OpenCores
URL https://opencores.org/ocsvn/or1k/or1k/trunk

Subversion Repositories or1k

[/] - Rev 1709

Rev

Go to most recent revision

Filtering Options

Clear current filter

Rev Log message Author Age Path
1709 Implement the READ_NATIVE_MAX_ADDRESS command. nogj 6729d 17h /
1708 Precalculate the size of the drive in sectors. Used very often in the code. nogj 6729d 17h /
1707 Allow ata_read_sect() to read more than one sector, reporting status by setting/
clearing appropriate status register bits.
nogj 6729d 17h /
1706 Introduce a helper function ata_calc_lba() to calculate the lba from the device
registers depending on what mode the device is in.
nogj 6729d 17h /
1705 Add a function that is called when the host has transfered all the queued data
from the device.
nogj 6729d 17h /
1704 Clear the _BSY status when commands complete. Set error if an invalid command
was sent.
nogj 6729d 17h /
1703 Introduce a device sub-section into the ata section and avoid confuseing
xx{0,1} parameters
nogj 6729d 17h /
1702 Make the default pio and dma settings be configurable from the config file. nogj 6729d 17h /
1701 Make the OCIDEC type and revision settable via the config file. nogj 6729d 17h /
1700 The ata section of the config file does not have the debug keywork anymore.
Remove it.
nogj 6729d 17h /
1699 Make ata_set_device_signature() take a boolean as its second argument, avoiding
a useless define
nogj 6729d 17h /
1698 Use proper types. nogj 6729d 17h /
1697 Move ata_device_execute_cmd() the the end of atadevice_cmdi.c avoiding
prototypes for the other functions. Mark some stuff static.
nogj 6729d 17h /
1696 Add pretty status register dumping code nogj 6729d 17h /
1695 Move configuration variables to ata_device.conf nogj 6729d 17h /
1694 Make ata_devices have an array of ata_devices instead of device0 and device1. nogj 6729d 17h /
1693 Rearange functions to be able to get rid of forward declarations. Also mark
some stuff static
nogj 6729d 17h /
1692 Instead of playing games with the esp when a jump needs to be executed, use
longjmp() to get to the jump handling code.
nogj 6729d 17h /
1691 Remove useless conditions from enter_dyn_code() nogj 6729d 17h /
1690 Move the main execution loop to execute.c. nogj 6729d 17h /
1689 Move do_scheduler() to sched.c. Doesn't result in any performance loss. nogj 6729d 17h /
1688 Don't inline sched_add/sched_remove. This only results in a minimal performance
loss.
nogj 6729d 17h /
1687 Don't use MMX registers to keep track of the pc and the number of cycles
remaining untill the next job. Keeping the two places that these where used at
concurrent bloated generated code quite a bit and there was no performance
increase resulting from their use.
nogj 6729d 17h /
1686 Remove the immu_ex_from_insn hack. nogj 6729d 17h /
1685 Remove the test_dyn_rec program. It's not that usefull anymore now that the
recompiler more-or-less works, it was starting to accumulate lots of hacks and
was a burden on maintance
nogj 6729d 17h /
1684 Instead of encoding the value of the raw instruction into the generated
instruction stream, put it into a malloced array that is accessed when
analysis() is called
nogj 6729d 17h /
1683 Protect l.jr instructions from having their delay slot on the page after the
instruction itself
nogj 6729d 17h /
1682 Move the call to find_jump_loc() to avoid a condition. nogj 6729d 17h /
1681 Use page_offset_mask instead of page_size - 1 nogj 6729d 17h /
1680 Reset opq->xref to zero after the page has been recompiled in preperation for
the next time recompile_page may be called
nogj 6729d 17h /

powered by: WebSVN 2.1.0

© copyright 1999-2024 OpenCores.org, equivalent to Oliscience, all rights reserved. OpenCores®, registered trademark.