This is an attempt to impose some sensible coherent ordering on the
cpu namespace, where previously there was none (we were at the mercy
of BIOS ordering, which varies wildly across systems).
Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
Change p2m infrastructure to operate on per-p2m instead of per-domain.
This allows us to use multiple p2m tables per-domain.
Signed-off-by: Christoph Egger <Christoph.Egger@amd.com>
Acked-by: Tim Deegan <Tim.Deegan@citrix.com>
Currently, there has a logic to check whether the EPT GFN is exceeding
guest physical address width. It uses right shift(>>) to implement the
check. But the right shift count is greater than the width of the
type(unsigned long = 32) under the PAE. And this will cause guest boot
fail under PAE with EPT supported.
Signed-off-by: Li Xin <xin.li@intel.com>
Signed-off-by: Zhang Yang <yang.z.zhang@intel.com>
Corrects an error introduced with cs 21422.
Without the patch my machine crashed when removing a cpu from a
cpupool other than Pool-0.
Signed-off-by: Juergen Gross <juergen.gross@ts.fujitsu.com>
VMs that don't use their full timeslice are guaranteed to flip back
and forth between "active" and "inactive". If we set credit to 0
when setting "inactive", then when the VM comes back to "active"
again, it will effectively be behind most other vcpus in credit.
This causes the credit1 to effectively discriminate *against*
VMs which use less than their full timeslice.
Instead of setting credit to 0, divide it in half. This gets rid of
some of the system credit while allowing non-cpu-bound VMs to keep
some priority advantage.
Signed-off-by: George Dunlap <george.dunlap@eu.citrix.com>
This patch implements 'yield' for credit1. It does this by attempting
to put yielding vcpu behind a single lower-priority vcpu on the
runqueue. If no lower-priority vcpus are in the queue, it will go at
the back (which if the queue is empty, will also be the front).
Runqueues are sorted every 30ms, so that's the longest this priority
inversion can happen.
For workloads with heavy concurrency hazard, and guest which implement
yield-on-spinlock, this patch significantly increases performance and
total system throughput.
Signed-off-by: George Dunlap <george.dunlap@eu.citrix.com>
When ioapic_rte_to_remap_entry fails, currently it just writes value
to ioapic. But the 'mask' bit may be changed if it writes to the upper
half of RTE. This patch ensures to recover the original value of 'mask'
bit in this case.
Signed-off-by: Weidong Han <weidong.han@intel.com>
At the end of io_apic_write_remap_rte, it writes new entry (remapped
interrupt) to ioapic. But it writes low 32 bits before high 32 bits,
it unmasks interrupt before writing high 32 bits if 'mask' bit in low
32 bits is cleared. Thus it may result in issues. This patch fixes
this issue by writing high 32 bits before low 32 bits.
Signed-off-by: Jiang, Yunhong <yunhong.jiang@intel.com>
Signed-off-by: Weidong Han <weidong.han@intel.com>
Extended PCI configuration space is not accessible on 32 bit
systems. This patch disables IBS feature on 32 bit to avoid
IBSCTL(0x1cc) accesses.
Signed-off-by: Wei Wang <wei.wang2@amd.com>
Creates a shared memory ring buffer and event channel in HVM domains
for passing debug messages from PV drivers on HVM guests
The console is used by windows pv drivers to send debug data to xenconsoled
Signed-off-by: Owen Smith <owen.smith@citrix.com>
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Fixes these valgrind reported leaks, found with "valgrind xl create -n ..."
==21170== 8 bytes in 3 blocks are definitely lost in loss record 1 of 3
==21170== at 0x4022F0A: malloc (vg_replace_malloc.c:236)
==21170== by 0x411B22F: strdup (in /lib/i686/cmov/libc-2.7.so)
==21170== by 0x4030085: xlu__cfgl_strdup (libxlu_cfg.c:290)
==21170== by 0x402F3C4: xlu__cfg_yylex (libxlu_cfg_l.l:37)
==21170== by 0x402DD86: xlu__cfg_yyparse (libxlu_cfg_y.c:1338)
==21170== by 0x40308AE: xlu_cfg_readdata (libxlu_cfg.c:85)
==21170== by 0x804DBE4: parse_config_data (xl_cmdimpl.c:591)
==21170== by 0x8056EE4: create_domain (xl_cmdimpl.c:1381)
==21170== by 0x80582AE: main_create (xl_cmdimpl.c:3178)
==21170== by 0x804B54B: main (xl.c:76)
==21170==
==21170== 57 bytes in 2 blocks are definitely lost in loss record 2 of 3
==21170== at 0x4022F0A: malloc (vg_replace_malloc.c:236)
==21170== by 0x402FE22: xlu__cfgl_dequote (libxlu_cfg.c:307)
==21170== by 0x402F4B4: xlu__cfg_yylex (libxlu_cfg_l.l:52)
==21170== by 0x402DD86: xlu__cfg_yyparse (libxlu_cfg_y.c:1338)
==21170== by 0x40308AE: xlu_cfg_readdata (libxlu_cfg.c:85)
==21170== by 0x804DBE4: parse_config_data (xl_cmdimpl.c:591)
==21170== by 0x8056EE4: create_domain (xl_cmdimpl.c:1381)
==21170== by 0x80582AE: main_create (xl_cmdimpl.c:3178)
==21170== by 0x804B54B: main (xl.c:76)
==21170==
==21170== 111 bytes in 6 blocks are definitely lost in loss record 3 of 3
==21170== at 0x4022F0A: malloc (vg_replace_malloc.c:236)
==21170== by 0x402FE22: xlu__cfgl_dequote (libxlu_cfg.c:307)
==21170== by 0x402F4ED: xlu__cfg_yylex (libxlu_cfg_l.l:56)
==21170== by 0x402DD86: xlu__cfg_yyparse (libxlu_cfg_y.c:1338)
==21170== by 0x40308AE: xlu_cfg_readdata (libxlu_cfg.c:85)
==21170== by 0x804DBE4: parse_config_data (xl_cmdimpl.c:591)
==21170== by 0x8056EE4: create_domain (xl_cmdimpl.c:1381)
==21170== by 0x80582AE: main_create (xl_cmdimpl.c:3178)
==21170== by 0x804B54B: main (xl.c:76)
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
1. Try to allocate from nodes containing CPUs which a guest can be
scheduled on.
2. Remember which node we allocated from last, and round-robin
allocations among above-mentioned nodes.
Signed-off-by: Keir Fraser <keir.fraser@citrix.com>
NOTE: This functionality depends on a corresponding qemu-dm patch to work as
expected. Should be safe to use with an un-patched qemu-dm as before.
libxl_wait_for_device_model can only wait for one status value, re-work the
API so that a callback function can chose between several different possible
status values for qemu-dm and fix up all callers appropriately.
In the case of PCI device insert we succeed if qemu-dm reports
"pci-device-inserted" and error out instead of hanging forever if it fails
since qemu-dm now reports a status of "pci-insert-failed".
Signed-off-by: Gianni Tedesco <gianni.tedesco@citrix.com>
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Move to state-machine parser for PCI BDF's now that the number of possible
sscanf expressions is sufficiently large to make it worth it. Also implement
parsing of virtual PCI slot numbers.
Signed-off-by: Gianni Tedesco <gianni.tedesco@citrix.com>
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Introduce a new libxl call libxl_device_pci_parse_bdf() and use it
consistently.
This patch also fixes an infinite loop bug in xl create. If there is a
parse-error on any pci config file entry then xl will attempt to skip it, but
since num_pcidevs is used to index the config list and is not incremented when
skipping, this caused an infinite loop. Solve that problem by introducing a new
loop counter variable.
Note that virtual PCI slots are not parsed by the new code. However this is
a step towards support for virtual slots and multi-function device assignment
since only one location in the code will need to be changed now.
Signed-off-by: Gianni Tedesco <gianni.tedesco@citrix.com>
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Get rid of scan_sys_pcidir() and open-code it inside
libxl_device_pci_list_assignable() since it's not a generically re-useable
function and we're not supporting pcistub driver now. Also use macros for sysfs
dirs in libxl_device_pci_reset
Signed-off-by: Gianni Tedesco <gianni.tedesco@citrix.com>
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Found using "valgrind xl create -n ..." and "valgrind xl create -e ..."
freeing config_data solves:
==18276== 944 bytes in 1 blocks are definitely lost in loss record 12 of 12
==18276== at 0x4022F0A: malloc (vg_replace_malloc.c:236)
==18276== by 0x404AEC1: libxl_read_file_contents (libxl_utils.c:258)
==18276== by 0x8056865: create_domain (xl_cmdimpl.c:1314)
==18276== by 0x8057E2D: main_create (xl_cmdimpl.c:3135)
==18276== by 0x804B2FB: main (xl.c:76)
==18276==
Adding free_domain_config() solves the following (plus presumably others
which didn't trigger because I have no devices of that type).
d_config->disks:
==18276== 61 (32 direct, 29 indirect) bytes in 1 blocks are definitely lost in loss record 9 of 12
==18276== at 0x4022F0A: malloc (vg_replace_malloc.c:236)
==18276== by 0x4022F94: realloc (vg_replace_malloc.c:525)
==18276== by 0x804E2D3: parse_config_data (xl_cmdimpl.c:715)
==18276== by 0x8056A7C: create_domain (xl_cmdimpl.c:1347)
==18276== by 0x8057E2D: main_create (xl_cmdimpl.c:3135)
==18276== by 0x804B2FB: main (xl.c:76)
d_config->vifs:
==18276== 76 (48 direct, 28 indirect) bytes in 1 blocks are definitely lost in loss record 10 of 12
==18276== at 0x4022F0A: malloc (vg_replace_malloc.c:236)
==18276== by 0x4022F94: realloc (vg_replace_malloc.c:525)
==18276== by 0x804E665: parse_config_data (xl_cmdimpl.c:779)
==18276== by 0x8056A7C: create_domain (xl_cmdimpl.c:1347)
==18276== by 0x8057E2D: main_create (xl_cmdimpl.c:3135)
==18276== by 0x804B2FB: main (xl.c:76)
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Found with "valgrind xl destroy ...":
==16272== 53,248 bytes in 1 blocks are definitely lost in loss record 6 of 6
==16272== at 0x4022249: calloc (vg_replace_malloc.c:467)
==16272== by 0x403FD4A: libxl_list_domain (libxl.c:490)
==16272== by 0x404B901: libxl_name_to_domid (libxl_utils.c:65)
==16272== by 0x804B4D2: domain_qualifier_to_domid (xl_cmdimpl.c:181)
==16272== by 0x804B50F: find_domain (xl_cmdimpl.c:198)
==16272== by 0x804D70C: destroy_domain (xl_cmdimpl.c:2104)
==16272== by 0x8054E4C: main_destroy (xl_cmdimpl.c:2912)
==16272== by 0x804B2FB: main (xl.c:76)
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
This has not caused crashes because generally use after free is OK
provided nothing else is going on. However the patch makes things
correct. It also allows us to use heap poisoning feature of valgrind on
tools linking to libblktapctl.
Signed-off-by: Gianni Tedesco <gianni.tedesco@citrix.com>
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Added a check which allows adding and removal of PCI devices for PV
guests in the absence of an IOMMU.
Signed-off-by: Mihir Nanavati <mihirn@cs.ubc.ca>
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Add IBS support for AMD family 10h processors. The major
implementation is derived from latest Linux. Two hypercalls are added,
which is necessary for IBS feature detection and user mode parameter
read.
Signed-off-by: Wei Wang <wei.wang2@amd.com>
The per-thread hypercall buffer is usually cleaned up on pthread_exit
by the destructor passed to pthread_key_create. However if the calling
application is not threaded then the destructor is never called.
This frees the data for the current thread only but that is OK since
any other threads will be cleaned up by the destructor.
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
Acked-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>
This allows the top level command dispatcher to cleanup some of its
own allocations.
This is a fairly mechanical conversion of exit(FOO) into return FOO
for the top-level command functions (i.e. main_*). There are still
code paths which will exit() further down the call chains which will
require actual thought.
At first glance not all the return codes are which one would normally
expect for process exit codes (e.g. some are negative) but I didn't
attempt to address that here.
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>
xm list takes an optional domain argument; make xl list do likewise.
This also gets rid of a small amount of code which was duplicated
between list_domains and list_domains_details.
Signed-off-by: Ian Jackson <Ian.Jackson@eu.citrix.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>
If the CPUID limit bit in MSR_IA32_MISC_ENABLE is set, clear it to
make all CPUID information available. This is required for some
features to work, such as MWAIT in cpuidle, get cpu topology, XSAVE,
etc.
Signed-off-by: Wei Gang <gang.wei@intel.com>
This tool reads the CPU save records, overwrites RIP with a bogus
value, and then restores them. This is, of course, not guaranteed
to crash the guest (since the CPUs may not be executing in kernel
at the time) but it's good for breaking into some tight loops that
would be hard to debug otherwise.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>
Implement a new libxl function libxl_device_pci_list_assignable. This is
used to implement the xl list-assignable-pci-devices command and part of
the implementation is used to make sure that PCI devices are not multiply
assigned to one or more domU's before doing the passthrough assignment.
The function libxl_device_pci_list changes to libxl_device_pci_list_assigned
due to a parameter change for consistency with pci_list_assignable.
Signed-off-by: Gianni Tedesco <gianni.tedesco@citrix.com>
Acked-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>
Keir Fraser
Stefano Stabellini
Juergen Gross
Ian Campbell
Ian Campbell
Gianni Tedesco
Christoph Egger
Mihir Nanavati
Zhigang Wang
Ian Jackson
Paul Durrant