28;
} else {
// write
cb.command = 0x2a;
}
cb.block = htonl (start);
cb.numblocks = htonw (n);
return execute_command (dev, dir, (u8 *) &cb, sizeof (cb), buf,
n * MSC_INST(dev)->blocksize, 0)
!= MSC_COMMAND_OK ? 1 : 0;
}
/**
* Reads or writes a number of sequential blocks on a USB storage device
* that is split into MAX_CHUNK_BYTES size requests.
*
* As it uses the READ(10) SCSI-2 command, it's limited to storage devices
* of at most 2TB. It assumes sectors of 512 bytes.
*
* @param dev device to access
* @param start first sector to access
* @param n number of sectors to access
* @param dir direction of access: cbw_direction_data_in == read,
* cbw_direction_data_out == write
* @param buf buffer to read into or write from.
* Must be at least n*sectorsize bytes
* @return 0 on success, 1 on failure
*/
int
readwrite_blocks (usbdev_t *dev, int start, int n, cbw_direction dir, u8 *buf)
{
int chunk_size = MAX_CHUNK_BYTES / MSC_INST(dev)->blocksize;
int chunk;
/* Read as many full chunks as needed. */
for (chunk = 0; chunk < (n / chunk_size); chunk++) {
if (readwrite_chunk (dev, start + (chunk * chunk_size),
chunk_size, dir,
buf + (chunk * MAX_CHUNK_BYTES))
!= MSC_COMMAND_OK)
return 1;
}
/* Read any remaining partial chunk at the end. */
if (n % chunk_size) {
if (readwrite_chunk (dev, start + (chunk * chunk_size),
n % chunk_size, dir,
buf + (chunk * MAX_CHUNK_BYTES))
!= MSC_COMMAND_OK)
return 1;
}
return 0;
}
/* Only request it, we don't interpret it.
On certain errors, that's necessary to get devices out of
a special state called "Contingent Allegiance Condition" */
static int
request_sense (usbdev_t *dev)
{
u8 buf[19];
cmdblock6_t cb;
memset (&cb, 0, sizeof (cb));
cb.command = 0x3;
libpayload: usbmsc: Set correct allocation length for REQUEST SENSE So I was debugging this faulty USB SD card reader that would just fail it's REQUEST SENSE response for some reason (sending the CSW immediately without the data), cursing those damn device vendors for building non-compliant crap like I always do... when I noticed that we do not actually set the Allocation Length field in our REQUEST SENSE command block at all! We set a length in the CBW, but the SCSI command still has its own length field and the SCSI spec specifically says that the device has to return the exact amount of bytes listed there (even if it's 0). I don't know what's more suprising: that we had such a blatant bug in this stack for so long, or that this card reader is really the first device to actually be spec compliant in that regard. This patch fixes the bug and changes the command block structures to be a little easier to read (why that field was called 'lun' before is beyond me... LUN is a transport level thing and should never appear in the command block at all, for any command). It also fixes a memcpy() in wrap_cbw() to avoid a read buffer overflow that might expose stack frame data to the device. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=The card reader works now (for it's first LUN at least). Original-Change-Id: I86fdcae2ea4d2e2939e3676d31d8b6a4e797873b Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198100 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 88943d9715994a14c50e74170f2453cceca0983b) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I3097c223248c07c866a33d4ab8f3db1a7082a815 Reviewed-on: http://review.coreboot.org/7903 Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com> Tested-by: build bot (Jenkins) Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com> Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
2014-05-02 23:51:03 +02:00
cb.length = sizeof (buf);
return execute_command (dev, cbw_direction_data_in, (u8 *) &cb,
libpayload: usbmsc: Set correct allocation length for REQUEST SENSE So I was debugging this faulty USB SD card reader that would just fail it's REQUEST SENSE response for some reason (sending the CSW immediately without the data), cursing those damn device vendors for building non-compliant crap like I always do... when I noticed that we do not actually set the Allocation Length field in our REQUEST SENSE command block at all! We set a length in the CBW, but the SCSI command still has its own length field and the SCSI spec specifically says that the device has to return the exact amount of bytes listed there (even if it's 0). I don't know what's more suprising: that we had such a blatant bug in this stack for so long, or that this card reader is really the first device to actually be spec compliant in that regard. This patch fixes the bug and changes the command block structures to be a little easier to read (why that field was called 'lun' before is beyond me... LUN is a transport level thing and should never appear in the command block at all, for any command). It also fixes a memcpy() in wrap_cbw() to avoid a read buffer overflow that might expose stack frame data to the device. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=The card reader works now (for it's first LUN at least). Original-Change-Id: I86fdcae2ea4d2e2939e3676d31d8b6a4e797873b Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198100 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 88943d9715994a14c50e74170f2453cceca0983b) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I3097c223248c07c866a33d4ab8f3db1a7082a815 Reviewed-on: http://review.coreboot.org/7903 Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com> Tested-by: build bot (Jenkins) Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com> Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
2014-05-02 23:51:03 +02:00
sizeof (cb), buf, sizeof (buf), 1);
}
static int request_sense_no_media (usbdev_t *dev)
{
u8 buf[19];
int ret;
cmdblock6_t cb;
memset (&cb, 0, sizeof (cb));
cb.command = 0x3;
libpayload: usbmsc: Set correct allocation length for REQUEST SENSE So I was debugging this faulty USB SD card reader that would just fail it's REQUEST SENSE response for some reason (sending the CSW immediately without the data), cursing those damn device vendors for building non-compliant crap like I always do... when I noticed that we do not actually set the Allocation Length field in our REQUEST SENSE command block at all! We set a length in the CBW, but the SCSI command still has its own length field and the SCSI spec specifically says that the device has to return the exact amount of bytes listed there (even if it's 0). I don't know what's more suprising: that we had such a blatant bug in this stack for so long, or that this card reader is really the first device to actually be spec compliant in that regard. This patch fixes the bug and changes the command block structures to be a little easier to read (why that field was called 'lun' before is beyond me... LUN is a transport level thing and should never appear in the command block at all, for any command). It also fixes a memcpy() in wrap_cbw() to avoid a read buffer overflow that might expose stack frame data to the device. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=The card reader works now (for it's first LUN at least). Original-Change-Id: I86fdcae2ea4d2e2939e3676d31d8b6a4e797873b Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198100 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 88943d9715994a14c50e74170f2453cceca0983b) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I3097c223248c07c866a33d4ab8f3db1a7082a815 Reviewed-on: http://review.coreboot.org/7903 Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com> Tested-by: build bot (Jenkins) Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com> Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
2014-05-02 23:51:03 +02:00
cb.length = sizeof (buf);
ret = execute_command (dev, cbw_direction_data_in, (u8 *) &cb,
libpayload: usbmsc: Set correct allocation length for REQUEST SENSE So I was debugging this faulty USB SD card reader that would just fail it's REQUEST SENSE response for some reason (sending the CSW immediately without the data), cursing those damn device vendors for building non-compliant crap like I always do... when I noticed that we do not actually set the Allocation Length field in our REQUEST SENSE command block at all! We set a length in the CBW, but the SCSI command still has its own length field and the SCSI spec specifically says that the device has to return the exact amount of bytes listed there (even if it's 0). I don't know what's more suprising: that we had such a blatant bug in this stack for so long, or that this card reader is really the first device to actually be spec compliant in that regard. This patch fixes the bug and changes the command block structures to be a little easier to read (why that field was called 'lun' before is beyond me... LUN is a transport level thing and should never appear in the command block at all, for any command). It also fixes a memcpy() in wrap_cbw() to avoid a read buffer overflow that might expose stack frame data to the device. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=The card reader works now (for it's first LUN at least). Original-Change-Id: I86fdcae2ea4d2e2939e3676d31d8b6a4e797873b Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198100 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 88943d9715994a14c50e74170f2453cceca0983b) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I3097c223248c07c866a33d4ab8f3db1a7082a815 Reviewed-on: http://review.coreboot.org/7903 Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com> Tested-by: build bot (Jenkins) Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com> Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
2014-05-02 23:51:03 +02:00
sizeof (cb), buf, sizeof (buf), 1);
if (ret)
return ret;
/* Check if sense key is set to NOT READY. */
if ((buf[2] & 0xf) != 2)
return MSC_COMMAND_FAIL;
/* Check if additional sense code is 0x3a. */
if (buf[12] != 0x3a)
return MSC_COMMAND_FAIL;
/* No media is present. Return MSC_COMMAND_OK while marking the disk
* not ready. */
usb_debug ("Empty media found.\n");
MSC_INST (dev)->ready = USB_MSC_NOT_READY;
return MSC_COMMAND_OK;
}
static int
test_unit_ready (usbdev_t *dev)
{
cmdblock6_t cb;
memset (&cb, 0, sizeof (cb)); // full initialization for T-U-R
return execute_command (dev, cbw_direction_data_out, (u8 *) &cb,
sizeof (cb), 0, 0, 0);
}
static int
spin_up (usbdev_t *dev)
{
cmdblock6_t cb;
memset (&cb, 0, sizeof (cb));
cb.command = 0x1b;
libpayload: usbmsc: Set correct allocation length for REQUEST SENSE So I was debugging this faulty USB SD card reader that would just fail it's REQUEST SENSE response for some reason (sending the CSW immediately without the data), cursing those damn device vendors for building non-compliant crap like I always do... when I noticed that we do not actually set the Allocation Length field in our REQUEST SENSE command block at all! We set a length in the CBW, but the SCSI command still has its own length field and the SCSI spec specifically says that the device has to return the exact amount of bytes listed there (even if it's 0). I don't know what's more suprising: that we had such a blatant bug in this stack for so long, or that this card reader is really the first device to actually be spec compliant in that regard. This patch fixes the bug and changes the command block structures to be a little easier to read (why that field was called 'lun' before is beyond me... LUN is a transport level thing and should never appear in the command block at all, for any command). It also fixes a memcpy() in wrap_cbw() to avoid a read buffer overflow that might expose stack frame data to the device. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=The card reader works now (for it's first LUN at least). Original-Change-Id: I86fdcae2ea4d2e2939e3676d31d8b6a4e797873b Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198100 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 88943d9715994a14c50e74170f2453cceca0983b) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I3097c223248c07c866a33d4ab8f3db1a7082a815 Reviewed-on: http://review.coreboot.org/7903 Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com> Tested-by: build bot (Jenkins) Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com> Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
2014-05-02 23:51:03 +02:00
cb.start = 1;
return execute_command (dev, cbw_direction_data_out, (u8 *) &cb,
sizeof (cb), 0, 0, 0);
}
static int
read_capacity (usbdev_t *dev)
{
cmdblock_t cb;
memset (&cb, 0, sizeof (cb));
cb.command = 0x25; // read capacity
u32 buf[2];
usb_debug ("Reading capacity of mass storage device.\n");
int count = 0, ret;
while (count++ < 20) {
switch (ret = execute_command
(dev, cbw_direction_data_in, (u8 *) &cb,
sizeof (cb), (u8 *)buf, 8, 0)) {
case MSC_COMMAND_OK:
break;
case MSC_COMMAND_FAIL:
continue;
default: /* if it's worse return */
return ret;
}
break;
}
if (count >= 20) {
// still not successful, assume 2tb in 512byte sectors, which is just the same garbage as any other number, but probably more usable.
usb_debug (" assuming 2 TB with 512-byte sectors as READ CAPACITY didn't answer.\n");
MSC_INST (dev)->numblocks = 0xffffffff;
MSC_INST (dev)->blocksize = 512;
} else {
MSC_INST (dev)->numblocks = ntohl(buf[0]) + 1;
MSC_INST (dev)->blocksize = ntohl(buf[1]);
}
usb_debug (" %d %d-byte sectors (%d MB)\n", MSC_INST (dev)->numblocks,
MSC_INST (dev)->blocksize,
/* round down high block counts to avoid integer overflow */
MSC_INST (dev)->numblocks > 1000000
? (MSC_INST (dev)->numblocks / 1000) * MSC_INST (dev)->blocksize / 1000 :
MSC_INST (dev)->numblocks * MSC_INST (dev)->blocksize / 1000 / 1000);
return MSC_COMMAND_OK;
}
static int
usb_msc_test_unit_ready (usbdev_t *dev)
{
int i;
time_t start_time_secs;
struct timeval tv;
/* SCSI/ATA specs say we have to wait up to 30s, but most devices
* are ready much sooner. Use a 5 sec timeout to better accommodate
* devices which fail to respond. */
const int timeout_secs = 5;
usb_debug (" Waiting for device to become ready...");
/* Initially mark the device ready. */
MSC_INST (dev)->ready = USB_MSC_READY;
gettimeofday (&tv, NULL);
start_time_secs = tv.tv_sec;
while (tv.tv_sec - start_time_secs < timeout_secs) {
switch (test_unit_ready (dev)) {
case MSC_COMMAND_OK:
break;
case MSC_COMMAND_FAIL:
mdelay (100);
usb_debug (".");
gettimeofday (&tv, NULL);
continue;
default:
/* Device detached, return immediately */
return USB_MSC_DETACHED;
}
break;
}
if (!(tv.tv_sec - start_time_secs < timeout_secs)) {
usb_debug ("timeout. Device not ready.\n");
MSC_INST (dev)->ready = USB_MSC_NOT_READY;
}
/* Don't bother spinning up the storage device if the device is not
* ready. This can happen when empty card readers are present.
* Polling will pick it back up if readiness changes. */
if (!MSC_INST (dev)->ready)
return MSC_INST (dev)->ready;
usb_debug ("ok.\n");
usb_debug (" spin up");
for (i = 0; i < 30; i++) {
usb_debug (".");
switch (spin_up (dev)) {
case MSC_COMMAND_OK:
usb_debug (" OK.");
break;
case MSC_COMMAND_FAIL:
mdelay (100);
continue;
default:
/* Device detached, return immediately */
return USB_MSC_DETACHED;
}
break;
}
usb_debug ("\n");
if (read_capacity (dev) == MSC_COMMAND_DETACHED)
return USB_MSC_DETACHED;
return MSC_INST (dev)->ready;
}
void
usb_msc_init (usbdev_t *dev)
{
configuration_descriptor_t *cd =
(configuration_descriptor_t *) dev->configuration;
interface_descriptor_t *interface =
(interface_descriptor_t *) (((char *) cd) + cd->bLength);
usb_debug (" it uses %s command set\n",
msc_subclass_strings[interface->bInterfaceSubClass]);
usb_debug (" it uses %s protocol\n",
msc_protocol_strings[interface->bInterfaceProtocol]);
if (interface->bInterfaceProtocol != 0x50) {
usb_debug (" Protocol not supported.\n");
usb_detach_device (dev->controller, dev->address);
return;
}
if ((interface->bInterfaceSubClass != 2) && // ATAPI 8020
(interface->bInterfaceSubClass != 5) && // ATAPI 8070
(interface->bInterfaceSubClass != 6)) { // SCSI
/* Other protocols, such as ATAPI don't seem to be very popular. looks like ATAPI would be really easy to add, if necessary. */
usb_debug (" Interface SubClass not supported.\n");
usb_detach_device (dev->controller, dev->address);
return;
}
usb_msc_force_init (dev, 0);
}
void usb_msc_force_init (usbdev_t *dev, u32 quirks)
{
int i;
/* init .data before setting .destroy */
dev->data = NULL;
dev->destroy = usb_msc_destroy;
dev->poll = usb_msc_poll;
dev->data = malloc (sizeof (usbmsc_inst_t));
if (!dev->data)
fatal("Not enough memory for USB MSC device.\n");
MSC_INST (dev)->bulk_in = 0;
MSC_INST (dev)->bulk_out = 0;
MSC_INST (dev)->usbdisk_created = 0;
MSC_INST (dev)->quirks = quirks;
for (i = 1; i <= dev->num_endp; i++) {
if (dev->endpoints[i].endpoint == 0)
continue;
if (dev->endpoints[i].type != BULK)
continue;
if ((dev->endpoints[i].direction == IN)
&& (MSC_INST (dev)->bulk_in == 0))
MSC_INST (dev)->bulk_in = &dev->endpoints[i];
if ((dev->endpoints[i].direction == OUT)
&& (MSC_INST (dev)->bulk_out == 0))
MSC_INST (dev)->bulk_out = &dev->endpoints[i];
}
if (MSC_INST (dev)->bulk_in == 0) {
usb_debug("couldn't find bulk-in endpoint.\n");
usb_detach_device (dev->controller, dev->address);
return;
}
if (MSC_INST (dev)->bulk_out == 0) {
usb_debug("couldn't find bulk-out endpoint.\n");
usb_detach_device (dev->controller, dev->address);
return;
}
usb_debug (" using endpoint %x as in, %x as out\n",
MSC_INST (dev)->bulk_in->endpoint,
MSC_INST (dev)->bulk_out->endpoint);
/* Some sticks need a little more time to get ready after SET_CONFIG. */
udelay(50);
libpayload: usbmsc: Implement limited LUN support I always thought the support for multiple logical SCSI units in the USB mass storage class was a dead feature. Turns out that it's actually used by SD card readers that provide multiple slots (e.g. one regular sized and one micro-SD). Implementing perfect support for that would require a major redesign of the whole MSC stack, since the one device -> one disk assumption is deeply embedded in our data structures. Instead, this patch implements a poor man's LUN support that will just cycle through all available LUNs (in multiple calls to usb_msc_poll()) until it finds a connected device. This should be reasonable enough to allow these card readers to be usable while only requiring superficial changes. Also removes the unused 'protocol' attribute of usb_msc_inst_t. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=Alternatively plug an SD or micro-SD card (or both) into my card reader, confirm that one of them is correctly detected at all times. Original-Change-Id: I3df4ca88afe2dcf7928b823aa2a73c2b0f599cf2 Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198101 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 960534a20e4334772c29355bb0d310b3f41b31ee) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I39909fc96e32c9a5d76651d91c2b5c16c89ace9e Reviewed-on: http://review.coreboot.org/7904 Tested-by: build bot (Jenkins) Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net> Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com>
2014-05-03 01:35:50 +02:00
initialize_luns (dev);
usb_debug (" has %d luns\n", MSC_INST (dev)->num_luns);
/* Test if unit is ready (nothing to do if it isn't). */
if (usb_msc_test_unit_ready (dev) != USB_MSC_READY)
return;
/* Create the disk. */
usb_msc_create_disk (dev);
}
static void
usb_msc_poll (usbdev_t *dev)
{
libpayload: usbmsc: Implement limited LUN support I always thought the support for multiple logical SCSI units in the USB mass storage class was a dead feature. Turns out that it's actually used by SD card readers that provide multiple slots (e.g. one regular sized and one micro-SD). Implementing perfect support for that would require a major redesign of the whole MSC stack, since the one device -> one disk assumption is deeply embedded in our data structures. Instead, this patch implements a poor man's LUN support that will just cycle through all available LUNs (in multiple calls to usb_msc_poll()) until it finds a connected device. This should be reasonable enough to allow these card readers to be usable while only requiring superficial changes. Also removes the unused 'protocol' attribute of usb_msc_inst_t. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=Alternatively plug an SD or micro-SD card (or both) into my card reader, confirm that one of them is correctly detected at all times. Original-Change-Id: I3df4ca88afe2dcf7928b823aa2a73c2b0f599cf2 Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198101 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 960534a20e4334772c29355bb0d310b3f41b31ee) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I39909fc96e32c9a5d76651d91c2b5c16c89ace9e Reviewed-on: http://review.coreboot.org/7904 Tested-by: build bot (Jenkins) Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net> Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com>
2014-05-03 01:35:50 +02:00
usbmsc_inst_t *msc = MSC_INST (dev);
int prev_ready = msc->ready;
if (usb_msc_test_unit_ready (dev) == USB_MSC_DETACHED)
return;
libpayload: usbmsc: Implement limited LUN support I always thought the support for multiple logical SCSI units in the USB mass storage class was a dead feature. Turns out that it's actually used by SD card readers that provide multiple slots (e.g. one regular sized and one micro-SD). Implementing perfect support for that would require a major redesign of the whole MSC stack, since the one device -> one disk assumption is deeply embedded in our data structures. Instead, this patch implements a poor man's LUN support that will just cycle through all available LUNs (in multiple calls to usb_msc_poll()) until it finds a connected device. This should be reasonable enough to allow these card readers to be usable while only requiring superficial changes. Also removes the unused 'protocol' attribute of usb_msc_inst_t. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=Alternatively plug an SD or micro-SD card (or both) into my card reader, confirm that one of them is correctly detected at all times. Original-Change-Id: I3df4ca88afe2dcf7928b823aa2a73c2b0f599cf2 Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198101 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 960534a20e4334772c29355bb0d310b3f41b31ee) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I39909fc96e32c9a5d76651d91c2b5c16c89ace9e Reviewed-on: http://review.coreboot.org/7904 Tested-by: build bot (Jenkins) Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net> Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com>
2014-05-03 01:35:50 +02:00
if (!prev_ready && msc->ready) {
usb_debug ("USB msc: not ready -> ready (lun %d)\n", msc->lun);
usb_msc_create_disk (dev);
libpayload: usbmsc: Implement limited LUN support I always thought the support for multiple logical SCSI units in the USB mass storage class was a dead feature. Turns out that it's actually used by SD card readers that provide multiple slots (e.g. one regular sized and one micro-SD). Implementing perfect support for that would require a major redesign of the whole MSC stack, since the one device -> one disk assumption is deeply embedded in our data structures. Instead, this patch implements a poor man's LUN support that will just cycle through all available LUNs (in multiple calls to usb_msc_poll()) until it finds a connected device. This should be reasonable enough to allow these card readers to be usable while only requiring superficial changes. Also removes the unused 'protocol' attribute of usb_msc_inst_t. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=Alternatively plug an SD or micro-SD card (or both) into my card reader, confirm that one of them is correctly detected at all times. Original-Change-Id: I3df4ca88afe2dcf7928b823aa2a73c2b0f599cf2 Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198101 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 960534a20e4334772c29355bb0d310b3f41b31ee) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I39909fc96e32c9a5d76651d91c2b5c16c89ace9e Reviewed-on: http://review.coreboot.org/7904 Tested-by: build bot (Jenkins) Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net> Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com>
2014-05-03 01:35:50 +02:00
} else if (prev_ready && !msc->ready) {
usb_debug ("USB msc: ready -> not ready (lun %d)\n", msc->lun);
usb_msc_remove_disk (dev);
libpayload: usbmsc: Implement limited LUN support I always thought the support for multiple logical SCSI units in the USB mass storage class was a dead feature. Turns out that it's actually used by SD card readers that provide multiple slots (e.g. one regular sized and one micro-SD). Implementing perfect support for that would require a major redesign of the whole MSC stack, since the one device -> one disk assumption is deeply embedded in our data structures. Instead, this patch implements a poor man's LUN support that will just cycle through all available LUNs (in multiple calls to usb_msc_poll()) until it finds a connected device. This should be reasonable enough to allow these card readers to be usable while only requiring superficial changes. Also removes the unused 'protocol' attribute of usb_msc_inst_t. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=Alternatively plug an SD or micro-SD card (or both) into my card reader, confirm that one of them is correctly detected at all times. Original-Change-Id: I3df4ca88afe2dcf7928b823aa2a73c2b0f599cf2 Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198101 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 960534a20e4334772c29355bb0d310b3f41b31ee) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I39909fc96e32c9a5d76651d91c2b5c16c89ace9e Reviewed-on: http://review.coreboot.org/7904 Tested-by: build bot (Jenkins) Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net> Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com>
2014-05-03 01:35:50 +02:00
} else if (!prev_ready && !msc->ready) {
u8 new_lun = (msc->lun + 1) % msc->num_luns;
usb_debug("USB msc: not ready (lun %d) -> lun %d\n", msc->lun,
libpayload: usbmsc: Implement limited LUN support I always thought the support for multiple logical SCSI units in the USB mass storage class was a dead feature. Turns out that it's actually used by SD card readers that provide multiple slots (e.g. one regular sized and one micro-SD). Implementing perfect support for that would require a major redesign of the whole MSC stack, since the one device -> one disk assumption is deeply embedded in our data structures. Instead, this patch implements a poor man's LUN support that will just cycle through all available LUNs (in multiple calls to usb_msc_poll()) until it finds a connected device. This should be reasonable enough to allow these card readers to be usable while only requiring superficial changes. Also removes the unused 'protocol' attribute of usb_msc_inst_t. BRANCH=rambi?,nyan BUG=chrome-os-partner:28437 TEST=Alternatively plug an SD or micro-SD card (or both) into my card reader, confirm that one of them is correctly detected at all times. Original-Change-Id: I3df4ca88afe2dcf7928b823aa2a73c2b0f599cf2 Original-Signed-off-by: Julius Werner <jwerner@chromium.org> Original-Reviewed-on: https://chromium-review.googlesource.com/198101 Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org> (cherry picked from commit 960534a20e4334772c29355bb0d310b3f41b31ee) Signed-off-by: Marc Jones <marc.jones@se-eng.com> Change-Id: I39909fc96e32c9a5d76651d91c2b5c16c89ace9e Reviewed-on: http://review.coreboot.org/7904 Tested-by: build bot (Jenkins) Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net> Reviewed-by: Edward O'Callaghan <eocallaghan@alterapraxis.com>
2014-05-03 01:35:50 +02:00
new_lun);
msc->lun = new_lun;
}
}