structure because the first field is 32 bits and the next is 64 bits,
whereas x86-32 doesn't pad. The quick fix is to read the cpu id
separately as a 32bit value, then read the rest of the trace record.
Here is a little patch that does that. Tested on x86/32 SMP and x86/64.
Signed-off-by: Rob Gardner <rob.gardner@hp.com>
a high-memory page. Otherwise stale writable mappings may
persist. This fixes a corner case for alloc_ldt (which can
vmalloc an LDT and hence use highmem pages).
Signed-off-by: Keir Fraser <keir@xensource.com>
1. Use update_va_mapping where possible (much faster)
2. If that fails, fall back to explicit pte update rather
than using writable pagetables (very slow for singleton
updates)
Also many x86/64 cleanups and fix to handle vmalloc space
correctly.
Signed-off-by: Keir Fraser <keir@xensource.com>
seriously harm Xen's health. Better to just disable it in the
tools until the outstanding issues are resolved.
Signed-off-by: Keir Fraser <keir@xensource.com>
This is the first change to the ramdisk since v0.1.0. So, we now build
the image as initrd-X.Y.img, and link initrd.img to it. This lets us
have a few checks to make sure that people rebuild their ramdisks when
necessary.
uses of goto that reduce down to return NULL in any case, breaking the common
code out for returning values based upon the success or failure of an
operation, and the common code that parses transaction and path parameters,
used by xspy_read, xspy_ls, and xspy_mkdir. Break out also the common code
that removes a watch token from the XsHandle watches list.
Use bool rather than int for result values, where this matches the same use
by tools/xenstore/xs.c.
Added missing free in xspy_get_permissions.
Use PySequence_SetItem inside xs_watch, rather than Py_INCREF followed by
PyList_SetItem. If there is no None entry in the watches list, the code goes
on to append to the list using PyList_Append, and that call takes its own
reference rather than stealing one like PyList_SetItem. This means that we
were previously leaking a reference if the list was full and so PyList_Append
was necessary.
Use PySequence_SetItem rather than Py_INCREF/PyList_SetItem also where we are
using assigning Py_None into the list, for neatness.
Remove meaningless pipe from xspy_release_domain's arg_spec.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
checks in entry.S. This will avoid a hard-to-debug loop
that we found ourselves in recently, involving
hypercall_preempt_check() spuriously returning TRUE, but
no event getting propagated to the guest (since mask != 0).
Signed-off-by: Keir Fraser <keir@xensource.com>
buffer that has grown beyond max_capacity. This fixes a Xenconsoled segfault
that was caused by stomping around above the buffer's allocated region. It
became possible to set the max_capacity with changeset 7431, and ever
since then we have been exposed to this bug. It would most easily be triggered
by running a domain without a client attached to the console, so that the
max_capacity was reached more easily.
Closes bug #380.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
one level-1 pagetable page on PAE and x86/64. Generalise
the existing code to allocate and map an appropriate number
of level-1 pagetable pages.
Signed-off-by: Jun Nakajima <jun.nakajima@intel.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
Twisted Matrix is no longer necessary for Xen. Removed from the
pre-requisites listed in the user documentation.
Signed-off-by: Andrew D. Ball <aball@us.ibm.com>
optional. The USB driver doesn't need this entrypoint because the state
machine driven by it is factored out of the driver code entirely into
xenidc.
Signed-off-by: Harry Butterworth butterwo@uk.ibm.com
We need to attach a console before rebooting to make sure the domain did
actually boot in the first place. Also, use a better method for pulling out
the uptime string.
internal state if the store is configured by an external tool.
We react properly to changes in name and on_{reboot,poweroff,crash}. Reacting
correctly to vcpus, vcpu_avail, memory, and maxmem will come soon (those code
aspects are already under review). cpu_weight and bootloader to be considered.
Removed unused and semantically invalid method XendDomainInfo.setDomid.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
