Included are drivers for buttons, backlight, lcd, audio and storage.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@31000 a1c6a512-1295-4272-9138-f99709370657
Move to plugin_crt0.c, plugins don't need PLUGIN_IRAM_* macros anymore
IRAM is no longered zeroed before copying (as it is at the same address
than BSS) -> Fix FS#11581
Use cpucache_invalidate() (and not cpucache_flush), needed for self-modifying code on cached IRAM
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@27948 a1c6a512-1295-4272-9138-f99709370657
It handles exit() properly, calling the handler also when the plugin returns
normally (also make exit() more standard compliant while at it).
It also holds PLUGIN_HEADER, so that it doesn't need to be in each plugin anymore.
To work better together with callbacks passed to rb->default_event_handler_ex() introduce exit_on_usb() which will call the exit handler before showing the usb screen and exit() after it.
In most cases rb->default_event_handler_ex() was passed a callback which was manually called at all other return points. This can now be done via atexit().
In future plugin_crt0.c could also handle clearing bss, initializing iram and more.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@27873 a1c6a512-1295-4272-9138-f99709370657
It handles exit() properly, calling the handler also when the plugin returns
normally (also it makes exit() more standard compliant while at it).
It also holds PLUGIN_HEADER, so that it doesn't need to be in each plugin anymore.
To work better together with callbacks passed to rb->default_event_handler_ex introduce exit_on_usb() which will call the exit handler before showing the usb screen and exit() after it.
In most cases it was passed a callback which was manually called at all other return points. This can now be done via atexit().
In future plugin_crt0.c could also handle clearing bss, initializing iram and more.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@27862 a1c6a512-1295-4272-9138-f99709370657
- reduce the size of some structs -> gain ~1kB
- since we run the code only once, we don't need to declare large
structure in bss and copy const data in them, declare them const
directly -> gain ~17kB
- steal audio buffer for encoded data buffer -> gain 64kB
this will stop playback when encoding, but it was stopped already on
targets using IRAM
- some const policy
Tested on Clipv1 & sim, 83012 bytes smaller on clipv1
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@26134 a1c6a512-1295-4272-9138-f99709370657
* Move strncpy() from core to the pluginlib
* Introduce strlcpy() and use that instead in most places (use memcpy in a few) in core and some plugins
* Drop strncpy() from the codec api as no codec used it
* Bump codec and plugin api versions
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@21863 a1c6a512-1295-4272-9138-f99709370657
struct plugin_api *rb is declared in PLUGIN_HEADER, and pointed to by
__header.api
the loader uses this pointer to initialize rb before calling entry_point
entry_point is no longer passed a pointer to the plugin API
all plugins, and pluginlib functions, are modified to refer to the
global rb
pluginlib functions which only served to copy the API pointer are
removed
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@19776 a1c6a512-1295-4272-9138-f99709370657
Use unmodified C200 keymap when possible, variation of it when not
Use Archos bitmaps when possible
Current problems:
- Pegbox plugin shows display corruption (corruption is different in the simulator)
- Bubbles plugin misses 1 or 2 bottom lines, but if we gain back these lines it occupies a too small area of the screen
- Splitted screen (yellow/blue) is a bit weird in most plugins
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@19252 a1c6a512-1295-4272-9138-f99709370657
breaks 64bit archs (like simulators) but also doing >8 bit reads from memory
addresses with assumptions on alignment without that being really sure seems a
bit error-prone too. So we use memcmp() to compare 4 bytes strings...
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@18356 a1c6a512-1295-4272-9138-f99709370657
