usb: storage: Fix TEST_UNIT_READY command size ...

TEST_UNIT_READY is 6 bytes long instead of 12. Fix it.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Default to a single LUN ...

For loop below is:

	for (lun=0; lun<=us->max_lun; lun++)
	    ...

so we need to set max_lun to 0 in order to default to having only one
LUN. Otherwise, we end up assuming the device has 2 valid LUNs.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Inline usb_limit_blk_cnt() ...

Usb_limit_blk_cnt() has only one user so we may as well inline it.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Drop unnecessary assignment ...

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Simplify I/O loop in usb_stor_blk_io() ...

Simplify I/O loop a bit, by re-arranging things and dropping
"sectors_done", "result" and "data" local variables.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Use simple boolean to speficy read vs. write operation ...

Use simple boolean to speficy read vs. write operation
usb_stor_blk_io() instead of having a custom anonymous enum.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Drop needless macro ...

There's only user of the to_usb_mass_storage() so we may as well
inline it.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Drop unnecessary check in usb_stor_blk_io() ...

Checking that sector_count is zero, shouldn't be necessary since block
layer won't call this function if there's no data to be read. Drop it.

Checking that blockbits is eqal to SECTOR_SHIFT isn't necessary, since
that field is filled by the driver and is not changed outsied of
it. We know it is going to be SECTOR_SHIFT. Drop it.

Checking sector_start > (ulong)-1 doesn't make sense at all since
sector start is 'int' and it can't possibly be greater that
ULONG_MAX. Drop it.

Checking for sector_start >= pblk_dev->blk.num_blocks isn't necessary
either, since we shouldn't receive request for invalid read from block
layer. Drop it.

Ditto for sector_count > INT_MAX and sector_start + sector_count >
pblk_dev->blk.num_blocks.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: Drop usb_disable_asynch() ...

There are no users of usb_control_msg() in the codebase that pass
timeout of 0, so it doesn't look like usb_disable_asynch() has any
effect on USB operation. Drop that function and remove all of its uses
to simplify things.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Use put_unaligned_be* helpers ...

Replace explicit endianness casts with put_unaligned_be* helpers.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Introduce usb_stor_transport() ...

Move retry logic found in all of the low-level usb_stor_* functions
into a standalone subroutine and convert the rest of the code to use
it.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Drop unused us_blkdev_list ...

There are no users of us_blkdev_list in the file. Drop it.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Drop unused fields in struct us_data ...

Only bulk-only transport is supported by the currennt codebase, so
ep_bInterval and recv_intr_ep are not really used. Remove them and all
related code. While at it remove flags and subclass as well since they
are not really used anywhere in the codebase.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Drop struct SCSI_cmd_block ...

All the info we need to pass to transport function can be captured in
a struct us_blk_dev and two byte arrays, so having a dedicated struct
with many unused fields doesn't really buy us anything. Drop the
struct and convert the rest of the code to pass needed data
explicitly.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Remove unused variables ...

Drop a number of variables no longer used in the code.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Make usb_stor_request_sense() a standalone function ...

Don't try to re-use caller's "ccb" and instead convert
usb_stor_request_sense() to use its own for clarity and to allow
futher code improvements. No functional change intended.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Share code for READ(10) and WRITE(10) ...

Both usb_stor_read_10() and usb_stor_write_10() do almost exactly the
same thing, so merge them into a signle routine and adjust all of the
users accordingly.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Drop extra call to transport in usb_stor_write_10() ...

There doesn't seem to be a particularly good reason to call
->trasport() one last time after 2 failures. Drop the call so allow
sharing this code with usb_stor_read_10() in the following commit.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Make usb_stor_write_10() a standalone function ...

Move special preparation steps done before the call to
usb_stor_write_10() inside the function itself for clarity and to
allow futher code improvements. No functional change intended.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Make usb_stor_read_10() a standalone function ...

Move special preparation steps done before the call to
usb_stor_read_10() inside the function itself for clarity and to allow
futher code improvements. No functional change intended.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Make usb_stor_test_unit_ready() a standalone function ...

Move special preparation steps done before the call to
usb_stor_test_unit_ready() inside the function itself for clarity and
to allow futher code improvements. No functional change intended.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Make usb_stor_inquiry() a standalone function ...

Move special preparation steps done before and after call to
usb_stor_inquiry() inside the function itself for clarity and to allow
futher code improvements. No functional change intended.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Make usb_stor_read_capacity() a standalone function ...

Move special preparation steps done before and after call to
usb_stor_read_capacity() inside the function itself for clarity and to
allow futher code improvements. No functional change intended.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Don't use "unsigned long" for 32-bit values ...

Unsignled long will expand to 64-bit unsigned integer on 64-bit
CPUs. This will break current code using it to read out two 32-bit
values returned by READ_CAPACITY. Fix the proble by using "u32"
explicitly.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Replace custom debug tracing with dev_dbg ...

Replace custom debug tracing macros with a call to dev_dbg to simplify
code and add device info into degug output.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: storage: Simplify memory allocation in usb_stor_probe() ...

Replace explicit malloc() + OOM check and memset() with a single call
to xzalloc().

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

string: Fix (v)asprintf prototypes ...

Our asprintf and vasprintf have different prototypes than the glibc
functions. This causes trouble when we want to share barebox code
with userspace code. Change the prototypes for (v)asprintf to match
the glibc prototypes. Since the current (v)asprintf are convenient
to use change the existing functions to b(v)asprintf.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

usb: abolish wait_ms() function ...

This function is only used in drivers/usb/*.
It is equivalent to mdelay().

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

USB: introduce usb_interface/usb_configuration structs ...

Currently we have two conflicting definitions of struct usb_config_descriptor
and struct usb_interface_descriptor in the tree. This is because the USB code
uses additional fields in the structs for internal housekeeping. Add
struct usb_interface and struct struct usb_configuration with the housekeeping
data and embed the corresponding hardware structs into them. This frees the
way to use the definitions from ch9.h in the next step.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

Treewide: remove address of the Free Software Foundation ...

The FSF address has changed in the past. Instead of updating it
each time the address changes, just drop it completely treewide.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>