barrelfish
2 years agokaluga: also starts net_sockets_server correctly for mlx4 release2018-02-23
Roni Häcki [Wed, 21 Feb 2018 09:12:34 +0000]
kaluga: also starts net_sockets_server correctly for mlx4

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

2 years agokernel,armv8: fix mapping code to use new mapping cap format
Simon Gerber [Tue, 20 Feb 2018 17:46:45 +0000]
kernel,armv8: fix mapping code to use new mapping cap format

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

2 years agokernel,armv7: fix mapping code to use new mapping cap format
Simon Gerber [Tue, 20 Feb 2018 17:43:27 +0000]
kernel,armv7: fix mapping code to use new mapping cap format

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

2 years agolibbarrelfish,arm: add missing cap invocation wrapper for vnode_modify_flags
Simon Gerber [Tue, 20 Feb 2018 17:42:54 +0000]
libbarrelfish,arm: add missing cap invocation wrapper for vnode_modify_flags

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

2 years agokernel: capabilities: mapping types: remove offset field and switch to ptable+entry
Simon Gerber [Mon, 11 Dec 2017 16:34:23 +0000]
kernel: capabilities: mapping types: remove offset field and switch to ptable+entry

To eliminate the need for looking up page table capabilities by address when
cleaning up mappings etc, we switch the mapping capability types to hold a
pointer to the page table capability and the entry in the capability instead
of just a pointer to the first page table entry which corresponds to the
mapping.

To make room for the additional two bytes, we remove the offset into the
mapped capability which was present in the mappings but never used.

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

2 years agokernel: cap_delete: cleanup_last: do unmap when deleting last copy of mapping
Simon Gerber [Thu, 21 Dec 2017 09:50:14 +0000]
kernel: cap_delete: cleanup_last: do unmap when deleting last copy of mapping

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

2 years agomonitor: capops: initialize delete stepping in BSP monitor before going into event...
Simon Gerber [Mon, 11 Dec 2017 16:27:04 +0000]
monitor: capops: initialize delete stepping in BSP monitor before going into event loop

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

2 years agolibbarrelfish: refill root slot allocator when two slots are free (up from one)
Simon Gerber [Mon, 11 Dec 2017 16:26:25 +0000]
libbarrelfish: refill root slot allocator when two slots are free (up from one)

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

2 years agolibvfs: vfs_fd: set errno to sensible value for more vfs_fd operations
Simon Gerber [Tue, 5 Dec 2017 16:43:57 +0000]
libvfs: vfs_fd: set errno to sensible value for more vfs_fd operations

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

2 years agolibbarrelfish,ram_alloc_fixed: fix assertion about amount of basecn caps
Simon Gerber [Tue, 5 Dec 2017 16:40:29 +0000]
libbarrelfish,ram_alloc_fixed: fix assertion about amount of basecn caps

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

2 years agohake: x86_64: remove verbose flag for linker
Simon Gerber [Tue, 5 Dec 2017 08:51:47 +0000]
hake: x86_64: remove verbose flag for linker

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

2 years agokernel,x86_64: fix refactored modify_flags()
Simon Gerber [Mon, 4 Dec 2017 14:01:41 +0000]
kernel,x86_64: fix refactored modify_flags()

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

2 years agotwolevel slot alloc: resize rootcn when one slot left instead of when out of slots
Simon Gerber [Wed, 15 Mar 2017 10:11:17 +0000]
twolevel slot alloc: resize rootcn when one slot left instead of when out of slots

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

2 years agoT166: fix vnode_modify_flags
Simon Gerber [Wed, 7 Oct 2015 16:54:43 +0000]
T166: fix vnode_modify_flags

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

2 years agoT166: x86_64: implement vnode_modify_flags
Simon Gerber [Wed, 7 Oct 2015 03:49:57 +0000]
T166: x86_64: implement vnode_modify_flags

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

2 years agolibc: properly propagate errno values for open,read,write,close,lseek
Simon Gerber [Tue, 5 Dec 2017 16:43:26 +0000]
libc: properly propagate errno values for open,read,write,close,lseek

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

2 years agokernel: paging_generic: remove declaration of non-existent function
Simon Gerber [Sat, 9 Dec 2017 16:44:00 +0000]
kernel: paging_generic: remove declaration of non-existent function

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

2 years agokernel: paging_generic: fix create_mapping_cap() argument type to properly require...
Simon Gerber [Sat, 9 Dec 2017 16:42:45 +0000]
kernel: paging_generic: fix create_mapping_cap() argument type to properly require lpaddr of pte

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

2 years agolibnet: dhcp fixed bug in case of polling
Roni Häcki [Wed, 17 Jan 2018 11:56:13 +0000]
libnet: dhcp fixed bug in case of polling

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

2 years agoKaluga: Fix rebase instantiate_driver_domain
Lukas Humbel [Fri, 22 Dec 2017 15:37:37 +0000]
Kaluga: Fix rebase instantiate_driver_domain

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

2 years agoKaluga: Make unified irq src caps work on armv7
Lukas Humbel [Fri, 22 Dec 2017 15:27:42 +0000]
Kaluga: Make unified irq src caps work on armv7

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

2 years agokaluga: Make sure start_decnet doesnt break x86
Lukas Humbel [Fri, 22 Dec 2017 14:38:01 +0000]
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>

2 years agoSockeye: enumerate node inputs, use it for interrupts
Lukas Humbel [Tue, 19 Dec 2017 18:03:27 +0000]
Sockeye: enumerate node inputs, use it for interrupts

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

2 years agoSockeye: Add root namespace to int net
Lukas Humbel [Fri, 15 Dec 2017 16:41:26 +0000]
Sockeye: Add root namespace to int net

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

2 years agoSockeye: add root namespace argument
Lukas Humbel [Fri, 15 Dec 2017 16:31:27 +0000]
Sockeye: add root namespace argument

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

2 years agoSockeye: add properties to skb definition
Lukas Humbel [Mon, 11 Dec 2017 12:27:49 +0000]
Sockeye: add properties to skb definition

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

2 years agoSockeye: Add properties
Lukas Humbel [Sat, 9 Dec 2017 15:28:29 +0000]
Sockeye: Add properties

It is now possible to specify properties behind each block spec (such as in
the accept blocks) and in the destination of map specs.

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

2 years agoSockeye: Remove distinction between address and natural
Lukas Humbel [Mon, 11 Dec 2017 20:38:53 +0000]
Sockeye: Remove distinction between address and natural

Addresses and natural module parameters can now be specified as
either hex or decimal number.

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

2 years agoAdd OMAP44xx interrupt model
Lukas Humbel [Mon, 11 Dec 2017 18:18:30 +0000]
Add OMAP44xx interrupt model

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

2 years agoKaluga: adding code to start new style drivers on x86
Roni Häcki [Fri, 22 Dec 2017 12:58:10 +0000]
Kaluga: adding code to start new style drivers on x86

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

2 years agoAdd comment in move_request__rx_handler()
Simon Gerber [Thu, 9 Nov 2017 15:09:30 +0000]
Add comment in move_request__rx_handler()

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

2 years agocapops: retype: if src and dest root cnode are identical only send one copy to monitor
Simon Gerber [Mon, 6 Nov 2017 17:36:29 +0000]
capops: retype: if src and dest root cnode are identical only send one copy to monitor

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

2 years agomonitor,capops: guard root cnode copy deletions in retype to make sure we never try...
Simon Gerber [Mon, 6 Nov 2017 17:30:54 +0000]
monitor,capops: guard root cnode copy deletions in retype to make sure we never try to delete the same capref twice

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

2 years agohake: Add opt-in list for libraries in addLibraries that need add extra include direc...
Simon Gerber [Wed, 4 Oct 2017 17:07:32 +0000]
hake: Add opt-in list for libraries in addLibraries that need add extra include directories

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

2 years agomonitor: capops: delete: do not use state after freeing it
Simon Gerber [Thu, 14 Sep 2017 06:56:54 +0000]
monitor: capops: delete: do not use state after freeing it

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

2 years agokernel: caps_mark_revoke: we never have to go backwards when marking descendants
Simon Gerber [Thu, 14 Sep 2017 06:55:34 +0000]
kernel: caps_mark_revoke: we never have to go backwards when marking descendants

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

2 years agokernel: caps_mark_revoke: correctly handle case where there are only descendants...
Simon Gerber [Thu, 7 Sep 2017 14:12:10 +0000]
kernel: caps_mark_revoke: correctly handle case where there are only descendants on a core

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

2 years agodevif: moved solaflare backend to driver
Roni Häcki [Mon, 27 Nov 2017 14:42:25 +0000]
devif: moved solaflare backend to driver

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

2 years agodevif: prepared backend library to move backends
Roni Häcki [Mon, 27 Nov 2017 13:48:16 +0000]
devif: prepared backend library to move backends

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

2 years agoskb: changing a driver for a Mellanox card
Adam Turowski [Tue, 31 Oct 2017 16:17:32 +0000]
skb: changing a driver for a Mellanox card

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

2 years agomlx4: a working driver
Adam Turowski [Mon, 30 Oct 2017 08:27:07 +0000]
mlx4: a working driver

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

2 years agotools/arm_boot: make compile with usual set of compiler warnings + -Werror
Simon Gerber [Fri, 27 Oct 2017 15:50:32 +0000]
tools/arm_boot: make compile with usual set of compiler warnings + -Werror

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

2 years agoomap44xx,usb: add host of missing mackerel files to hakefile
Simon Gerber [Fri, 27 Oct 2017 15:27:15 +0000]
omap44xx,usb: add host of missing mackerel files to hakefile

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

2 years agobfscope: Explicitly add monitor flounder interface
Simon Gerber [Fri, 27 Oct 2017 15:20:51 +0000]
bfscope: Explicitly add monitor flounder interface

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

2 years agofix harness reprocess and fix armv7 build
Simon Gerber [Fri, 27 Oct 2017 13:32:37 +0000]
fix harness reprocess and fix armv7 build

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

2 years agomonitor: comment out printfs in bfscope code
Simon Gerber [Fri, 13 Oct 2017 15:32:35 +0000]
monitor: comment out printfs in bfscope code

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

2 years agokernel,x86_64: print rsp,rbp when taking kernel fault
Simon Gerber [Tue, 10 Oct 2017 16:38:56 +0000]
kernel,x86_64: print rsp,rbp when taking kernel fault

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

2 years agobfscope_nfs: properly signal dump_done when not dumping anything
Simon Gerber [Fri, 13 Oct 2017 15:31:35 +0000]
bfscope_nfs: properly signal dump_done when not dumping anything

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

2 years agokernel: caps_retype: check offset into source based on requested destination cap
Simon Gerber [Fri, 27 Oct 2017 13:34:24 +0000]
kernel: caps_retype: check offset into source based on requested destination cap

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

2 years agoexamples: Updated README.txt to reflect changes to platforms/Hakefile
Simon Gerber [Mon, 9 Oct 2017 08:27:21 +0000]
examples: Updated README.txt to reflect changes to platforms/Hakefile

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

2 years agox86: add compiler memory barrier to rdtsc inline assembly
Simon Gerber [Fri, 6 Oct 2017 16:36:57 +0000]
x86: add compiler memory barrier to rdtsc inline assembly

Debugging benchmarking code showed that the compiler would reorder some rdtsc
calls around the code they were supposed to bracket in order to do performance
measurements.  This behavior is prohibited by adding "memory" in the inline
assembly statement clobber list in the functions rdtsc() and rdtscp().

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

2 years agotrace,bfscope_nfs: don't rely on returned bytes to figure out whether more dumping...
Simon Gerber [Fri, 6 Oct 2017 16:36:26 +0000]
trace,bfscope_nfs: don't rely on returned bytes to figure out whether more dumping necessary

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

2 years agoplatforms/Hakefile: add section with programs from usr/examples
Simon Gerber [Fri, 6 Oct 2017 16:04:27 +0000]
platforms/Hakefile: add section with programs from usr/examples

Programs that currently do not compile are commented out in
platforms/Hakefile.

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

2 years agokernel: caps_delete: properly return RETRY_THROUGH_MONITOR when trying to delete...
Simon Gerber [Fri, 6 Oct 2017 16:03:14 +0000]
kernel: caps_delete: properly return RETRY_THROUGH_MONITOR when trying to delete already locked cap

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

2 years agobfscope_nfs: increase local buffer to 128M, redump trace if necessary, and cleanup...
Simon Gerber [Fri, 6 Oct 2017 15:58:05 +0000]
bfscope_nfs: increase local buffer to 128M, redump trace if necessary, and cleanup unused/unnecessary globals

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

2 years agotracing: trace_dump fix loop to not continue dumping when buffer is used up
Simon Gerber [Fri, 6 Oct 2017 15:51:56 +0000]
tracing: trace_dump fix loop to not continue dumping when buffer is used up

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

2 years agotracing: make larger trace buffers possible
Simon Gerber [Fri, 6 Oct 2017 15:50:53 +0000]
tracing: make larger trace buffers possible

Init allocates the Frame cap for the shared trace buffer. As we only gave init
16MB for it's memory allocator, this could fail for larger machines (e.g.
babybel). This commit increases the maximum number of cores for tracing to 32,
and the per-core trace buffer to 20'000 events.

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

2 years agolibnfs: rpc: make retransmit spam RPC_DEBUGP
Simon Gerber [Wed, 4 Oct 2017 17:11:02 +0000]
libnfs: rpc: make retransmit spam RPC_DEBUGP

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

2 years agotrace: up TRACE_COREID_LIMIT to 10 and reduce TRACE_MAX_EVENTS to 8000
Simon Gerber [Wed, 4 Oct 2017 17:09:50 +0000]
trace: up TRACE_COREID_LIMIT to 10 and reduce TRACE_MAX_EVENTS to 8000

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

2 years agolibbarrelfish: distops: retry getting monitor RPC client if it is currently NULL
Simon Gerber [Wed, 4 Oct 2017 17:08:09 +0000]
libbarrelfish: distops: retry getting monitor RPC client if it is currently NULL

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

2 years agopleco: generate actually valid JSON
Simon Gerber [Wed, 4 Oct 2017 17:03:27 +0000]
pleco: generate actually valid JSON

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

2 years agobfscope: Implement variant that dumps traces over NFS
Simon Gerber [Wed, 4 Oct 2017 17:02:51 +0000]
bfscope: Implement variant that dumps traces over NFS

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

2 years agolibvfs: nfs client: update write #define parameters to work with new network stack
Simon Gerber [Wed, 4 Oct 2017 17:01:42 +0000]
libvfs: nfs client: update write #define parameters to work with new network stack

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

2 years agoSockeye: Move overlay translation to backend
Daniel Schwyn [Wed, 18 Oct 2017 14:59:45 +0000]
Sockeye: Move overlay translation to backend

Signed-off-by: Daniel Schwyn <schwyda@student.ethz.ch>

2 years agoTN: Device queue added implementation details
Roni Häcki [Thu, 26 Oct 2017 12:35:00 +0000]
TN: Device queue added implementation details

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

2 years agoTN: devif documentation extended
Roni Häcki [Wed, 25 Oct 2017 14:56:27 +0000]
TN: devif documentation extended

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

2 years agoTN: update to devif technote
Roni Häcki [Tue, 24 Oct 2017 15:38:00 +0000]
TN: update to devif technote

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

2 years agodevif: added flag to header
Roni Häcki [Tue, 24 Oct 2017 11:58:45 +0000]
devif: added flag to header

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

2 years agoTN: first part of Device Queue technote
Roni Häcki [Tue, 24 Oct 2017 09:19:37 +0000]
TN: first part of Device Queue technote

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

2 years agomlx4: adding the enqueuing functions
Adam Turowski [Wed, 4 Oct 2017 15:21:48 +0000]
mlx4: adding the enqueuing functions

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

2 years agoIRQ: Split into MSIx and Legacy test.
Lukas Humbel [Mon, 23 Oct 2017 14:18:57 +0000]
IRQ: Split into MSIx and Legacy test.

Backwards compatible qemu-wrapper (now uses e1000e if available).

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

2 years agodevif: debug backend added some missing checks
Roni Häcki [Mon, 23 Oct 2017 12:48:23 +0000]
devif: debug backend added some missing checks

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

2 years agodevif: benchmark now dumps results to nfs
Roni Häcki [Mon, 23 Oct 2017 11:33:48 +0000]
devif: benchmark now dumps results to nfs

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

2 years agomlx4: fixed compilation in release mode
Roni Häcki [Mon, 23 Oct 2017 10:57:49 +0000]
mlx4: fixed compilation in release mode

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

2 years agovfs: added missing dependency
Roni Häcki [Tue, 17 Oct 2017 09:38:02 +0000]
vfs: added missing dependency

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

2 years agoMerge branch 'irq'
Lukas Humbel [Mon, 16 Oct 2017 15:27:47 +0000]
Merge branch 'irq'

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

2 years agodevif: added benchmark for stacking queues
Roni Häcki [Mon, 16 Oct 2017 15:25:22 +0000]
devif: added benchmark for stacking queues

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

2 years agoIRQ: Restore legacy interrupt test
Lukas Humbel [Mon, 16 Oct 2017 14:39:12 +0000]
IRQ: Restore legacy interrupt test

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

2 years agoIRQ: Successfully convince e1000 to trigger MSIx interrupts
Lukas Humbel [Mon, 16 Oct 2017 14:15:12 +0000]
IRQ: Successfully convince e1000 to trigger MSIx interrupts

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

2 years agoIRQ: Getting the 82574 to trigger MSIxs...
Lukas Humbel [Fri, 6 Oct 2017 16:48:55 +0000]
IRQ: Getting the 82574 to trigger MSIxs...

The msix controller gets instantiated in the card, it gets
the correct message. The card is being set up to use MSIx.
But something is not working yet...

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

2 years agoIRQ: Better interrupt model selection logic
Lukas Humbel [Thu, 5 Oct 2017 15:08:41 +0000]
IRQ: Better interrupt model selection logic

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

2 years agoIRQ: Make ioapic controller understand new format
Lukas Humbel [Wed, 4 Oct 2017 16:35:37 +0000]
IRQ: Make ioapic controller understand new format

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

2 years agoFix ACPI debug statements
Lukas Humbel [Wed, 4 Oct 2017 16:34:38 +0000]
Fix ACPI debug statements

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

2 years agoIRQ: Separate e1ks by msix capability
Lukas Humbel [Wed, 4 Oct 2017 15:12:23 +0000]
IRQ: Separate e1ks by msix capability

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

2 years agoIRQ: Parse skb output (again)
Lukas Humbel [Wed, 4 Oct 2017 14:02:04 +0000]
IRQ: Parse skb output (again)

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

2 years agoIRQ: Fixup skb output parsing
Lukas Humbel [Tue, 3 Oct 2017 16:54:05 +0000]
IRQ: Fixup skb output parsing

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

2 years agoIRQ: Fixup rebase errors
Lukas Humbel [Tue, 3 Oct 2017 08:24:38 +0000]
IRQ: Fixup rebase errors

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

2 years agoIRQ: Try to parse MSIX mem_write terms from skb
Lukas Humbel [Mon, 2 Oct 2017 12:04:12 +0000]
IRQ: Try to parse MSIX mem_write terms from skb

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

2 years agodevif: missing dependency added
Roni Häcki [Mon, 16 Oct 2017 15:21:45 +0000]
devif: missing dependency added

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

2 years agoIRQ: Added irq_idx to route call
Lukas Humbel [Tue, 8 Nov 2016 09:45:58 +0000]
IRQ: Added irq_idx to route call

When given an interrupt source capability with a range, this
allows you to specify exactly which vector you want routed.

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

2 years agoIRQ: Instantiate MSIx controller and tell driver
Lukas Humbel [Fri, 4 Nov 2016 16:17:02 +0000]
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>

2 years agodevif: added some code to benchmark backends
Roni Häcki [Mon, 16 Oct 2017 15:00:40 +0000]
devif: added some code to benchmark backends

To use define #BENCH

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

2 years agoproc_mgmt: fixed race condition between proc_mgmt and spawnd on other
Roni Häcki [Fri, 13 Oct 2017 14:36:19 +0000]
proc_mgmt: fixed race condition between proc_mgmt and spawnd on other
cores

Needs a better solution when cores can come and go since we just
wait for the spawnd on the core that we want to spawn something on

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

2 years agodevif: loopback fixed bug
Roni Häcki [Fri, 13 Oct 2017 09:29:47 +0000]
devif: loopback fixed bug

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

2 years agoMerge remote-tracking branch 'origin/proc_mgmt_unify' release2017-10-12
Roni Häcki [Tue, 10 Oct 2017 15:35:41 +0000]
Merge remote-tracking branch 'origin/proc_mgmt_unify'

Unified way of spawning processes through the process manger
instead of directly using spawnd

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

2 years agoharness: devif test load only e10k driver on babybel2
Roni Häcki [Tue, 10 Oct 2017 14:34:50 +0000]
harness: devif test load only e10k driver on babybel2

With the new net_sockets_server for both e10k and solarflare
get an IP which might confuse thy harness code since it
parses the target IP from the console output.

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

2 years agousb manager: converted spawn_program call to new cap based interface
Roni Häcki [Tue, 10 Oct 2017 14:06:03 +0000]
usb manager: converted spawn_program call to new cap based interface

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

2 years agoxeonphi: wait for spawnd 0 to be up before trying to spawn on core 0
Roni Häcki [Tue, 10 Oct 2017 10:10:32 +0000]
xeonphi: wait for spawnd 0 to be up before trying to spawn on core 0

Since we now use the process manager as proxy to the spawnds on
different cores, the process manger returns an error when it does not
have a connection to the spawnd on the core it tries to spawn something

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

2 years agoxeonphi: adapted to new spawn calls
Roni Häcki [Tue, 10 Oct 2017 09:21:14 +0000]
xeonphi: adapted to new spawn calls

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

2 years agoproc_mgmt: xeon phi translation did to cap implemented
Roni Häcki [Tue, 10 Oct 2017 08:52:27 +0000]
proc_mgmt: xeon phi translation did to cap implemented

Currently the interphi channel usues domain ids to identify
a domain. In the rest of the code caps are used.

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