It seems connecting an iPod Video to a Mac triggers the null
request check, resulting in a panic. Ignoring the error with
a bare return "fixes" it and allows the iPod to connect. This
isn't ideal though, because it could silently introduce bugs
on other targets.
The likely cause of this is the host sending control requests
too fast, or a driver problem (the Video uses the ARC driver,
which is still on the legacy interface), with multiple requests
getting queued at once. Since the USB core expects to deal with
only one request at a time, the second response trips the check.
Try to handle this situation a bit more gracefully by detecting
overlapped requests and returning a STALL to the host when it
occurs. At this point the USB stack is able to safely handle a
new request.
Link: https://forums.rockbox.org/index.php/topic,54414.0.html
Change-Id: I9a2b7e35620ff540ebdb39f81671377062a4917d
allow buflib_free to check for invalid or already freed handles
within the function -- remove all the invalid handle guards thru core_free
Change-Id: Ibdcbc82760fc93b674c42283fca420d94907df8e
There are various allocations that can't be moved or shrunk.
Provide a global callback struct for this use case instead of
making each caller declare its own dummy struct.
Also fixed ROLO and x1000 installer code which incorrectly
used movable allocations.
Change-Id: I00088396b9826e02e69a4a33477fe1a7816374f1
Due to how the old control request API worked, the request
pointer (and the pointed-to contents) passed by the driver
must remain valid for the lifetime of the request.
The buffered copy wasn't used consistently anyway, so it
should be safe to just get rid of it.
(This only affects the old API compatibility layer.)
Change-Id: I00294c718a7515a91b84f7c7cae5220c73aa5960
In the old API, usb_core_transfer_complete() would ignore
any completions on EP0. The legacy control API will also
ignore most completions on EP0, but it only did so after
filling in the completion event.
If the driver submits a spurious completion on EP0, then
this could clobber a previously queued completion event.
Fix this by deciding whether to ignore the completion
*before* modifying the event data.
Change-Id: I68526898a96efa594ada4e94bb6851aaaf12e92a
The address from the packet needs to be saved before sending
the response -- after the response the request being pointed
to could get overwritten. This used to be done correctly but
I unintentionally broke it when updating the handler to the
new control request API.
Change-Id: I9b11548baf20dce44a82301731405dc8e8f1cef3
Because usb_core_legacy_control_request() is called by an
interrupt handler the global variables used to track the
current control request at the very least need to be volatile.
It might also be necessary to disable IRQs but I'm not sure
of that.
Change-Id: I0f0bb86d0ad63734e8d3c73cb1c52fedf5a98120
When performing an OUT transfer which is not a multiple of the
max packet size, the last packet of the OUT transfer should be
a short packet. However, there's no guarantee the host sends
the expected amount of data in the final packet.
The DWC2 USB controller handles this case by accepting any size
packet and copying it out to memory. So if the packet is bigger
than expected, it'll overrun the caller's buffer and Bad Things
will happen.
The USB 2.0 spec seems to endorse this behavior. Section 8.5.1
says "an ACK handshake indicates the endpoint has space for a
wMaxPacketSize data payload." So it is possible that other USB
controllers share the DWC2's behavior.
The simplest solution is to force all USB RX buffers to be big
enough to hold the transfer size, rounded up to a multiple of
the max packet size. For example, a transfer of 700 bytes would
require a 1024-byte buffer if the MPS = 512 bytes.
Change-Id: Ibb84d2b2d53aec8800a3a7c2449f7a17480acbcf
All existing USB drivers now define USB_LEGACY_CONTROL_API to
enable the emulation layer.
Control request handlers will be ported in follow-up commits.
Change-Id: I4be1ce7c372f2f6fee5978a4858c841b72e77405
This makes it possible for macros of conditionally included string
descriptors to get a correct index no matter what other usb drivers
are enabled or disabled due to the nature behavior of enums.
Change-Id: I8ccebbd316605bed0f5d90b6b73fab4a333c02fa
Upon getting a USB reset, the USB core will update charging
current by calling usb_charging_maxcurrent_change(). On all
current X1000 targets this may cause a hang, since changing
the charge current involves a blocking I2C transaction.
Eg. if the host issues a reset when we're already configured
as part of error recovery, the change from 500 mA -> 100 mA
will cause a hang.
Change-Id: I5b45272c01fa16b179ae3d16bbc50c7fab9a416b
IMHO the current name is somewhat misleading:
- usb_drv_send() is blocking and we have usb_drv_send_nonblocking()
for the non-blocking case. This inconsistent naming can only
promote confusion. (And what would we call a blocking receive?)
- Other hardware abstraction APIs in Rockbox are usually blocking:
storage, LCD, backlight, audio... in other words, blocking is the
default expected behavior, with non-blocking calls being a rarity.
Change-Id: I05b41088d09eab582697674f4f06fdca0c8950af
The funny thing is that the memset() for the field in question was
redundant, as the overall inquiry structure was memset(0) already.
Change-Id: I8bec0c93c9317823ff39cf7133535e3bf58fb987
Atmel AT88SC6416C CryptoMemory is almost I2C compatible. The device
is connected to bitbanged I2C bus shared with compliant I2C devices.
Change-Id: Iec54702db1bdfb93c01291eef18ec60391c63b16
On devices that can assign interrupt IN, bulk IN and bulk OUT endpoints
this change results in the serial interface working out of the box on
Linux and Windows. On Linux it is registered as ttyACM device and on
Windows it is assigned a COM port number.
On devices that cannot assign the interrupt IN this change won't have
any effect.
Implement minimum required interface control requests. Respond with
whatever line coding was set to make terminal programs happy.
Change-Id: Id7d3899d8546e45d7cb4ecc3fe464908cb59e810
They were never finished, never saw any release ever, and haven't
compiled for the better part of a decade. Given their HW capabilities [1],
they are not worth trying to fix.
[1] 1-2MB RAM, ~256MB onboard flash, no expandability
Change-Id: I7b2a5806d687114c22156bb0458d4a10a9734190
This uses the new unicode string literal feature that is available
now to greatly simplify the initialization of these special string
types. This makes them much more readable at a quick glance.
Change-Id: Iad8b49aa763486608e3bb7e83fb8abfb48ce0a7b
This means we will no longer send them routinely and instead rely on
the HID driver to send them when the host requests it. This also
moves the reporting out of the power management code where it probably
did not belong in the first place.
Change-Id: I9c8420e81897f1f6caaa55ffacc7525589f1ef75
This allows rockbox to report its battery status through the
HID Battery Strength method that is available through the
Device Controls usage page.
Change-Id: Ia7a7dd9b9d476dd9df5a5f5becabc5ae823e9a89
For s5l8701, s5l8702 and as3525v2 targets.
The crash occurs when USB is inserted or extracted while the playlist
is being loaded or updated (it could take a few seconds for huge
playlists), at this point all buflib memory is allocated and not freed
before usb_starage_init_connection() is executed.
This workaround mitigates this panic by using static memory for USB
buffers, so this memory cannot be used for other tasks, in addition
the problem still persist when playlist load is 'paused' by USB
insertion and then updated after USB extraction.
Change-Id: Iff1db5a949361fd543e0b494924d1f2906c84b5e
Many includes of fat.h are pointless. Some includes are just for
SECTOR_SIZE. Add a file 'firmware/include/fs_defines.h' for that
and to define tuneable values that were scattered amongst various
headers.
Remove some local definitions of SECTOR_SIZE since they have to be
in agreement with the rest of the fs code anyway.
(We'll see what's in fact pointless in a moment ;)
Change-Id: I9ba183bf58bd87f5c45eba7bd675c7e2c1c18ed5
Based on g#844 and g#949, it is intended as a replacement for the
current s3c6400x USB driver.
The DesignWare USB OTG core is integrated into many SoC's, however
HW core version and capabilities (mainly DMA mode, Tx FIFO mode,
FIFO size and number of available IN/OUT endpoins) may differ:
CPU targets HW ver DMA NPTX FIFO FIFO sz #IN/OUT
-------- ------------- ------ --- --------- ------- -------
as3525v2 sansaclipplus 2.60a Yes Dedicated 0x535 4/4
sansaclipv2
sansaclipzip
sansafuzev2
s5l8701 ipodnano2g 2.20a Yes Shared 0x500 4/5
s5l8702 ipod6g 2.60a Yes Dedicated 0x820 7/7
ipodnano3g
s5l8720 ipodnano4g ? ? ? ? ?
Functionality supported by this driver:
- Device mode, compatible with USB 1.1/2.0 hosts.
- Shared FIFO (USB_DW_SHARED_FIFO) or dedicated FIFOs.
- No DMA (USB_DW_ARCH_SLAVE) or internal DMA mode.
- Concurrent transfers: control, bulk (usb_storage, usb_serial) and
interrupt (usb_hid).
Actually this driver is not used by any CPU, it will be enabled for
each individual CPU/target in next patches.
Change-Id: I74a1e836d18927a31f6977d71115fb442477dd5f
This patch redoes the filesystem code from the FAT driver up to the
clipboard code in onplay.c.
Not every aspect of this is finished therefore it is still "WIP". I
don't wish to do too much at once (haha!). What is left to do is get
dircache back in the sim and find an implementation for the dircache
indicies in the tagcache and playlist code or do something else that
has the same benefit. Leaving these out for now does not make anything
unusable. All the basics are done.
Phone app code should probably get vetted (and app path handling
just plain rewritten as environment expansions); the SDL app and
Android run well.
Main things addressed:
1) Thread safety: There is none right now in the trunk code. Most of
what currently works is luck when multiple threads are involved or
multiple descriptors to the same file are open.
2) POSIX compliance: Many of the functions behave nothing like their
counterparts on a host system. This leads to inconsistent code or very
different behavior from native to hosted. One huge offender was
rename(). Going point by point would fill a book.
3) Actual running RAM usage: Many targets will use less RAM and less
stack space (some more RAM because I upped the number of cache buffers
for large memory). There's very little memory lying fallow in rarely-used
areas (see 'Key core changes' below). Also, all targets may open the same
number of directory streams whereas before those with less than 8MB RAM
were limited to 8, not 12 implying those targets will save slightly
less.
4) Performance: The test_disk plugin shows markedly improved performance,
particularly in the area of (uncached) directory scanning, due partly to
more optimal directory reading and to a better sector cache algorithm.
Uncached times tend to be better while there is a bit of a slowdown in
dircache due to it being a bit heavier of an implementation. It's not
noticeable by a human as far as I can say.
Key core changes:
1) Files and directories share core code and data structures.
2) The filesystem code knows which descriptors refer to same file.
This ensures that changes from one stream are appropriately reflected
in every open descriptor for that file (fileobj_mgr.c).
3) File and directory cache buffers are borrowed from the main sector
cache. This means that when they are not in use by a file, they are not
wasted, but used for the cache. Most of the time, only a few of them
are needed. It also means that adding more file and directory handles
is less expensive. All one must do in ensure a large enough cache to
borrow from.
4) Relative path components are supported and the namespace is unified.
It does not support full relative paths to an implied current directory;
what is does support is use of "." and "..". Adding the former would
not be very difficult. The namespace is unified in the sense that
volumes may be specified several times along with relative parts, e.g.:
"/<0>/foo/../../<1>/bar" :<=> "/<1>/bar".
5) Stack usage is down due to sharing of data, static allocation and
less duplication of strings on the stack. This requires more
serialization than I would like but since the number of threads is
limited to a low number, the tradoff in favor of the stack seems
reasonable.
6) Separates and heirarchicalizes (sic) the SIM and APP filesystem
code. SIM path and volume handling is just like the target. Some
aspects of the APP file code get more straightforward (e.g. no path
hashing is needed).
Dircache:
Deserves its own section. Dircache is new but pays homage to the old.
The old one was not compatible and so it, since it got redone, does
all the stuff it always should have done such as:
1) It may be update and used at any time during the build process.
No longer has one to wait for it to finish building to do basic file
management (create, remove, rename, etc.).
2) It does not need to be either fully scanned or completely disabled;
it can be incomplete (i.e. overfilled, missing paths), still be
of benefit and be correct.
3) Handles mounting and dismounting of individual volumes which means
a full rebuild is not needed just because you pop a new SD card in the
slot. Now, because it reuses its freed entry data, may rebuild only
that volume.
4) Much more fundamental to the file code. When it is built, it is
the keeper of the master file list whether enabled or not ("disabled"
is just a state of the cache). Its must always to ready to be started
and bind all streams opened prior to being enabled.
5) Maintains any short filenames in OEM format which means that it does
not need to be rebuilt when changing the default codepage.
Miscellaneous Compatibility:
1) Update any other code that would otherwise not work such as the
hotswap mounting code in various card drivers.
2) File management: Clipboard needed updating because of the behavioral
changes. Still needs a little more work on some finer points.
3) Remove now-obsolete functionality such as the mutex's "no preempt"
flag (which was only for the prior FAT driver).
4) struct dirinfo uses time_t rather than raw FAT directory entry
time fields. I plan to follow up on genericizing everything there
(i.e. no FAT attributes).
5) unicode.c needed some redoing so that the file code does not try
try to load codepages during a scan, which is actually a problem with
the current code. The default codepage, if any is required, is now
kept in RAM separarately (bufalloced) from codepages specified to
iso_decode() (which must not be bufalloced because the conversion
may be done by playback threads).
Brings with it some additional reusable core code:
1) Revised file functions: Reusable code that does things such as
safe path concatenation and parsing without buffer limitations or
data duplication. Variants that copy or alter the input path may be
based off these.
To do:
1) Put dircache functionality back in the sim. Treating it internally
as a different kind of file system seems the best approach at this
time.
2) Restore use of dircache indexes in the playlist and database or
something effectively the same. Since the cache doesn't have to be
complete in order to be used, not getting a hit on the cache doesn't
unambiguously say if the path exists or not.
Change-Id: Ia30f3082a136253e3a0eae0784e3091d138915c8
Reviewed-on: http://gerrit.rockbox.org/566
Reviewed-by: Michael Sevakis <jethead71@rockbox.org>
Tested: Michael Sevakis <jethead71@rockbox.org>
Windows will try to retrieve such a descriptor on first connect.
If the device returns STALL or a regular string descriptor (i.e.
not one that follows the Microsoft OS Descriptor spec), things
will continue normally.
Unfortunately some of our low-level USB drivers have issues with
STALL so any other valid descriptor is the next best solution.
Change-Id: I59eb09eea157e4e14bec0197a898be378a5559f2
Reviewed-on: http://gerrit.rockbox.org/680
Reviewed-by: Frank Gevaerts <frank@gevaerts.be>
Tested: Frank Gevaerts <frank@gevaerts.be>
usb_storage needs a fairly reasonable amount of memory. Allocating
what we need and no more allows other (future) USB drivers to get
something too, and is much cleaner in general.
Change-Id: Iec9573c0f251f02400f92d92727cbf2969785de0
disconnect() needs to be called exactly once per call to init_connection().
In case of bus resets, disconnect() was not called, which led to leaking
alloc_maximum() allocated buflib handles, which led to buflib running out
of memory to allocate.
Change-Id: I03025da578dc54e48b6de6bd3e3f40feae7220a6
This fixes the radioart crash that was the result of buffering.c working
on a freed buffer at the same time as buflib (radioart uses buffering.c for the
images). With this change the buffer is owned by buflib exclusively so this
cannot happen.
As a result, audio_get_buffer() doesn't exist anymore. Callers should call
core_alloc_maximum() directly. This buffer needs to be protected as usual
against movement if necessary (previously it was not protected at all which
cased the radioart crash), To get most of it they can adjust the willingness of
the talk engine to give its buffer away (at the expense of disabling voice
interface) with the new talk_buffer_set_policy() function.
Change-Id: I52123012208d04967876a304451d634e2bef3a33
When using variadic macros there's no need for IF_MD2/IF_MV2 to deal
with function parameters. IF_MD/IF_MV are enough.
Throw in IF_MD_DRV/ID_MV_VOL that return the parameter if MD/MV, or 0
if not.
Change-Id: I7605e6039f3be19cb47110c84dcb3c5516f2c3eb
Currently we don't know where the serial number is stored on the
stmp3600. It is probably using the laser fuses but this needs to
be investigated
Change-Id: I1ac25e38b8f65635abb68788ceb65df0a740dabd
