The filesystem API often passes in unaligned receive buffers, and some
code (eg BMP reader) processes data in-place, leading to data loss when
we dropped the cache.
(And document exactly what we're doing, so we don't go through this again
at $future_date)
Change-Id: If47a7f2148a5a1a43777f0bd3be1bdfe8239e91e
Since that encompasses _all_ of our native targets in a post-archos world,
either replace it with #if (CONFIG_PLATFORM & PLATFORM_NATIVE) or
delete it altogher as appropriate.
Change-Id: I9128a456e850d5c96a9e05806aad3acd923f90c5
Discovered on ClipZip but applicable to other LOW MEM targets
Back off voice buffer on playlist_resume
Change-Id: Ia316c4b7fc1bcb089d3069a13dd7c193edf2ba1e
In fixing the original bug I tried to optimize discard_dcache_range()
to minimize writeback and inadvertently introduced a second bug, which
typically ends in a TLB refill panic.
It occurs only if the range fits within one cache line, and when both
the start and end of the range are not aligned to a cache line. This
causes ptr to be incremented and end to be decremented, so ptr > end,
and the loop can't terminate.
Change-Id: Ibaac072f1369268d3327d534ad08ef9dcee3db65
- The range-based cache operations on MIPS were broken and only worked
properly when BOTH the address and size were multiples of the cache
line size. If this was not the case, the last cache line of the range
would not be touched!
Fix is to align start/end pointers to cache lines before iterating.
- To my knowledge all MIPS processors have a cache, so I enabled
HAVE_CPU_CACHE_ALIGN by default. This also allows mmu-mips.c to use
the CACHEALIGN_UP/DOWN macros.
- Make jz4760/system-target.h define its cache line size properly.
Change-Id: I1fcd04a59791daa233b9699f04d5ac1cc6bacee7
Open plugin needs the context menu to allow the user to add an item
on the first run there are no items and therefore nothing to pop the
context menu from
Change-Id: I7a78454ff29b20c7b60db44349dd23d500ed887c
Some flac encoded files contain junk that our decoder
picked up
upstream has some sign and overflow fixes too
Change-Id: I5857b2fe56906a48f04944cdfee8fe2306f2c3fd
I think the assumption is that during playback, the DSP is already running
so it assumes PBE needs to be flushed before it is properly enabled.
Change-Id: I2bac3d02c80f97c8d9ce26a575175f6344a8e86a
0ce42df prevented read_color_theme_file() from being
executed when applying settings (from a cfg file)
if the filetype colors setting was set to "-". Resetting
custom filetype colors happens in that function though and
thus wouldn't work. You had to restart Rockbox for the
setting to be applied (at which point read_color_theme_file()
is called from filetype_init() which is called by the init()
function in main.c)
The safety check has been moved from settings.c directly into
read_color_theme_file(). This has the added
advantage that it will also apply during the init process
where there existed no check before (prevents the problem
mentioned in 2e3a8c7).
Change-Id: I547fe180f0bac79889bf2c44778bb1cc4f9d4307
This is essentially an expanded version of jz4760_tools/usbboot,
able to support both X1000 and JZ4760 CPUs and easily extended to
handle other Ingenic CPUs using the same boot protocol.
Change-Id: I70ce3acc3531d65390c6bbae4d2b3352140acf0a
When playback is stopped and the user inserts songs, Rockbox will not erase a dynamic playlist anymore, unless it has finished playing.
Change-Id: I50bd62fbfe3929a0b651b5136e9305b9bc9278f3
It seems that 'toolset=' in tools/configure was set but not used, with
the build system only generating a hardcoded set of tools. This change
fixes the issue by exporting the required tools in the TOOLSET variable,
and having root.make append those to the TOOLS list.
Change-Id: I07024bdf1dd2260c32cfeecaba0d3bb5bf00346d
If the custom UI viewport (for displaying lists) changes size when entering dirbrowse, other viewports won't be correctly rendered and require a redraw.
The following is a minimal test case for an example SBS where (the specified background color for) the viewport at the top of the screen won't appear after the user has entered the (root menu of the) file or database browser:
%?if(%cs,=,1)<%VI(main)|%VI(other)>
%V(0,0,-,21,-)
%Vb(ededed)
%Vi(main,0,22,-,-,-)
%Vi(other,0,60,-,-,-)
Change-Id: I1aeed0561f16531802d0fb8dc5fd18d65ac8f25a
ATA DMA was enabled for all PP502x targets in d118f47 after previously reported instabilities were thought to have been fixed. The iPod 4G target remains unstable when UDMA 2 is enabled. File system corruption will eventually occur even using stock hardware in normal usage, according to both my own experience and that of several other forum users. UDMA 1 appears to be stable.
Change-Id: I8526bad9e879f5dad5174cfe07cd8828d8b72406
Reshuffling a playlist using the WPS context menu for the current playlist is a destructive operation since the previous order of items can't be restored. A warning should be displayed, so that the user can still cancel the operation in case they intended to select another menu item (e.g. one of the harmless "Playing Time" or "Save Current Playlist" items right next to it).
Change-Id: Ib9fc379439d155be907872d77cf54d5035bd08a0
Freetype-config has been deprecated and is no longer included in libfreetype-dev. Pkg-config serves as its replacement.
Convttf isn't built as part of Rockbox by default, but manually doing so will currently fail even when libfreetype-dev is installed.
Change-Id: Iafc31df3b40c11d7c790189eb9d3788ba97d3e54
mp4 files can have multiple 'mdat' chunks. This is common for
audiobooks, where there is often a secondary mdat containing the
chapter names, but it's also legal to have multiple mdat chunks
for a single logical "track"
This confuses the mp4 metadata parser, which assumes there is
only a single mdat, and always uses the last mdat seen to
determine the "filesize" of the data we're trying to decode.
Work around this by appending each mdat's size to result in the final
"filesize"
Change-Id: I3e7a7efb0f05ef965e8d77f79e450c957524a480
After using “Insert Next” to insert multiple songs at once (e.g. an album from the database browser or folder from the file browser), subsequent Insert operations will incorrectly insert items after the first song of all items that were previously inserted, instead of after the last song of the previously inserted items.
A bug fix was originally written by Costas Calamvokis for the file browser only. I adopted the original fix and added code analogous to it so that it works from the database browser as well.
See FS#7898, FS#7363 or this forum post for more info:
https://forums.rockbox.org/index.php/topic,53741.0.html
Change-Id: Ie2718e136df0b340000f7a171e9e806cf23a27b4
Rockbox would unexpectedly and without warning erase the current dynamic playlist (when it hadn’t finished playing) if the user inserted or queued up items while playback was stopped.
This fix ensures that the playlist is only deleted if it has either finished playing or can't be resumed.
Change-Id: I73b7bd56e6399f9be4bb000bae39c8b6a5e1b79b
Music currently doesn't begin playing after the user has selected a new song in a playlist, unless another item was already playing or the previous player state was "stopped" (but not "paused"). This results in an inconsistent user experience. The user's likely intention is to immediately listen to the song they select for playback regardless of the previous player state.
Change-Id: I68a8da01b06a81d8c3b61e351710431152bfdfaa
I wanted to build into a 'toolchain' directory like this:
$ tools/rockboxdev.sh --dlwhere=toolchain --builddir=toolchain/build --prefix=toolchain/bin
This failed twice:
a) the extraction stage could not uncompress because of the relative
paths, so convert them to absolute paths in the script to fix the issue.
b) patches could not be applied because 'rockboxdevdir' was assigned a
string with the directory twice and '\n' in the middle. Using readlink,
this now works.
Change-Id: I9c7c30812e1fb4b58db7e6561d73678f34714c91
I hit this error and its message confused me because I interpreted it as
"bzip2 executable could not be found". The message was also outdated
because xz is supported meanwhile. Update the message to something
hopefully more helpful.
Change-Id: Ide2f8809dc0f8159d7d04fa22c29607dd6b35792
'SUBDIRS' file has puzzles inside a 'HAVE_LCD_COLOR' block. I found this
issue because sgt-puzzles was in the Clip+ docs while they were not
installed on the target.
Change-Id: I8ddf43fa75c8429d2f1a99f13689a5442913748e