Commit Graph

98642 Commits

Author SHA1 Message Date
Roger Pau Monné
24f7e474cc xen: add missing files
Commit missing files that actually belong to previous commits.

Sponsored by: Citrix Systems R&D
Approved by: gibbs
2014-06-16 08:54:04 +00:00
Roger Pau Monné
79cd455edb isa: allow ISA bus to attach to xenpv bus
This is needed because syscons depends on ISA.

Sponsored by: Citrix Systems R&D
Approved by: gibbs

x86/isa/isa.c:
 - Allow the ISA bus to attach to xenpv.
2014-06-16 08:49:16 +00:00
Roger Pau Monné
a8f2f559f3 xen: unify gnttab initialization for PVHVM and PVH
Switch the initialization of gnttab to use an unused physical memory
range for both PVHVM and PVH.

In the past PVHVM was using the xenpci BAR, but there's no reason to
do that, and in fact FreeBSD was probably doing it because it was the
way it was done in Windows, were drivers cannot probably request for
unused physical memory ranges, but it was never enforced in the
hypervisor.

Sponsored by: Citrix Systems R&D
Approved by: gibbs

xen/gnttab.c:
 - Allocate contiguous physical memory for grant table frames for both
   PVHVM and PVH.
 - Since gnttab is not a device, use the xenpv device in order to
   request for this allocation.

dev/xen/xenpci/xenpcivar.h:
dev/xen/xenpci/xenpci.c:
 - Remove the now unused xenpci_alloc_space and xenpci_alloc_space_int
   functions.

xen/gnttab.h:
 - Change the prototype of gnttab_init and gnttab_resume, that now
   takes a device_t parameter.

dev/xen/control/control.c:
x86/xen/xenpv.c:
 - Changes to accomodate the new prototype of gnttab_init and
   gnttab_resume.
2014-06-16 08:48:42 +00:00
Roger Pau Monné
3593fa95ef xen: init gnttab from xenpv
Currently the grant table is initialized from xenstore, but a better
place to do this would be xenpv, so move grant table initialization
there.

Sponsored by: Citrix Systems R&D
Approved by: gibbs

x86/xen/xenpv.c:
 - Add gnttab initialization.

xen/xenstore/xenstore.c:
 - Remove gnttab initialization.
2014-06-16 08:48:06 +00:00
Roger Pau Monné
9a522f8203 xen: xenstore changes to support PVH
For PVH guests the xenstore parameters are fetched from the start_info
struct, just like on PV.

Sponsored by: Citrix Systems R&D
Approved by: gibbs

xen/xenstore/xenstore.c:
 - Fetch xenstore event channel port from start_info.
2014-06-16 08:47:15 +00:00
Roger Pau Monné
bd851e636f xen: add shutdown hook for PVH
Add the PV shutdown hook to PVH.

Sponsored by: Citrix Systems R&D
Approved by: gibbs

dev/xen/control/control.c:
 - Make xen_pv_shutdown_final available on XENHVM builds.
 - Register the Xen PV shutdown hook for PVH guests.
2014-06-16 08:46:31 +00:00
Roger Pau Monné
e048c70674 xen: create a Xen nexus to use in PV/PVH
Introduce a Xen specific nexus that is going to be used by Xen PV/PVH
guests.

Sponsored by: Citrix Systems R&D
Approved by: gibbs

x86/xen/xen_nexus.c:
 - Introduce a Nexus to use on Xen PV(H) guests, this prevents PV(H)
   guests from using the legacy Nexus.

conf/files.amd64:
conf/files.i386:
 - Add the xen nexus to the build.
2014-06-16 08:45:51 +00:00
Roger Pau Monné
34ec15da90 xen: create a PV CPU device for PVH guests
Since there's no ACPI on PVH guests, we need to create a dummy CPU
device in order to fill the pcpu->pc_device field.

Sponsored by: Citrix Systems R&D
Approved by: gibbs

dev/xen/pvcpu/pvcpu.c:
 - Create a dummy CPU device for PVH guests in order to fill the
   per-cpu pc_device field.

conf/files:
 - Add the pvcpu device to kernels using XEN or XENHVM options.
2014-06-16 08:45:12 +00:00
Roger Pau Monné
aa64d12bcd xen: introduce xenpv bus
Create a dummy bus so top level Xen devices can attach to it (instead
of attaching directly to the nexus). This allows to have all the Xen
related devices grouped under a single bus.

Sponsored by: Citrix Systems R&D
Approved by: gibbs

x86/xen/xenpv.c:
 - Attach the xenpv bus when running as a Xen guest.
 - Attach the ISA bus if needed, in order to attach syscons.

conf/files.amd6:
conf/files.i386:
 - Include the xenpv.c file in the build of i386/amd64 kernels using
   XENHVM.

dev/xen/console/console.c:
dev/xen/timer/timer.c:
xen/xenstore/xenstore.c:
 - Attach to the xenpv bus instead of the Nexus.

dev/xen/xenpci/xenpci.c:
 - Xen specific devices on PVHVM guests are no longer attached to the
   xenpci device, they are instead attached to the xenpv bus, remove
   the now unused methods.
2014-06-16 08:44:33 +00:00
Roger Pau Monné
842471b331 xen: add hooks for Xen PV APIC
Create the necessary hooks in order to provide a Xen PV APIC
implementation that can be used on PVH. Most of the lapic ops
shouldn't be called on Xen, since we trap those operations at a higher
layer.

Sponsored by: Citrix Systems R&D
Approved by: gibbs

x86/xen/hvm.c:
x86/xen/xen_apic.c:
 - Move IPI related code to xen_apic.c

x86/xen/xen_apic.c:
 - Introduce Xen PV APIC implementation, most of the functions of the
   lapic interface should never be called when running as PV(H) guest,
   so make sure FreeBSD panics when trying to use one of those.
 - Define the Xen APIC implementation in xen_apic_ops.

xen/xen_pv.h:
 - Extern declaration of the xen_apic struct.

x86/xen/pv.c:
 - Use xen_apic_ops as apic_ops when running as PVH guest.

conf/files.amd64:
conf/files.i386:
 - Include the xen_apic.c file in the build of i386/amd64 kernels
   using XENHVM.
2014-06-16 08:43:45 +00:00
Roger Pau Monné
ef409ede7b amd64/i386: introduce APIC hooks for different APIC implementations.
This is needed for Xen PV(H) guests, since there's no hardware lapic
available on this kind of domains. This commit should not change
functionality.

Sponsored by: Citrix Systems R&D
Reviewed by: jhb
Approved by: gibbs

amd64/include/cpu.h:
amd64/amd64/mp_machdep.c:
i386/include/cpu.h:
i386/i386/mp_machdep.c:
 - Remove lapic_ipi_vectored hook from cpu_ops, since it's now
   implemented in the lapic hooks.

amd64/amd64/mp_machdep.c:
i386/i386/mp_machdep.c:
 - Use lapic_ipi_vectored directly, since it's now an inline function
   that will call the appropiate hook.

x86/x86/local_apic.c:
 - Prefix bare metal public lapic functions with native_ and mark them
   as static.
 - Define default implementation of apic_ops.

x86/include/apicvar.h:
 - Declare the apic_ops structure and create inline functions to
   access the hooks, so the change is transparent to existing users of
   the lapic_ functions.

x86/xen/hvm.c:
 - Switch to use the new apic_ops.
2014-06-16 08:43:03 +00:00
Roger Pau Monné
5f35f84fa0 xen: fix style in pv.c
Fix the lenght of some comments, and also add proper indentation to
xen_init_ops

Sponsored by: Citrix Systems R&D
Approved by: gibbs
2014-06-16 08:41:57 +00:00
Bryan Venteicher
bae486f5d7 Force two byte alignment for all control message headers
The header structure consists of two 1-byte elements, but it must always
be describable by a single SG entry. Note for consistency, specify the
alignment everywhere, even if the structure has the appropriate natural
alignment since it contains a uint16_t.

Obtained from:	DragonFlyBSD
MFC after:	1 week
2014-06-16 04:32:27 +00:00
Bryan Venteicher
fd5b395117 Make the feature negotiation code easier to follow
MFC after:	1 week
2014-06-16 04:29:28 +00:00
Bryan Venteicher
45543f0751 Move the VIRTIO_RING_F_* defines out of virtqueue.h into virtio_config.h
These defines are applicable to userland too, but virtqueue.h contains
the kernel virtqueue interface, and is therefore not usable in userland.

Note that Linux places these defines in virtio_ring.h, but I don't want
the drivers including this header file to keep the VirtIO ring opaque to
everything but the virtqueue.

MFC after:	1 week
2014-06-16 04:25:04 +00:00
Bryan Venteicher
e026de111e Remove kernel specific macro out of the VirtIO PCI header file
The eventual goal is to share this file with userland, so
remove the macro that is only specific for virtio_pci(4).
Instead, add the VIRTIO_PCI_CONFIG_OFF macro from Linux to
get the config size whether MSIX is enabled or not.

MFC after:	1 week
2014-06-16 04:16:31 +00:00
Bryan Venteicher
add526c613 - Remove two write-only local variables
- Remove unused element in the vtnet_rxq structure

MFC after:	1 week
2014-06-16 04:12:33 +00:00
Alexander Motin
e86a414238 Make backends track completion by processed number of sectors instead of
total transfer size.

Commands such as VERIFY or COMPARE AND WRITE may have transfer size not
matching directly to number of sectors.
2014-06-15 20:14:11 +00:00
Alexander Motin
66df9136e3 Remove memcpy() from ctl_private[] accesses.
That union is aligned enough to access data directly.
2014-06-15 18:16:51 +00:00
Alexander Motin
9c71cd5aae Move kern_total_len setting from backend to core code. 2014-06-15 17:14:52 +00:00
Ruslan Bukin
b9402e3d56 Add very basic driver for 12-bit Analog to Digital Converter (ADC). 2014-06-15 12:08:33 +00:00
Alexander Motin
20a5f2d963 Format Portal Group Tag same as istgt does -- %4.4x instead of %x.
SPC-4 spec tells it should be "two or more hexadecimal digits".
RFC3720 tells it is 16-bit value.

MFC after:	2 weeks
2014-06-15 10:04:44 +00:00
Alexander Motin
eb3687a6a6 Remove custom processing for "file" option. 2014-06-15 09:37:06 +00:00
Alexander Motin
5777f09019 Respect "vendor" option in all places.
MFC after:	2 weeks
2014-06-15 08:43:52 +00:00
Alan Cox
e1f92ccc73 One of the intentions behind r267254 was that the global variable "sgrowsiz"
would be read once and cached in a local variable so that the resource limit
check and map entry insertion would be guaranteed to use the same value.
However, the value being passed to vm_map_insert() is still from "sgrowsiz"
and not the local variable.  Correct this oversight.

Reviewed by:	kib
2014-06-15 07:52:59 +00:00
Alexander Motin
0c934f7f89 Add "vendor", "product" and "revision" options to control inquiry data.
MFC after:	2 weeks
2014-06-15 06:56:10 +00:00
Konstantin Belousov
31241fa778 Fix some cosmetic issues with the use of kmem_malloc() in the i386 LDT
sysarch(2) code.

Use M_ZERO instead of explicit bzero(9).  Do not check for failed
allocation when M_WAITOK is specified (which is specified always).
Use malloc(9) when allocating memory for the intermediate copy of the
user-supplied buffer.

Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2014-06-15 05:01:35 +00:00
Konstantin Belousov
2e501b0a9e Use vn_io_fault for the writes from core dumping code. Recursing into
VM due to copyin(9) faulting while VFS locks are held is
deadlock-prone there in the same way as for the write(2) syscall.

Reported and tested by:	pho
Sponsored by:	The FreeBSD Foundation
MFC after:	2 weeks
2014-06-15 04:51:53 +00:00
Bryan Venteicher
9945063d70 Don't check the dma map address for a static DMA mapping against NULL
to determine if the mapping is valid.

Submitted by:	jhb
2014-06-14 23:36:17 +00:00
Alexander Motin
ad9cb3314a Remove non-functional remnants of control LUN -- 18MB of RAM for nothing. 2014-06-14 20:25:14 +00:00
Gavin Atkinson
deee077e0c Stop telling people to use send-pr; instead point them towards Bugzilla.
Hat:		bugmeister@
MFC after:	3 days
2014-06-14 18:43:48 +00:00
Alexander Motin
57a5db13b7 Implement small KPI to access LUN options instead doing it by hands.
MFC after:	2 weeks
2014-06-14 17:47:44 +00:00
Alexander Motin
475c3a653d Fix/improve fhe_stats sysctl output.
MFC after:	2 weeks
2014-06-14 12:26:12 +00:00
Adrian Chadd
43f2450f5d Shadow RAM support is apparently not at all working on the Intel 5300 Wifi.
Tested:

* Intel 5300 3x3 wifi
* Intel 5100
2014-06-14 06:54:03 +00:00
Xin LI
e1e766bcc4 Correct variable for loader tunable variable hw.mfi.mrsas_enable.
Submitted by:	Sascha Wildner (DragonFly)
MFC after:	3 days
2014-06-13 22:33:41 +00:00
John Baldwin
0f698bd800 - Don't check the dma map address for a static DMA mapping against NULL
to determine if the mapping is valid.
- Don't explicitly destroy the dmap map for static DMA mappings.

Reviewed by:	adrian
2014-06-13 19:40:02 +00:00
John Baldwin
8d461aafd8 Don't check maps for static DMA allocations against NULL to see if an
allocation succeeded.  Instead, check the tag against NULL.
2014-06-13 19:34:34 +00:00
John Baldwin
aced5239b2 Add missing calls to bus_dmamap_unload() when freeing static DMA
allocations.

Reviewed by:	scottl
2014-06-13 18:20:44 +00:00
Hans Petter Selasky
36716419f3 Attach the CUSE library and kernel module to the default FreeBSD
builds. Bump the FreeBSD version number.
2014-06-13 08:53:49 +00:00
Alexander Motin
9e005bbcc9 Fix some leaks on LUN creation error.
MFC after:	2 weeks
2014-06-12 21:50:46 +00:00
John-Mark Gurney
4f4bd09dd1 clear the write bit... This allows my AVILA board to survive a
portsnap extract, where previously it would panic..  clearly someone
who knows pmap should optimize this code per alc's comment...

Submitted by:	alc
MFC after:	probably
2014-06-12 16:31:15 +00:00
Hans Petter Selasky
648ad2e720 - Fix out of range shifting bug in bitops.h.
- Make code a bit easier to read by adding parenthesis.

MFC after:	3 days
Sponsored by:	Mellanox Technologies
2014-06-12 13:33:01 +00:00
Alexander Motin
781c93d405 Implement simple direct-mapped cache for popular filesystem identifiers to
avoid congestion on global mountlist_mtx mutex in vfs_busyfs(), while
traversing through the list of mount points.

This change significantly improves NFS server scalability, since it had
to do this translation for every request, and the global lock becomes quite
congested.

This code is more optimized for relatively small number of mount points.
On systems with hundreds of active mount points this simple cache may have
many collisions.  But the original traversal code in that case should also
behave much worse, so we are not loosing much.

Reviewed by:	attilio
MFC after:	2 weeks
Sponsored by:	iXsystems, Inc.
2014-06-12 12:43:48 +00:00
Alexander Motin
1aa6c75827 Introduce new "256 Bucket" zone to split requests and reduce congestion
on "128 Bucket" zone lock.

MFC after:	2 weeks
Sponsored by:	iXsystems, Inc.
2014-06-12 11:57:07 +00:00
Ruslan Bukin
d17024dfd7 Enable SMP for both Exynos5 models we support. 2014-06-12 11:44:26 +00:00
Ruslan Bukin
f0687d366a Activate IRQ 30 (non-secure private timer IRQ) for case we are running
in non-secure state.
2014-06-12 11:41:04 +00:00
Ruslan Bukin
f098889377 Add machine-dependent SMP part for Exynos5420 2014-06-12 11:37:38 +00:00
Alexander Motin
20d3ab87cd Allocating new bucket for bucket zone, never take it from the zone itself,
since it will almost certanly fail.  Take next bigger zone instead.

This situation should not happen with original bucket zones configuration:
"32 Bucket" zone uses "64 Bucket" and vice versa.  But if "64 Bucket" zone
lock is congested, zone may grow its bucket size and start biting itself.

MFC after:	2 weeks
Sponsored by:	iXsystems, Inc.
2014-06-12 11:36:22 +00:00
John Baldwin
896ae82d59 Remove some Giant unlocks I missed in r267290. 2014-06-12 11:21:18 +00:00
John Baldwin
46e89834dc - Don't compare bus_dma map pointers for static DMA allocations against
NULL to determine if bus_dmamap_unload() or bus_dmamem_free() should be
  called.  Instead, check the associated bus and virtual addresses.
- Don't clear static DMA maps to NULL.

Reviewed by:	jfv
2014-06-12 11:15:19 +00:00