errno is supposed to be thread local and we do that here in a rather
makeshift way by implementing a function that returns a pointer to
the __errno variable in the thread block.
If more serious TLS is required, perhaps it would be worthwhile
implementing it using the linker and the "__thread" storage attribute.
That's a bit overkill just for this. I'm also not liking what I saw
the compiler producing.
Change-Id: I03bc0bd6a89f6e3d6bae7653284ee01054614f9a
Reviewed-on: http://gerrit.rockbox.org/803
Reviewed-by: Michael Sevakis <jethead71@rockbox.org>
Tested: Michael Sevakis <jethead71@rockbox.org>
The graphical editor can now display and editor description files.
The library has been improved to provide more useful function.
The XML format has been slightly changed: only one soc is allowed per file
(this is was already de facto the case since <soc> was the root tag).
Also introduce a DTD to validate the files.
Change-Id: If70ba35b6dc0242bdb87411cf4baee9597798aac
The write buffer should not be modified but the current code does and then
forget to restore it to its original content. I'm not sure if any code relies
to the write buffer to not be modifies by the write function but this seems like
a reasonable assumption in general so it's better not to break it.
Change-Id: I449a01db2ec51d2273e59b69c59db0e7d2eed3db
Previously %Vs styles only took effect upon the first character of text, i.e.
on the right if the line is %arXXX. Now the style is applied for the whole
line width.
Fixes FS#12973.
Change-Id: I95f2cc76a90bfc0779e5a667d5bfa54c33896c89
The plugin buffer was used only to avoid reparsing the playlist, so non-essential.
But when it was used it conflicted with the playlist viewer which already uses
the plugin buffer for playlist purposes simultaneously. It only works by
accident.
Since the reparse avoidance is non-essential don't do it for now. A temp buffer
can be passed to playlist_save() to enable it but the only caller (as of now)
does not do that.
Change-Id: I3f75f89d8551e1ec38800268b273105faba0efbf
Implicit promotion of integer literals to unsigned long introduced a subtle bug
on 64-bit systems due to weird sign extensions (leads to audible glitches in a
few files). The table is originally designed for unsigned 32bit integers, and
it works with those so use them. As a consequence the lookup table size is
halved as well.
Change-Id: I35d878d6df03300387f0e403e0f3c3bdc73eea00
Scrollstrip works as quadrature encoder. We set pin change
interrupt for both edges of signal both lines and sample the state
in ISR. From time to time the sequence is out-of-sync which was
properly detected but erroneously handled leaving scrollstip
interrupts disabled.
Change-Id: I08e4f99c6c27df0f8180aa16e6e1d9e4203bafa7
Reviewed-on: http://gerrit.rockbox.org/782
Reviewed-by: Marcin Bukat <marcin.bukat@gmail.com>
Tested: Marcin Bukat <marcin.bukat@gmail.com>
- drop support for PP500x: it's very different from other PP and although
it would be possible to support them, I don't have one to test the code
- make sure only the CPU is started
- add PP descriptor to report chip ID and revision
- add code in shell and lua to support pp (no register description yet)
- compile for ARMv4 because PP502x is an ARM7TDMI
Change-Id: I36c4e465dfc2cfdfe7433b2f65cc8f6f0720fe62
Some plugins grab the whole audio buffer and still want to start playback
somehow (e.g. random_folder_advance_config). Since 22e802e the plugin buffer
is allocated via buflib and has to be released explicitely. For these plugins
the automatic free on exit is not sufficient and they need an API function for
that.
Fixes OOM panic on random_folder_advance_config when using start
shuffled playback.
Change-Id: I0d351daa782cb829f4ff80d34c05f40a2e0c142f
It is very similar to how e200tool from MrH works but uses the framework
of hwstub which is makes it completely trivial since we already have the
USB driver written.
Change-Id: I61cdc245d3f828c2682bcd6ecfed5a1cc0094139
This is actually the trivial part of e200tool from MrH: it simply writes the
code on the bulk endpoint. Code was mostly copied from imxtools/sbloader.
Change-Id: I6c208840d23553aaf3bd8b9374e6b0337e54f3b0
This patch completes the plugin keymaps for the Zen X-Fi3 and enables those plugins for compilation.
One key was changed in "button-target.h" for compatibility with Rockboy.
This also caused the changes to "keymap-zenxfi3.c", to keep the stock functionality (no further changes in here).
Change-Id: Ic222faf89e9a9a2332a49d6e532cedb6eb16d3d7
Reviewed-on: http://gerrit.rockbox.org/762
Reviewed-by: Amaury Pouly <amaury.pouly@gmail.com>
Originating from 3661581
Some build clients finding their "standard" string.h's that don't declare
strlcpy?
Change-Id: I50d19c7cecf5ae96ee1855f77d3c2e1f42620108
Apart from the fact that the original settings were much
to sensitive for my taste, they are now easier configurable.
Change-Id: If1772367fc1f34fa1255f57b1831d1f33dc34558
Reviewed-on: http://gerrit.rockbox.org/772
Reviewed-by: Marcin Bukat <marcin.bukat@gmail.com>
It is trivial to obtain all required information from the allocated
handles without maintaining global indexes. In fact, it is less
complicated and increases general thread safety.
Other miscellaneous changes (some are nice to do at this time due to
required alterations, with some particularly more relevant than others):
* Handle value 0 will no longer be returned as a valid handle but all
failures will still return a negative value. Creates consistency with
buflib and removes the need to explicitly initialize them.
* Linking a new handle is delayed until explicitly
added by the code that called add_handle, keeping it invisible
until every operation succeeds, which is safer thread-wise. If anything
fails, the handle itself may just be abandoned rather than reqiring it
be freed.
* Dump the special handling to slow buffering when the PCM buffer
is low that calls PCM buffer functions. It doesn't seem to help much
of anything these days and it's a bit of a nasty hack to directly
tie those bits together. It can of course be put back (again!) if
there really is a need for it.
* Make data waiters ping the buffering thread more than just once if
the request is taking too long. Somehow I figured out how the requests
could get forgotten about but can't remember why months later after
making the change in my branch. :-)
* Neaten up some code by using (inline) functions and packing down
parameters; remember handle allocation and movement attributes in the
handle itself rather than figuring it out each time they're needed.
Change-Id: Ibf863370da3dd805132fc135e0ad104953365183
Reviewed-on: http://gerrit.rockbox.org/764
Reviewed-by: Michael Sevakis <jethead71@rockbox.org>
Tested: Michael Sevakis <jethead71@rockbox.org>
FILE typedef to *void needs more work to not break sim and
application builds. I checked only a few random native builds
unfortunately. Sorry for inconvenience.
Prior to this patch the Lua plugin used version 5.1.4. This change
reduces the number of modifications in the Lua source using some new
defines and because the upstream source is now more flexible.
Unless otherwise stated, l*.[ch] files are taken unmodified from the
upstream lua-5.2.3.
fscanf.c:
file descriptors in rockbox are just ints, they are hidden behind a
void* now so liolib requires less modifications. fscanf is updated to
use void* too.
getc.c: this is a new file required for getc implementation in lauxlib.c
lauxlib.c: LoadF replaced FILE* with int, the rockbox file
descriptor int are cast to FILE* (actually void* due to typedef).
getc uses the PREFIX version. stdin is not used, as per 5.1.4.
lbaselib.c: now uses strspn in the number parsing. print uses DEBUGF now
rather than being commented out.
lbitlib.c: use the built-in version from 5.2.3 rather than Reuben
Thomas's external library. Backwards compatible and adds some new bit
operations.
ldo.c: the LUAI_THROW/TRY defines are now in the core lua code, so have
been removed from rockconf.h
liolib.c: here the implementation has changed to use the LStream from
the original source, and cast the FILE* pointers to int. This has
reduced the number of modifications from the upstream version.
llex.c: the only change from upstream is to remove the locale include.
lmathlib.c: updated from the 5.2.3 version and re-applied the changes
that were made vs 5.1.4 for random numbers and to remove unsupported
float functions.
loadlib.c: upstream version, with the 5.1.4 changes for missing
functions.
lobject.c: upstream version, with ctype.h added and sprintf changed to
snprintf.
loslib.c: upstream version with locale.h removed and 5.1.4 changes for
unsupportable functions.
lstrlib.c: sprintf changed to snprintf.
ltable.c: upstream with the hashnum function from 5.1.4 to avoid frexp
in luai_hashnum.
luaconf.h: updated to 5.2.3 version, restored relevant parts from the
original 5.1.4 configuration. The COMPAT defines that are no longer
available are not included.
lundump.c: VERSION macro conflicts with the core Rockbox equivalent.
rocklib.c: luaL_reg is no longer available, replaced by luaL_Reg
equivalent. Moved checkboolean/optboolean functions to this file and out
of core lua files. luaL_getn is no longer available, replaced by
luaL_rawlen. luaL_register is deprecated, use the newlib/setfuncs
replacements. rli_init has to be called before setting up the newlib to
avoid overwriting the rb table.
rocklib_aux.pl: use rli_checkboolean from rocklib.c.
rocklua.c: new default bits library used, update the library loading
code with idiomatic 5.2 code.
strcspn.c: no longer needed, but strspn.c is required for strspn in
lbaselib.c
Change-Id: I0c7945c755f79083afe98ec117e1e8cf13de2651
Reviewed-on: http://gerrit.rockbox.org/774
Tested: Richard Quirk <richard.quirk@gmail.com>
Reviewed-by: Marcin Bukat <marcin.bukat@gmail.com>
Before, the Cypress chip power is always turn on either from internal battery
or from usb host. In this regard, the internal battery energy was spend on the
inactive chip. Now power of the Cypress chip is enabled only on the external usb host.
Change-Id: Iae7cddd1237ea9ed93fee5404575cc01543cb00c
Reviewed-on: http://gerrit.rockbox.org/433
Reviewed-by: Marcin Bukat <marcin.bukat@gmail.com>
After placing the firmware.mi4 file in the root dir of the player in UMS mode of
the OF, Sansa should do stupid blinking with the backlight and buttonlight
alternately. Recovering from this state is possible through the recovery mode
(see Wiki), by putting an original copy of the firmware.mi4.
Change-Id: Ia913442b97e8c405f55c4676b9a2bf0b1b1d05d6
