adding disable_ide_dma true in the kernel settings file

From a quick look, the ATA driver only looks to safe mode settings for
that, and not the kernel file. The attached patch should address that.

ticket : #10253

Change-Id: I5e5c2474fecbd441a36a3600f7f16a964e0945dd
Reviewed-on: https://review.haiku-os.org/c/haiku/+/56
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>

Various 64-bit fixes to SCSI and ATA bus managers/drivers.

Mostly compilation fixes, as well as a few 64-bit safety fixes. I've
briefly looked through everything for any obvious issues and fixed
the ones I've found, and it seems like they're working properly, though
there could be some more well hidden ones that I've missed.

* Added a new B_GET_DEVICE_NAME ioctl - this should be implemented by all
drivers in the future, such that NetworkStatus and similar software can show
nice names for the devices. The device manager should implement this and
return the B_DEVICE_PRETTY_NAME of the device (and in turn, new style drivers
should actually set this).
* Implemented handling of this ioctl in the scsi_periph to return the vendor/
product strings.
* Implemented this in the ATA bus manager to return the model from the info
block.
* KDiskDevice now fills in the partition_data::name if the B_GET_DEVICE_NAME
succeeds.
* As a side effect, at least BootManager now shows the drive name; maybe
DriveSetup does as well for the raw device.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@40231 a95241bf-73f2-0310-859d-f6bbb57e9c96

Merged the signals branch into trunk, with these changes:
* The team and thread kernel structures have been renamed to Team and Thread
respectively and moved into the new BKernel namespace.
* Several (kernel add-on) sources have been converted from C to C++ since
private kernel headers are included that are no longer C compatible.

Changes after merging:
* Fixed gcc 2 build (warnings mainly in the scary firewire bus manager).


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@40196 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Changed physical_entry::{address,size} to phys_{addr,size}_t and changed
map_physical_memory()'s physicalAddress parameter type from void* to
phys_addr_t. This breaks source compatibility, but -- as long as
phys_{addr,size}_t remain 32 bit wide -- keeps binary compatibility with
BeOS.
* Adjusted all code using the affected interfaces (Oh what fun!). Added a few
TODOs in places where the wrong types (e.g. void* for physical addresses
are used). Looks like quite a few drivers aren't 64 bit safe and others
will break with PAE.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36960 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Got rid of the ATA_BLOCK_SIZE define.
* The logical and physical block sizes are now retrieved from the info block,
though only the logical one is actually used (where ATA_BLOCK_SIZE was used
before).
* Also added retrieving the offset of the first logical block inside the first
physical one.
* All this stupidity just for Windows XP, it's really sad.
* Added "const" to some methods, removed some superfluous ';'.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36416 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Moved the VM headers into subdirectory vm/.
* Renamed vm_cache.h/vm_address_space.h to VMCache.h/VMAddressSpace.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34449 a95241bf-73f2-0310-859d-f6bbb57e9c96

Fix for ticket #4543. This fixes the ATA
timeout waiting for interrupt / device busy / status 0x58


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33626 a95241bf-73f2-0310-859d-f6bbb57e9c96

Write device identification into syslog.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33042 a95241bf-73f2-0310-859d-f6bbb57e9c96

Clear pending interrupts after reset and timeouts


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33039 a95241bf-73f2-0310-859d-f6bbb57e9c96

Add more debug code. I see spurius device selection timeouts.
When this happens, no recovery is possible right now.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32989 a95241bf-73f2-0310-859d-f6bbb57e9c96

Faster bus scan by reusing device presence information that is available before
resetting the bus. This allows to skip devices that are not present. Works for
most but not all devices. Sometimes device 0 pretends that a device 1 exists,
but it's usefull anyway.
Allow working device 1 only configurations by not selecting device 0 anymore
for identification. Moved indentification from Reset into BusScan to do everything
in a single step. A SATA emulation in combined mode might have a slave (device 1)
only configuration when only one device is connected, depending on the port.
Also tested with PATA.
Recovery for lost interrupts. Works but is dead slow because of the timeout.
Bus reset isn't possible anymore, this needs to be implemented in a save way.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32978 a95241bf-73f2-0310-859d-f6bbb57e9c96

Modified initial ata device detection. Works for me. Please test.
If you get an "error waiting for interrupt" during boot, try to disabled IDE DMA.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32769 a95241bf-73f2-0310-859d-f6bbb57e9c96

Initialize last part of byte to 0, just in case.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32633 a95241bf-73f2-0310-859d-f6bbb57e9c96

Update and enable the debugging device selection check. Might shed some light on
the problem in #4308.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32616 a95241bf-73f2-0310-859d-f6bbb57e9c96

By popular request: Change the invalid device number and add a comment
explaining the reasons behind it. Oh well, at least the commit count raises...


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30644 a95241bf-73f2-0310-859d-f6bbb57e9c96

Add three more error checks for paranoias sake.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30642 a95241bf-73f2-0310-859d-f6bbb57e9c96

Workaround for Gigabyte i-RAM, which always reports 0x00 as error code after bus reset,


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30474 a95241bf-73f2-0310-859d-f6bbb57e9c96

According to ATA-4 specification, the error and device-fault
bit (and all other status register bits) are only valid when
the busy bit is zero. This makes PIO mode work on a WD Raptor.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30443 a95241bf-73f2-0310-859d-f6bbb57e9c96

Make debugging easier by tracing only status changes.
Always first disable interrupts on a device that can be
selected, then check for invalid status.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30426 a95241bf-73f2-0310-859d-f6bbb57e9c96

Check the status value for validity before waiting for a device. If it has an
invalid value, assume the device is not present. Should fix long boot pauses
in VMWare and maybe also on some hardware.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30381 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Fixed wrong indentation, and number of blank lines.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30367 a95241bf-73f2-0310-859d-f6bbb57e9c96

Only use 2 bytes of the signature instead of all four. Even though the specs
clearly state that the other two bytes need to be 0x01 (and they always did,
even back to the very first ATA/ATAPI specs), I have one device here that
doesn't set the sector count register to 1. Since those two bytes are the unique
ones of the signatures anyway, it shouldn't harm to just ignore the others.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30311 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Explicitly check device selection success. If device selection failed, assume
there is no device. Should fix long reset timeouts when only device 1 is
present and therefore device 0 can't be selected.
* In case a device reset error is reported, don't try to identify/use the device.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30286 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Wait slightly longer before falling into passive poll mode.
* Cause scheduler invokation when we handled an interrupt as the transfering
thread can continue immediately. This brings small block transfer rates to
a comparable level to the old IDE bus_manager. For large block transfers there
is no speed difference at all.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30257 a95241bf-73f2-0310-859d-f6bbb57e9c96

Apply block count maximum according to device capability. If the device supports
LBA48 we can allow a maximum block count of 65535 as we have a 16bit sector
count field available.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30254 a95241bf-73f2-0310-859d-f6bbb57e9c96

For packet devices we want to actually return a check condition when the device
indicates an error. Also we then need to pass through the request sense command
to the device. This should fix media status reporting for ATAPI.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30253 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Reworked device detection, we'll see how well this works...
* Reworked some of the transfer handling after reading the specs.
* Ensure that the device selection bit is set correctly for all commands.
* Generally disable interrupts and enable them only when expecting a DMA one.
* Renamed disk failure to device fault according to specs.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30248 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Rework ATAPI handling, removing all PIO interrupts. As Marcus pointed out the
complication is not really worth it and it would in any case be unreliable
in the shared interrupt case (even though you could always check the poll
condition in the interrupt handler).
* Probably broke ATAPI with those changes though.
* Change blocksLeft to bytesLeft to allow using it for arbitrary ATAPI requests.
* Check some more error conditions and actually let those states get through
to the SCSI layer.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30109 a95241bf-73f2-0310-859d-f6bbb57e9c96

Move the lock into the request.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30100 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Fix race condition between finishing a request and returning a busy status
when the channel is in use, which would cause the SCSI scheduler to stop
sending requests.
* Indeed we need to return a status from the interrupt handler, as for PIO
transfers there is no way of knowning whether or not the interrupt at hand
was ours.
* Add some debug output.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30091 a95241bf-73f2-0310-859d-f6bbb57e9c96

Adding ATAPI support to the new ATA bus_manager. Only tested in QEMU.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30080 a95241bf-73f2-0310-859d-f6bbb57e9c96

Implement DMA transfers in the new ATA bus_manager.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30069 a95241bf-73f2-0310-859d-f6bbb57e9c96

Adding DMA detection and generally allow DMA. Not used yet though.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30054 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Adding/updating license headers.
* Add the standalone ata module to the Jamfile.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30053 a95241bf-73f2-0310-859d-f6bbb57e9c96

Additional to the normal detection of device presence, use the trick previously
present in the IDE bus_manager when detecting devices to avoid long timeouts
when waiting for possible ATAPI devices.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30052 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Work in progress commit of a reworked ATA bus_manager. It's now object based,
but doesn't really do anything more than before.
* It also replaces everything IDE with ATA counterparts and cleans up a lot
of the definitions.
* Cleaning up the old ATA bus_manager as well as some license headers missing.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30049 a95241bf-73f2-0310-859d-f6bbb57e9c96

Various 64-bit fixes to SCSI and ATA bus managers/drivers.

Mostly compilation fixes, as well as a few 64-bit safety fixes. I've
briefly looked through everything for any obvious issues and fixed
the ones I've found, and it seems like they're working properly, though
there could be some more well hidden ones that I've missed.

* Added a new B_GET_DEVICE_NAME ioctl - this should be implemented by all
drivers in the future, such that NetworkStatus and similar software can show
nice names for the devices. The device manager should implement this and
return the B_DEVICE_PRETTY_NAME of the device (and in turn, new style drivers
should actually set this).
* Implemented handling of this ioctl in the scsi_periph to return the vendor/
product strings.
* Implemented this in the ATA bus manager to return the model from the info
block.
* KDiskDevice now fills in the partition_data::name if the B_GET_DEVICE_NAME
succeeds.
* As a side effect, at least BootManager now shows the drive name; maybe
DriveSetup does as well for the raw device.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@40231 a95241bf-73f2-0310-859d-f6bbb57e9c96

Merged the signals branch into trunk, with these changes:
* The team and thread kernel structures have been renamed to Team and Thread
respectively and moved into the new BKernel namespace.
* Several (kernel add-on) sources have been converted from C to C++ since
private kernel headers are included that are no longer C compatible.

Changes after merging:
* Fixed gcc 2 build (warnings mainly in the scary firewire bus manager).


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@40196 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Changed physical_entry::{address,size} to phys_{addr,size}_t and changed
map_physical_memory()'s physicalAddress parameter type from void* to
phys_addr_t. This breaks source compatibility, but -- as long as
phys_{addr,size}_t remain 32 bit wide -- keeps binary compatibility with
BeOS.
* Adjusted all code using the affected interfaces (Oh what fun!). Added a few
TODOs in places where the wrong types (e.g. void* for physical addresses
are used). Looks like quite a few drivers aren't 64 bit safe and others
will break with PAE.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36960 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Got rid of the ATA_BLOCK_SIZE define.
* The logical and physical block sizes are now retrieved from the info block,
though only the logical one is actually used (where ATA_BLOCK_SIZE was used
before).
* Also added retrieving the offset of the first logical block inside the first
physical one.
* All this stupidity just for Windows XP, it's really sad.
* Added "const" to some methods, removed some superfluous ';'.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36416 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Moved the VM headers into subdirectory vm/.
* Renamed vm_cache.h/vm_address_space.h to VMCache.h/VMAddressSpace.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34449 a95241bf-73f2-0310-859d-f6bbb57e9c96

Fix for ticket #4543. This fixes the ATA
timeout waiting for interrupt / device busy / status 0x58


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33626 a95241bf-73f2-0310-859d-f6bbb57e9c96

Write device identification into syslog.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33042 a95241bf-73f2-0310-859d-f6bbb57e9c96

Clear pending interrupts after reset and timeouts


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33039 a95241bf-73f2-0310-859d-f6bbb57e9c96

Add more debug code. I see spurius device selection timeouts.
When this happens, no recovery is possible right now.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32989 a95241bf-73f2-0310-859d-f6bbb57e9c96

Faster bus scan by reusing device presence information that is available before
resetting the bus. This allows to skip devices that are not present. Works for
most but not all devices. Sometimes device 0 pretends that a device 1 exists,
but it's usefull anyway.
Allow working device 1 only configurations by not selecting device 0 anymore
for identification. Moved indentification from Reset into BusScan to do everything
in a single step. A SATA emulation in combined mode might have a slave (device 1)
only configuration when only one device is connected, depending on the port.
Also tested with PATA.
Recovery for lost interrupts. Works but is dead slow because of the timeout.
Bus reset isn't possible anymore, this needs to be implemented in a save way.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32978 a95241bf-73f2-0310-859d-f6bbb57e9c96

Modified initial ata device detection. Works for me. Please test.
If you get an "error waiting for interrupt" during boot, try to disabled IDE DMA.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32769 a95241bf-73f2-0310-859d-f6bbb57e9c96

Initialize last part of byte to 0, just in case.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32633 a95241bf-73f2-0310-859d-f6bbb57e9c96

Update and enable the debugging device selection check. Might shed some light on
the problem in #4308.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32616 a95241bf-73f2-0310-859d-f6bbb57e9c96

By popular request: Change the invalid device number and add a comment
explaining the reasons behind it. Oh well, at least the commit count raises...


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30644 a95241bf-73f2-0310-859d-f6bbb57e9c96

Add three more error checks for paranoias sake.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30642 a95241bf-73f2-0310-859d-f6bbb57e9c96

Workaround for Gigabyte i-RAM, which always reports 0x00 as error code after bus reset,


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30474 a95241bf-73f2-0310-859d-f6bbb57e9c96

According to ATA-4 specification, the error and device-fault
bit (and all other status register bits) are only valid when
the busy bit is zero. This makes PIO mode work on a WD Raptor.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30443 a95241bf-73f2-0310-859d-f6bbb57e9c96

Make debugging easier by tracing only status changes.
Always first disable interrupts on a device that can be
selected, then check for invalid status.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30426 a95241bf-73f2-0310-859d-f6bbb57e9c96

Check the status value for validity before waiting for a device. If it has an
invalid value, assume the device is not present. Should fix long boot pauses
in VMWare and maybe also on some hardware.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30381 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Fixed wrong indentation, and number of blank lines.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30367 a95241bf-73f2-0310-859d-f6bbb57e9c96

Only use 2 bytes of the signature instead of all four. Even though the specs
clearly state that the other two bytes need to be 0x01 (and they always did,
even back to the very first ATA/ATAPI specs), I have one device here that
doesn't set the sector count register to 1. Since those two bytes are the unique
ones of the signatures anyway, it shouldn't harm to just ignore the others.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30311 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Explicitly check device selection success. If device selection failed, assume
there is no device. Should fix long reset timeouts when only device 1 is
present and therefore device 0 can't be selected.
* In case a device reset error is reported, don't try to identify/use the device.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30286 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Wait slightly longer before falling into passive poll mode.
* Cause scheduler invokation when we handled an interrupt as the transfering
thread can continue immediately. This brings small block transfer rates to
a comparable level to the old IDE bus_manager. For large block transfers there
is no speed difference at all.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30257 a95241bf-73f2-0310-859d-f6bbb57e9c96

Apply block count maximum according to device capability. If the device supports
LBA48 we can allow a maximum block count of 65535 as we have a 16bit sector
count field available.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30254 a95241bf-73f2-0310-859d-f6bbb57e9c96

For packet devices we want to actually return a check condition when the device
indicates an error. Also we then need to pass through the request sense command
to the device. This should fix media status reporting for ATAPI.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30253 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Reworked device detection, we'll see how well this works...
* Reworked some of the transfer handling after reading the specs.
* Ensure that the device selection bit is set correctly for all commands.
* Generally disable interrupts and enable them only when expecting a DMA one.
* Renamed disk failure to device fault according to specs.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30248 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Rework ATAPI handling, removing all PIO interrupts. As Marcus pointed out the
complication is not really worth it and it would in any case be unreliable
in the shared interrupt case (even though you could always check the poll
condition in the interrupt handler).
* Probably broke ATAPI with those changes though.
* Change blocksLeft to bytesLeft to allow using it for arbitrary ATAPI requests.
* Check some more error conditions and actually let those states get through
to the SCSI layer.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30109 a95241bf-73f2-0310-859d-f6bbb57e9c96

Move the lock into the request.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30100 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Fix race condition between finishing a request and returning a busy status
when the channel is in use, which would cause the SCSI scheduler to stop
sending requests.
* Indeed we need to return a status from the interrupt handler, as for PIO
transfers there is no way of knowning whether or not the interrupt at hand
was ours.
* Add some debug output.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30091 a95241bf-73f2-0310-859d-f6bbb57e9c96

Adding ATAPI support to the new ATA bus_manager. Only tested in QEMU.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30080 a95241bf-73f2-0310-859d-f6bbb57e9c96

Implement DMA transfers in the new ATA bus_manager.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30069 a95241bf-73f2-0310-859d-f6bbb57e9c96

Adding DMA detection and generally allow DMA. Not used yet though.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30054 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Adding/updating license headers.
* Add the standalone ata module to the Jamfile.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30053 a95241bf-73f2-0310-859d-f6bbb57e9c96

Additional to the normal detection of device presence, use the trick previously
present in the IDE bus_manager when detecting devices to avoid long timeouts
when waiting for possible ATAPI devices.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30052 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Work in progress commit of a reworked ATA bus_manager. It's now object based,
but doesn't really do anything more than before.
* It also replaces everything IDE with ATA counterparts and cleans up a lot
of the definitions.
* Cleaning up the old ATA bus_manager as well as some license headers missing.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@30049 a95241bf-73f2-0310-859d-f6bbb57e9c96