replacing umlaute and fixing address in headers Haldeneggsteig -> Universitaetsstrasse

Signed-off-by: Reto Achermann <reto.achermann@inf.ethz.ch>

armv8: fix boot into qemu with make target

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

HPET: Work on modularized hpet driver

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

hpet source code cleanup

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

Fixing some debug messages

Signed-off-by: Adam Turowski <adam.turowski@inf.ethz.ch>

HPET Driver which now can configure FSB interrupts and receive them

Signed-off-by: Rana Afifi <afrana@student-ethz.ch>

kaluga: better startup synchronization with acpi/pci

Gottardo got stuck because we have some depenencies that might result
in a deadlock. Did not happen on other machines.

Further checked error of oct_get_names() when checking for iommus to
avoid failing when there is no IOMMU record

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

kaluga: fixed compilation with debug flags

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

kaluga: several checks for setup when iommu is not enabled

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

kaluga: orchestrate the boot up process differently & fix endpoint creation

Signed-off-by: Reto Achermann <reto.achermann@inf.ethz.ch>

Retry on oct_mset and cleanup some model instantiations

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

Integrate new model interface part 2

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

pci_client: added functionality to query this connections device info
(when connection setup using EP cap)

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

kaluga: check if an element if returned from the list

Signed-off-by: Reto Achermann <reto.achermann@inf.ethz.ch>

iommu: Use queries to determine buffer

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

kaluga: request endpoint when at least one iommu driver is started

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

kaluga: start pci with all iommu endpoints

Required changes to order in discovery of skb facts acpi/pci/kaluga etc.
Further reversed the direction of pci/iommu channel setup

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

pci/kauga: actually delete the caps not just free the slots

Signed-off-by: Reto Achermann <reto.achermann@inf.ethz.ch>

kaluga: gracefully handle if iommu is not present

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

kaluga: removed some hardcoded values for e10k_vf

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

kaluga: requesting iommu endpoint cap for device drivers

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

driverkit: renamed devid cap to iommu cap

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

pci: channel to iommu setup

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

kaluga: some cleanup regarding kaluga channel setup

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

kaluga: hand off endpoint to itself to driver it starts

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

kaluga: generalized naming of endpoint requesting interface

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

kaluga: correctyl initalized pci endpoint cap

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

pci: kaluga requests PCI endpoints to add at startup of drivers

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

kaluga: when starting PCI add endpoint to kaluga to arguments

We have a boostrap problem when starting PCI from Kaluga.
Kaluga starts PCI and Kaluga requires a connection i.e.
and endpoint to PCI. The easier way is to hand off
an endpoint cap to PCI and use a different interface
to request device resources

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

kaluga: changed driver startup conditions

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

kaluga: starting e10k PF and VF driver correctly

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

kaluga: using driverkit defined slots

Signed-off-by: Reto Achermann <reto.achermann@inf.ethz.ch>

skb: added module name to driver facts

Prevsiously only binary name was used, which limited the
finding out the names for the modules which can
be started in the same domain.

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

driverkit: adding functions to extract the caps from the argcn

Signed-off-by: Reto Achermann <reto.achermann@inf.ethz.ch>

pci: remove endpoint creation and add deviceid cap

Signed-off-by: Reto Achermann <reto.achermann@inf.ethz.ch>

kaluga: allow starting of multiple IOMMU drivers

Signed-off-by: Reto Achermann <reto.achermann@inf.ethz.ch>

acpi: include iommu idx in SKB records

Signed-off-by: Reto Achermann <reto.achermann@inf.ethz.ch>

Kaluga: obtainign cap from ACPI when starting the IOMMU driver

Signed-off-by: Reto Achermann <reto.achermann@inf.ethz.ch>

kaluga: register for iommu events

Signed-off-by: Reto Achermann <reto.achermann@inf.ethz.ch>

irq: enable msi-x in conf header on route setup

Same mechanism as with legacy interrupts, pci_int_ctrl is now managing
which interrupt model will be activated in the confspace.

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

irqtest: Make sure its started only once

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

irqtest: Fix test. Enable multi instance

Enable to instantiate multiple irqtest in the same domain.

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

irqtest: Fix test. Enable multi instance

Enable to instantiate multiple irqtest in the same domain.

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

kaluga: removed unneeded debug statement

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

Pass IRQ caps to e1k devif backend, add PCI int ctrl

We add a dedicated legacy PCI interrupt controller, that will
activate interrupts in the PCI conf space. It works analogous
to the MSIx case then. The controller driver is not yet implemented.

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

kaluga: fixed bug in copying driver arguments

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

kaluga: removed driver domain hacks for e1000

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

Start of pci driver client library.

Starting to implement the pci_driver_client lib and adopting e1000
driverdriver to use it. This is the interface
that newstyle pci drivers should use. This commit introduces a bug,
that kaluga gets a trigger twice for the same device, which confuses
kaluga.

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

Removing multiple caps per bar

Our capabilites now allow byte-granularity, therefore we can always
have one cap per PCI BAR.

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

Unify use of struct pci_address, refactor kaluga driver startup

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

e1000 irqtest run in driverdomain, newstyle PCI startup

Kaluga starts PCI drivers in a newstyle driver domain. With this commit,
interrupts capabilities are passed to the init function. PCI bars and EP
not yet.

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

Start of pci driver client library.

Starting to implement the pci_driver_client lib and adopting e1000
driverdriver to use it. This is the interface
that newstyle pci drivers should use. This commit introduces a bug,
that kaluga gets a trigger twice for the same device, which confuses
kaluga.

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

Removing multiple caps per bar

Our capabilites now allow byte-granularity, therefore we can always
have one cap per PCI BAR.

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

Unify use of struct pci_address, refactor kaluga driver startup

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

e1000 irqtest run in driverdomain, newstyle PCI startup

Kaluga starts PCI drivers in a newstyle driver domain. With this commit,
interrupts capabilities are passed to the init function. PCI bars and EP
not yet.

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

kaluga: Make sure start_decnet doesnt break x86

Also refactor irq src cap management int int_caps.c/h

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

IRQ: Instantiate MSIx controller and tell driver

The PCI client library now needs to be adapted. It still assumes
that it gets only one interrupt.

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

Removed if/octopus_defs.h from public header file.

Referenced in definitions.h and trigger.h.

trigger.h is problematic because it provides helper functions for
types definied in the interface. Clients that want to use trigger.h
now have to make sure to include the interface file.

Signed-off-by: Gerd Zellweger <mail@gerdzellweger.com>

flounder: making const pointers in receiving handlers, using CONST_CAST as a temporary solution in few places but finally it should be removed

Signed-off-by: Adam Turowski <adam.turowski@inf.ethz.ch>

kaluga: Do not try to obtain IRQ cap on ARMv8 because it is not implemented

Signed-off-by: Moritz Hoffmann <moritz.hoffmann@inf.ethz.ch>

IRQ: Allow retyping of IRQSrc capability

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

IRQ: Add range to IRQSrc cap

To prepare for MSIx, we change the IRQSrc cap so that it does not
represent a single interrupt source but a whole range. This commit
changes it to a range.

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

Trying to fix vmkit test.

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

Kaluga: Print message for each driver startup

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

T283: Fix irqtest on tomme

So far, I ignored the Pin requested by the device. The code is now
also fault tolerant and will not produce an error in case the ACPI
entries for a PCI device are missing.

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

SKB: Better error output for skb clients

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

Kaluga: Do not panic on skb error

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

T258: kaluga: use cnode_create_l2()

Signed-off-by: Simon Gerber <simon.gerber@inf.ethz.ch>

IRQ: Better compatibility with existing drivers

When no int_model is specified in device_db.pl, the driver
will now not receive any additional arguments. However,
the caps are still passed, so that the pci client functions
work as expected.

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

IRQ: Kaluga start driver with IRQ caps.

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

IRQ: Refactor into int_route_client library, add legacy int range

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

IRQ: Add int_arg parameter to kalugas driver start

As we need to tell the driver what interrupt numbers to use,
we add an additional command line argument. This should
be done one day by directly passing caps instead of command
line arguments.

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

IRQ: Kaluga: put int ctrl startup in own file

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

IRQ: Kaluga, react on octopus events and start ctrl driver

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

IRQ: Use kaluga to start interrupt controllers

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

Adjustments of IPC to match new convention (static buffers).
(Flounder) Fixing a token passing in 32-bit architectures.

Signed-off-by: Adam Turowski <adam.turowski@inf.ethz.ch>

Fix printf's remove dead code.

Fixed shadowed local variables, and added missing THC include files

Fix spantest and ump_bench by making sure all_spawnd_up is set correctly again.

Add trigger_exisiting_and_watch to octopus.

Removed duplicated code in e1000 and kaluga.

Adapted kaluga to deal with multiple device instances based on entries in the SKB.

kaluga/octopus: make it compile for arm

Compiles again after merge with current tip.

Make sure everything compiles on 32bit. Still need to run tests though.

renaming defines

--HG--
branch : octopus.renaming

Renaming errors and header defines...

--HG--
branch : octopus.renaming

More renaming

--HG--
branch : octopus.renaming

rename dist2 interface to octopus

--HG--
branch : octopus.renaming
rename : if/dist2.if => if/octopus.if

dist2 include rename

--HG--
branch : octopus.renaming

Renaming header files.

--HG--
branch : octopus.renaming

Fixed NPE in start_pci. Added ahcid to device db.

Remove sleep/reset handlers from PCI interface and
adjust fish to use acpi client.

--HG--
rename : usr/kaluga/startup.h => usr/kaluga/driver_startup.h

Starting PCI.

Extended Kaluga: Added custom handler function for boot modules. Also improved some code areas and fixed some bugs.

Added functionality to start PCI driver to Kaluga.

--HG--
rename : usr/kaluga/app_cores.c => usr/kaluga/start_cpu.c