rockbox/bootloader/ipod.c

439 lines
13 KiB
C
Raw Normal View History

/***************************************************************************
* __________ __ ___.
* Open \______ \ ____ ____ | | _\_ |__ _______ ___
* Source | _// _ \_/ ___\| |/ /| __ \ / _ \ \/ /
* Jukebox | | ( <_> ) \___| < | \_\ ( <_> > < <
* Firmware |____|_ /\____/ \___ >__|_ \|___ /\____/__/\_ \
* \/ \/ \/ \/ \/
* $Id$
*
* Copyright (C) 2005 by Dave Chapman
*
* Based on Rockbox iriver bootloader by Linus Nielsen Feltzing
* and the ipodlinux bootloader by Daniel Palffy and Bernard Leach
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version 2
* of the License, or (at your option) any later version.
*
* This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
* KIND, either express or implied.
*
****************************************************************************/
#include "config.h"
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <stdarg.h>
#include "cpu.h"
#include "system.h"
#include "lcd.h"
#include "../kernel-internal.h"
#include "ata.h"
#include "fat.h"
Rewrite filesystem code (WIP) 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>
2013-08-06 02:02:45 +00:00
#include "file_internal.h"
#include "disk.h"
#include "font.h"
#include "adc.h"
#include "backlight.h"
#include "panic.h"
#include "power.h"
#include "file.h"
#include "common.h"
#include "rb-loader.h"
#include "loader_strerror.h"
#include "hwcompat.h"
#include "usb.h"
#include "version.h"
#define XSC(X) #X
#define SC(X) XSC(X)
/* Maximum allowed firmware image size. The largest known current
(December 2006) firmware is about 7.5MB (Apple's firmware for the ipod video)
so we set this to 8MB. */
#define MAX_LOADSIZE (8*1024*1024)
/* A buffer to load the Linux kernel or Rockbox into */
unsigned char *loadbuffer = (unsigned char *)DRAM_START;
#if CONFIG_KEYPAD == IPOD_4G_PAD && !defined(IPOD_MINI)
/* check if number of seconds has past */
int timer_check(int clock_start, unsigned int usecs)
{
if ((USEC_TIMER - clock_start) >= usecs) {
return 1;
} else {
return 0;
}
}
static void ser_opto_keypad_cfg(int val)
{
int start_time;
GPIOB_ENABLE &=~ 0x80;
outl(inl(0x7000c104) | 0xc000000, 0x7000c104);
outl(val, 0x7000c120);
outl(inl(0x7000c100) | 0x80000000, 0x7000c100);
GPIOB_OUTPUT_VAL &=~ 0x10;
GPIOB_OUTPUT_EN |= 0x10;
start_time = USEC_TIMER;
do {
if ((inl(0x7000c104) & 0x80000000) == 0) {
break;
}
} while (timer_check(start_time, 1500) != 0);
outl(inl(0x7000c100) & ~0x80000000, 0x7000c100);
GPIOB_ENABLE |= 0x80;
GPIOB_OUTPUT_VAL |= 0x10;
GPIOB_OUTPUT_EN &=~0x10;
outl(inl(0x7000c104) | 0xc000000, 0x7000c104);
outl(inl(0x7000c100) | 0x60000000, 0x7000c100);
}
int opto_keypad_read(void)
{
int loop_cnt, had_io = 0;
for (loop_cnt = 5; loop_cnt != 0;)
{
int key_pressed = 0;
int start_time;
unsigned int key_pad_val;
ser_opto_keypad_cfg(0x8000023a);
start_time = USEC_TIMER;
do {
if (inl(0x7000c104) & 0x4000000) {
had_io = 1;
break;
}
if (had_io != 0) {
break;
}
} while (timer_check(start_time, 1500) != 0);
key_pad_val = inl(0x7000c140);
if ((key_pad_val & ~0x7fff0000) != 0x8000023a) {
loop_cnt--;
} else {
key_pad_val = (key_pad_val << 11) >> 27;
key_pressed = 1;
}
outl(inl(0x7000c100) | 0x60000000, 0x7000c100);
outl(inl(0x7000c104) | 0xc000000, 0x7000c104);
if (key_pressed != 0) {
return key_pad_val ^ 0x1f;
}
}
return 0;
}
#endif
static int key_pressed(void)
{
unsigned char state;
#if CONFIG_KEYPAD == IPOD_4G_PAD
#ifdef IPOD_MINI /* mini 1G only */
state = GPIOA_INPUT_VAL & 0x3f;
if ((state & 0x10) == 0) return BUTTON_LEFT;
if ((state & 0x2) == 0) return BUTTON_MENU;
if ((state & 0x4) == 0) return BUTTON_PLAY;
if ((state & 0x8) == 0) return BUTTON_RIGHT;
#else
state = opto_keypad_read();
if ((state & 0x4) == 0) return BUTTON_LEFT;
if ((state & 0x10) == 0) return BUTTON_MENU;
if ((state & 0x8) == 0) return BUTTON_PLAY;
if ((state & 0x2) == 0) return BUTTON_RIGHT;
#endif
#elif (CONFIG_KEYPAD == IPOD_3G_PAD) || (CONFIG_KEYPAD == IPOD_1G2G_PAD)
state = GPIOA_INPUT_VAL;
if ((state & 0x08) == 0) return BUTTON_LEFT;
if ((state & 0x10) == 0) return BUTTON_MENU;
if ((state & 0x04) == 0) return BUTTON_PLAY;
if ((state & 0x01) == 0) return BUTTON_RIGHT;
#endif
return 0;
}
bool button_hold(void)
{
#if CONFIG_KEYPAD == IPOD_1G2G_PAD
return (GPIOA_INPUT_VAL & 0x20);
#else
return !(GPIOA_INPUT_VAL & 0x20);
#endif
}
void fatal_error(void)
{
extern int line;
bool holdstatus=false;
/* System font is 6 pixels wide */
#if defined(IPOD_1G2G) || defined(IPOD_3G)
printf("Insert Firewire cable, or");
printf("hold MENU+PLAY to reboot");
printf("then REW+FF for disk mode");
#elif LCD_WIDTH >= (30*6)
printf("Insert USB cable, or");
printf("hold MENU+SELECT to reboot");
printf("then SELECT+PLAY for disk mode");
#else
printf("Insert USB cable, or");
printf("hold MENU+SELECT to");
printf("reboot then SELECT+PLAY");
printf("for disk mode");
#endif
lcd_update();
usb_init();
while (1) {
if (button_hold() != holdstatus) {
if (button_hold()) {
holdstatus=true;
lcd_puts(0, line, "Hold switch on!");
} else {
holdstatus=false;
lcd_puts(0, line, " ");
}
lcd_update();
}
if (usb_detect() == USB_INSERTED) {
ata_sleepnow(); /* Immediately spindown the disk. */
sleep(HZ*2);
#if CONFIG_CPU == PP5020
memcpy((void *)0x40017f00, "diskmode\0\0hotstuff\0\0\1", 21);
#elif CONFIG_CPU == PP5022
memcpy((void *)0x4001ff00, "diskmode\0\0hotstuff\0\0\1", 21);
#elif CONFIG_CPU == PP5002
memcpy((void *)0x40017f00, "diskmodehotstuff\1", 17);
#endif /* CONFIG_CPU */
system_reboot(); /* Reboot */
}
udelay(100000); /* 100ms */
}
}
/* The bootloader is started from the OSOS image on the firmware
* partition. There are several ways it can be installed there:
* appended to the Apple firmware, on its own, or appended to
* Rockbox itself. The Apple ROM loader loads the entire OSOS
* image to DRAM_START, whatever it contains. If the bootloader
* is appended to another image then it will've modified the
* entry point in the OSOS header such that the ROM will call the
* bootloader rather than the main image.
*
* So, once the bootloader has control:
*
* 1) If the hold switch is on, or the menu button is being held,
* try to boot the Apple firmware.
* 1a) First, it looks for apple_os.ipod on the FAT32 partition,
* in .rockbox or the root directory. If found it loads that
* without further checking and runs it.
* 1b) Next, it checks to see if the OSOS image already loaded
* into RAM is in fact the Apple firmware with the bootloader
* appended. It looks at DRAM_START+0x20 for the string
* "portalplayer", and if it's there, just jumps back to
* DRAM_START where the entry point was before the bootloader
* was appended.
* 1c) If neither of those worked, it displays an error and dies.
*
* 2) If the play button is being held, try to boot Linux. It looks
* for linux.bin in the root directory, and if it's not there,
* it displays an error and dies.
*
* 3) Otherwise, try to boot Rockbox.
* 3a) First, it looks for rockbox.ipod on the FAT32 partition,
* in .rockbox or the root directory. If found it loads that
* without further checking and runs it.
* 3b) Next, it checks to see if the OSOS image already loaded
* into RAM is in fact Rockbox with the bootloader appended.
* It looks at DRAM_START+0x20 for the string "Rockbox\1"
* (which is inserted there in crt0-pp.S), and if it's there,
* just humps back to DRAM_START where the entry point was
* before the bootloader was appended.
* 3c) If neither of those worked, it displays an error and dies.
*
* The result is that any of the three install configurations work,
* and that images of apple_os.ipod or rockbox.ipod on the FAT32
* partition take priority over the contents of OSOS (this avoids
* upgrades failing to work if OSOS is not updated).
*
* Loading from OSOS is somewhat faster than loading from FAT32,
* because the Apple ROM doesn't have to deal with filesystems or
* fragmentation, and is already loading from OSOS anyway. Thus,
* the fastest boot configuration that still allows dual booting
* is to install Rockbox into OSOS with the bootloader appended
* (and delete/rename rockbox.ipod from the FAT32 partition).
*
* It is of course faster to just install Rockbox to OSOS alone,
* but then it's impossible to boot the Apple firmware.
*/
void* main(void)
{
char buf[256];
int i;
int btn;
int rc;
bool haveramos;
bool button_was_held;
Rewrite filesystem code (WIP) 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>
2013-08-06 02:02:45 +00:00
struct partinfo pinfo;
unsigned short* identify_info;
/* Check the button hold status as soon as possible - to
give the user maximum chance to turn it off in order to
reset the settings in rockbox. */
button_was_held = button_hold();
system_init();
kernel_init();
#ifndef HAVE_BACKLIGHT_INVERSION
backlight_init(); /* Turns on the backlight */
#endif
lcd_init();
font_init();
#ifdef HAVE_LCD_COLOR
lcd_set_foreground(LCD_WHITE);
lcd_set_background(LCD_BLACK);
lcd_clear_display();
#endif
#if 0
/* ADC and button drivers are not yet implemented */
adc_init();
button_init();
#endif
btn=key_pressed();
/* Enable bootloader messages */
if (btn==BUTTON_RIGHT)
verbose = true;
lcd_setfont(FONT_SYSFIXED);
printf("Rockbox boot loader");
printf("Version: %s", rbversion);
printf("IPOD version: 0x%08x", IPOD_HW_REVISION);
i=ata_init();
if (i==0) {
identify_info=ata_get_identify();
/* Show model */
for (i=0; i < 20; i++) {
((unsigned short*)buf)[i]=htobe16(identify_info[i+27]);
}
buf[40]=0;
for (i=39; i && buf[i]==' '; i--) {
buf[i]=0;
}
printf(buf);
} else {
printf("ATA: %d", i);
}
Rewrite filesystem code (WIP) 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>
2013-08-06 02:02:45 +00:00
filesystem_init();
rc = disk_mount_all();
if (rc<=0)
{
printf("No partition found");
fatal_error();
}
Rewrite filesystem code (WIP) 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>
2013-08-06 02:02:45 +00:00
disk_partinfo(1, &pinfo);
printf("Partition 1: 0x%02x %ld sectors",
Rewrite filesystem code (WIP) 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>
2013-08-06 02:02:45 +00:00
pinfo.type, pinfo.size);
if (button_was_held || (btn==BUTTON_MENU)) {
/* If either the hold switch was on, or the Menu button was held, then
try the Apple firmware */
printf("Loading original firmware...");
/* First try an apple_os.ipod file on the FAT32 partition
(either in .rockbox or the root)
*/
rc=load_firmware(loadbuffer, "apple_os.ipod", MAX_LOADSIZE);
if (rc > 0) {
printf("apple_os.ipod loaded.");
return (void*)DRAM_START;
} else if (rc == EFILE_NOT_FOUND) {
/* If apple_os.ipod doesn't exist, then check if there is an Apple
firmware image in RAM */
haveramos = (memcmp((void*)(DRAM_START+0x20),"portalplayer",12)==0);
if (haveramos) {
/* We have a copy of the retailos in RAM, lets just run it. */
return (void*)DRAM_START;
}
} else {
printf("Error!");
printf("Can't load apple_os.ipod:");
printf(loader_strerror(rc));
}
/* Everything failed - just loop forever */
printf("No RetailOS detected");
} else if (btn==BUTTON_PLAY) {
printf("Loading Linux...");
rc=load_raw_firmware(loadbuffer, "/linux.bin", MAX_LOADSIZE);
if (rc <= EFILE_EMPTY) {
printf("Error!");
printf("Can't load linux.bin:");
printf(loader_strerror(rc));
} else {
return (void*)DRAM_START;
}
} else {
printf("Loading Rockbox...");
rc=load_firmware(loadbuffer, BOOTFILE, MAX_LOADSIZE);
if (rc > 0) {
printf("Rockbox loaded.");
return (void*)DRAM_START;
} else if (rc == EFILE_NOT_FOUND) {
/* if rockbox.ipod doesn't exist, then check if there is a Rockbox
image in RAM */
haveramos = (memcmp((void*)(DRAM_START+0x20),"Rockbox\1",8)==0);
if (haveramos) {
/* We have a copy of Rockbox in RAM, lets just run it. */
return (void*)DRAM_START;
}
}
printf("Error!");
printf("Can't load " BOOTFILE ": ");
printf(loader_strerror(rc));
}
/* If we get to here, then we haven't been able to load any firmware */
fatal_error();
/* We never get here, but keep gcc happy */
return (void*)0;
}