Kernel doesn't use it, and it could be regenerated in the kernel if it did need it.
This also unlocks the apic range the bios can use. Previously the apic ids would have
to fit within 0..MAX_CPUS or it'd reject the cpu. Some boxes (mine in particular)
seem to sparsely populate the apic id so that the range is pretty large.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37108 a95241bf-73f2-0310-859d-f6bbb57e9c96
phys_addr_t. IOW, if PAE is enabled, that memory should be put to use now.
Apparently we report an incorrect amount of total memory (also counting
memory gaps), which also suggests that we need another method to manage the
vm_page structures (currently a huge array with indexes proportional to
physical page addresses, i.e. wasting memory for the gaps).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37107 a95241bf-73f2-0310-859d-f6bbb57e9c96
need a spinlock per CPU; a single variable suffices.
* Extended call_all_cpus[_sync]() to work before smp_wake_up_non_boot_cpus()
(even before smp_init()).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37105 a95241bf-73f2-0310-859d-f6bbb57e9c96
maps. Now we can at least fully boot in qemu with one CPU. A few things
still need to be implemented, though.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37090 a95241bf-73f2-0310-859d-f6bbb57e9c96
page with a 32 bit physical address (needed for the PDPTs). A small set of
free pages is cached, so the rather expensive vm_page_allocate_page_run() can
be avoided most of the time.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37088 a95241bf-73f2-0310-859d-f6bbb57e9c96
* The 32 bit paging structures inherited from the boot loader are translated
to PAE structures and PAE is enabled.
* Implemented the initialization of the initial PhysicalPageSlotPool, so that
the physical page mapper can be initialized.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37076 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Made all helper function operating on paging structures static methods of
X86PagingMethod32Bit.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37064 a95241bf-73f2-0310-859d-f6bbb57e9c96
accessed from anywhere. Added static X86PagingMethod32Bit::Method()
returning it as the subtype pointer -- to be used in the code related to
that method only, of course.
* Made a bunch of static variables non-static members of
X86PagingMethod32Bit and added accessors for them. This makes them
accessible in other source files (allowing for more refactoring) and saves
memory, when we actually have another paging method implementation.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37062 a95241bf-73f2-0310-859d-f6bbb57e9c96
was used.
* Renamed X86VMTranslationMap to X86VMTranslationMap32Bit and pulled the paging
method agnostic part into new base class X86VMTranslationMap.
* Moved X86PagingStructures into its own header/source pair.
* Moved pgdir_virt from X86PagingStructures to X86PagingStructures32Bit where
it is actually used.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37055 a95241bf-73f2-0310-859d-f6bbb57e9c96
agnostic:
* Changed return value of X86PhysicalPageMapper::[Interrupt]GetPageTableAt()
from page_table_entry* to void*.
* Made PhysicalPageSlotPool an abstract base class. The paging method provides
an implementation and creates the pools now.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37054 a95241bf-73f2-0310-859d-f6bbb57e9c96
Introduced the interface X86PagingMethod which is used by those. ATM
there's one implementing class, X86PagingMethod32Bit.
* Made X86PagingStructures a base class, with one derived class,
X86PagingStructures32Bit.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37050 a95241bf-73f2-0310-859d-f6bbb57e9c96
would be skipped. Concerned non-x86 only, as for x86 the methods are
overridden.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37031 a95241bf-73f2-0310-859d-f6bbb57e9c96
the potential input parameter (virtual address) rather than its address. As
suggested by Andreas Faerber in #6141.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37029 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Renamed i386_context_switch() to x86_context_switch().
* x86_context_switch() no longer sets the page directory.
arch_thread_context_switch() does that explicitly, now. This allows to solve
the TODO by reordering releasing the previous paging structures reference and
setting the new page directory.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37024 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Renamed vm_translation_map_arch_info to X86PagingStructures, and all
members and local variables of that type accordingly.
* arch_thread_context_switch(): Added TODO: The still active paging structures
can indeed be deleted before we stop using them.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37022 a95241bf-73f2-0310-859d-f6bbb57e9c96
That makes the automatic stack traces in case of kernel panics work.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37021 a95241bf-73f2-0310-859d-f6bbb57e9c96
vm_translation_map_arch_info to X86VMTranslationMap where they actually
belong.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37014 a95241bf-73f2-0310-859d-f6bbb57e9c96
vm_free_unused_boot_loader_range(): Don't free any memory beyond the given
range.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37010 a95241bf-73f2-0310-859d-f6bbb57e9c96
even have to be available in the generic case. See bug #6105, patch 1 & 2.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37007 a95241bf-73f2-0310-859d-f6bbb57e9c96
warnings, but also some oversights from earlier changes.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37000 a95241bf-73f2-0310-859d-f6bbb57e9c96
that are wide enough for both virtual and physical addresses.
* DMABuffer, IORequest, IOScheduler,... and code using them: Use
generic_io_vec and generic_{addr,size}_t where necessary.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36997 a95241bf-73f2-0310-859d-f6bbb57e9c96
of having the logic be triggered by IOScheduler::SetDeviceCapacity(), as that
one might actually be called more often (for each call to update_capacity(),
ie. each B_GET_GEOMETRY/B_GET_DEVICE_SIZE will trigger it), and there is no
reason to throw away the cache every time (will make a difference during
partition/file system detection).
* In cd_init_device() just call update_capacity() instead of duplicating its
code.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36986 a95241bf-73f2-0310-859d-f6bbb57e9c96
map_physical_memory()'s physicalAddress parameter type from void* to
phys_addr_t. This breaks source compatibility, but -- as long as
phys_{addr,size}_t remain 32 bit wide -- keeps binary compatibility with
BeOS.
* Adjusted all code using the affected interfaces (Oh what fun!). Added a few
TODOs in places where the wrong types (e.g. void* for physical addresses
are used). Looks like quite a few drivers aren't 64 bit safe and others
will break with PAE.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36960 a95241bf-73f2-0310-859d-f6bbb57e9c96
address ranges, and a set of support functions working with it.
* Changed the type of the kernel_args physical address range arrays to
phys_addr_range and adjusted the code working with those.
* Removed a bunch of duplicated address range code in the PPC's mmu.cpp.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36947 a95241bf-73f2-0310-859d-f6bbb57e9c96
where appropriate.
* Typedef'ed page_num_t to phys_addr_t and used it in more places in
vm_page.{h,cpp}.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36937 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Don't keep any memory mappings from the OF for the time being. We can't
keep mappings < 2 GB, since those aren't in the kernel address space and
we don't handle memory mapped registers or the like correctly either.
Ticket #5193.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36886 a95241bf-73f2-0310-859d-f6bbb57e9c96
created, and moved the heap's grow and VIP heap initialization to it. Should
fix#5956.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36855 a95241bf-73f2-0310-859d-f6bbb57e9c96
The old mechanism to route an apic id back to a cpu id is faulty, built with the assumption that
the bios will 'pack' the apic ids from 0-num_cpus. In systems that dont do that, the code would
randomly corrupt the bootloader. Fatal in this case.
This quick fix simply rejects all apic ids >= MAX_CPUS (8). No way it would have worked before
if you had a box that started with >= 8 or anything, so it shouldn't regress any existing system.
Better solution is to allow any apic id to exist (0-255).
On this particular box the ids (from lunix dmesg):
SRAT: PXM 0 -> APIC 0 -> Node 0
SRAT: PXM 1 -> APIC 16 -> Node 1
SRAT: PXM 0 -> APIC 2 -> Node 0
SRAT: PXM 0 -> APIC 4 -> Node 0
SRAT: PXM 0 -> APIC 6 -> Node 0
SRAT: PXM 1 -> APIC 18 -> Node 1
SRAT: PXM 1 -> APIC 20 -> Node 1
SRAT: PXM 1 -> APIC 22 -> Node 1
SRAT: PXM 0 -> APIC 1 -> Node 0
SRAT: PXM 0 -> APIC 3 -> Node 0
SRAT: PXM 0 -> APIC 5 -> Node 0
SRAT: PXM 0 -> APIC 7 -> Node 0
SRAT: PXM 1 -> APIC 17 -> Node 1
SRAT: PXM 1 -> APIC 19 -> Node 1
SRAT: PXM 1 -> APIC 21 -> Node 1
SRAT: PXM 1 -> APIC 23 -> Node 1
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36836 a95241bf-73f2-0310-859d-f6bbb57e9c96
platforms that are ignorant of the debug syslog feature.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36748 a95241bf-73f2-0310-859d-f6bbb57e9c96
map_backing_store() doesn't commit memory when this flag is given.
* Used the new flag vm_copy_area(): We no longer commit memory for read-only
areas. This prevents read-only mapped files from suddenly requiring memory
after fork(). Might improve the situation on machines with very little RAM
a bit.
We should probably mark writable copies over-committing, since the usual
case is fork() + exec() where the child normally doesn't need more than a
few pages until calling exec(). That would significantly reduce the memory
requirement for jamming the Haiku tree.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36651 a95241bf-73f2-0310-859d-f6bbb57e9c96
idle_scan_active_pages() cannot satisfy this assertion anymore. Added TODO
for a better fix which can wait until after the release.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36649 a95241bf-73f2-0310-859d-f6bbb57e9c96
array at least work with the part intersecting with the array. Log those
quasi error cases.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36633 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Now it checks whether the given physical range exists at all.
* Not only check whether any range of the given size is free, but also check
whether the given range is free.
* B_BAD_VALUE is not a particularly good return value for a function supposed
to return a bool
Fixes#5911. With only 64 MB RAM the usual debug syslog buffer location may
be outside the actually existing physical memory.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36632 a95241bf-73f2-0310-859d-f6bbb57e9c96
CREATE_AREA_DONT_WAIT flag; waiting at this point is not allowed.
* I hope I found all occurences, but there might be some areas left (note,
only those that don't use B_ALREADY_WIRED are problematic).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36624 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Separated the other stuff previously done in debug_init_post_vm() to the new
debug_init_post_settings().
* Removed superfluous status_t return codes - they are ignored, anyway, and if
there really is a show stopper in the init process, panicking would be the
thing one should do.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36623 a95241bf-73f2-0310-859d-f6bbb57e9c96
Since we also force uncacheability via the PTEs, it is no longer necessary to
use MTRRs for uncacheable ranges. It is recommended, which is why we still do
it, if possible, but when we run out of MTRRs, the algorithm ignores
uncacheable ranges now -- iteratively it increases the minimum size of ranges
to drop until the available MTRRs suffice.
Should fix#5703 and #5865.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36591 a95241bf-73f2-0310-859d-f6bbb57e9c96
memory. This avoids implementation defined behavior on Pentium Pro/II when
intersecting with an write-combining or write-protected MTRR range.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36589 a95241bf-73f2-0310-859d-f6bbb57e9c96
PAT support (i.e. uncacheable, write-through, and write-back). Has pretty
much no effect ATM, as the MTRRs restrict the types to what is actually
requested.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36583 a95241bf-73f2-0310-859d-f6bbb57e9c96
implemented for any architecture yet.
* vm_set_area_memory_type(): Call VMTranslationMap::ProtectArea() to change the
memory type for the already mapped pages.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36574 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Don't set the VMArea's memory type in arch_vm_set_memory_type(), but let the
callers do that.
* vm_set_area_memory_type(): Does nothing, if the memory type doesn't change.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36573 a95241bf-73f2-0310-859d-f6bbb57e9c96
of consistency.
* Moved the B_OVERCOMMITTING_AREA flag from B_KERNEL_AREA_FLAGS to
B_USER_AREA_FLAGS, since we really allow it to be passed from userland.
* Most VM syscalls check the provided protection against B_USER_AREA_FLAGS
instead of B_USER_PROTECTION, now. This way they allow for
B_OVERCOMMITTING_AREA as well.
* _user_map_file(), _user_set_memory_protection(): Check the protection like
the other syscalls do and use fix_protection() instead of doing that
manually.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36572 a95241bf-73f2-0310-859d-f6bbb57e9c96
continuing to run. This must not happen since the thread is the boot CPU's
idle thread performing the early kernel initialization (before the scheduler
is started).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36560 a95241bf-73f2-0310-859d-f6bbb57e9c96
fixes the problem that the CPU initiating the call could make the next call
and reset sCpuRendezvous2 before the other CPUs have returned from their
smp_cpu_rendezvous(). Probably virtually impossible on real hardware, but I
could almost reliably reproduce it with qemu -smp 2 (would hang the late boot
process without ability to enter KDL).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36559 a95241bf-73f2-0310-859d-f6bbb57e9c96
boot CPU wait until all other CPUs are ready to wait. This solves a
theoretical problem in main(): The boot CPU could run fully through the early
initialization and reset sCpuRendezvous2 before the other CPUs left
smp_cpu_rendezvous(). It's very unlikely on real hardware that the non-boot
CPUs are so much slower, but it might be a concern in emulation.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36558 a95241bf-73f2-0310-859d-f6bbb57e9c96
kernel from committing memory for all read-only segments until we're done
relocating. This allows Haiku to boot on machines with less RAM and swap
disabled. At least in qemu I could boot with 100 MB and start WebPositive.
Probably fixed#5822.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36552 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Commit(): Unreserve memory when asked to shrink the commitment.
* Fault(): The whole logic is flawed, since this is always called by
vm_soft_fault(), even, if the page is finally mapped from a lower cache.
Now we do at least limit our commitment to (page_count + 1) * B_PAGE_SIZE
instead of always reserving memory for another page.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36551 a95241bf-73f2-0310-859d-f6bbb57e9c96
maps. The new cache will be created in over-committing mode.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36550 a95241bf-73f2-0310-859d-f6bbb57e9c96
scheduler. This avoids the need to use the send_signal_etc() work-around for
resume_thread() during the early kernel initialization. Might fix#5851.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36530 a95241bf-73f2-0310-859d-f6bbb57e9c96
supporting ID array (just that it only contains one entry).
* Fixed missing malloc() result check in embedded_controller_init_driver().
* Style fixes.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36529 a95241bf-73f2-0310-859d-f6bbb57e9c96
already initialized in syslog_init(), if the debug syslog feature is enabled.
Since the area would never be created, the pointer becomes invalid, however,
and could cause a double fault on entering KDL.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36528 a95241bf-73f2-0310-859d-f6bbb57e9c96
precondition for calling the function it read/write accessed the fAreaHint
attribute in a non-atomic manner. I.e. if executed concurrently by two
threads of the same team one could return the wrong area. The most likely
problem could be caused in vm_soft_fault(), leading to pages being added to
the wrong caches.
The bug itself is inherited from NewOS, but is triggered way more likely
since the page daemon unmaps inactive pages (r35485). Probably fixes#5413.
Might also fix the sporadically occurring crashes/asserts in the hoard
allocator -- at least inactive pages being replaced by zeroed ones would be
an excellent explanation.
[Thanks to Matt for providing remote debug access to his machine.]
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36523 a95241bf-73f2-0310-859d-f6bbb57e9c96
vm_page_schedule_write_page_range() to prevent more memory pressure.
* While it seems to help a bit, it doesn't solve the problem of bug #5777.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36519 a95241bf-73f2-0310-859d-f6bbb57e9c96
it is non-NULL first.
* syslog_init_post_vm(): Make sure the area for the debug syslog could actually
be created and fail otherwise.
* syslog_init_post_threads(): Fixed the cleanup on error. In case of a debug
syslog we need to delete the respective area instead of free()ing the
memory.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36497 a95241bf-73f2-0310-859d-f6bbb57e9c96
free respective free and cached pages.
* Removed the unused vm_page_allocate_page_run_no_base().
* vm_page_allocate_page_run() (and allocate_page_run()):
- Use vm_page_reserve_pages() instead of vm_page_try_reserve_pages(), i.e.
wait until the reservation succeeds.
- Now we iterates two times through the pages to find a suitable page run. In
the first iteration it only looks for free/clear pages, in the second
iteration it also considers cached pages. This increases the chance of the
function to succeed, when a lot of caching is going on.
This reduces the amount of memory required to use the IOCache when booting
off the anyboot Live CD to around 160 MB in qemu. It also seems to work with
128 MB, but the syslog indicates that some memory allocations fail, which
is not exactly inspiring confidence.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36489 a95241bf-73f2-0310-859d-f6bbb57e9c96
before locking the address space. That gives us a bit more flexibility in
vm_page_allocate_page_run().
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36486 a95241bf-73f2-0310-859d-f6bbb57e9c96
caching (similar to the file cache) instead of contiguous areas. This is
probably a little bit slower, but integrates better with the VM -- the
caching doesn't increase memory pressure and the least recently used pages
will automatically be recycled when needed.
There are still memory allocation issues on machines with little memory. The
USB stack apparently tries to allocate a rather big chunk of contiguous
memory, which fails when all not otherwise bound memory is used for caches,
since the VM functions for allocating contiguous memory consider only free
pages ATM.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36458 a95241bf-73f2-0310-859d-f6bbb57e9c96
as a drop-in replacement for IOScheduler, processing IORequests synchronously
in FIFO order. It stores cache lines of user-defined size. Currently for each
cache line an area of contiguous memory is used, which is not optimal.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36436 a95241bf-73f2-0310-859d-f6bbb57e9c96
loop causes the function to return before setting the _numEntries. The
overflow case is checked at the end of the function anyway, so we can just
break out.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36425 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Removed irq_routing struct from ACPI.h to where it is used. Might be able to use acrestype.h instead of duplication.
* Disable old embedded controller and remove it from image.
* Move embedded controller code to busmanager module.
* Remove (some) code duplication
Should work as before, but is now initialized right after the bus manager. Can probably remove further code duplication.
I hope I've done it correctly, feedback is most welcome.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36420 a95241bf-73f2-0310-859d-f6bbb57e9c96
respective constants.
* compare_image_boot(): Compare volume names case insensitively.
* DiskBootMethod::IsBootDevice(): For BOOT_METHOD_CD only accept devices with
removable media.
* DiskBootMethod::IsBootPartition(): Added special recognition for anyboot
CDs. Since their partition types aren't kPartitionTypeDataSession,
compare_cd_boot() didn't prefer them and it was more or less random whether
it was chosen when other Haiku installations where available. Moreover
selecting the CD in the boot loader menu would cause the kernel not to find
it.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36403 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Moved IOScheduler::_IOCallbackWrapper() to IOCallback::WrapperFunction().
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36394 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Changed the semantics of VMCache::HasPage(). It was interpreted
inconsistently by the derived classes. Now it returns whether the backing
store can provide the page (via Read()). The default implementation returns
false. VNodeCache::HasPage() only returns true, if the given offset is
within the cache (i.e. file) bounds. This prevents vm_soft_fault() from
adding clean pages to vnode caches on faults beyond the file bounds.
Probably fixes#5473 -- at least mmap_resize_test behaves correctly, now.
* Removed redundant HasPage() and Fault() overrides in VMCache derived
classes.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36374 a95241bf-73f2-0310-859d-f6bbb57e9c96
swap space when the cache shrinks. Currently the implementation stil leaks
swap space of busy pages.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36373 a95241bf-73f2-0310-859d-f6bbb57e9c96
the CPU info, if x86_exit_user_debug_at_kernel_entry() was executed before,
i.e. if the debug exception occurred in userlands. In all other cases we need
to read the current register values. Fixes#5742, a regression introduced in
r35951.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36340 a95241bf-73f2-0310-859d-f6bbb57e9c96
- Reimplemented mutex to use the _kern_mutex*() syscalls.
- The initializer functions cannot fail anymore -- changed their return type
to void.
- Changed the initializer function semantics to not copy the name by default
anymore (as in the kernel). Also added *_etc() versions of them that take an
additional flags.
- Added static initializer macros.
- Made the mutex (and thus recursive_lock) lock functions non-interruptable.
- Got rid of the "lazy" version. They are no longer needed, since the
initialization of the standard types can be done statically and cannot fail.
* Adjusted libroot, runtime loader, and other code using the private libroot
locking primitives to the new semantics.
* pthreads mutexes and condition variables:
- Reimplemented using the _kern_mutex*() syscalls.
- Consistently use POSIX error codes.
- Fixed some not quite POSIX compliant behavior.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36323 a95241bf-73f2-0310-859d-f6bbb57e9c96
* user_mutex_unlock_locked(): Set or clear the locked flag depending on whether
we wake up a waiter.
* _user_mutex_switch_lock(): The syscall cannot be restartable.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36319 a95241bf-73f2-0310-859d-f6bbb57e9c96
it failed...).
* Moved IRQ table reading much earlier (before starting to program the I/O
APIC), though it currently fails, possibly because the device manager isn't
up yet, and there is no embedded controller driver.
* The kernel now enables I/O APICs by default, but the boot loader disables
them - you can now enable them using the safe mode menu, but it currently
won't have any (positive) effect.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36293 a95241bf-73f2-0310-859d-f6bbb57e9c96
can safely be used.
* Since using the I/O APIC is disabled by default, I've removed the "return"
that prevented its use when enabled. Let's see if it already does anything.
* Adapted other arch_int.cpp with a bit of cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36290 a95241bf-73f2-0310-859d-f6bbb57e9c96
double/deadlock when we delete heap areas because of other area deletions.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36284 a95241bf-73f2-0310-859d-f6bbb57e9c96
The O_CLOEXEC open mode wasn't actually set in the close-on-exec bitmap causing
all files opened with O_CLOEXEC (like done in the storage kit classes) to still
be inherited. This caused for example to be unable to unmount volumes when
opening apps while Tracker touched some files (i.e. copying some large files)
since these apps would inherit the file descriptor and therefore keep the
volume busy.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36273 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Add file descriptor and IO context tracing.
* Some minor cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36272 a95241bf-73f2-0310-859d-f6bbb57e9c96
via exit() (the calling thread will drop into the debugger instead).
* The DebugServer now uses this flag by default.
* Added TODO comment: the default debugger should already be able to set a flag
like this in order to close a race condition between dropping a thread into
the debugger and setting the flag.
* Cleaned up the debug_server sources a bit.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36265 a95241bf-73f2-0310-859d-f6bbb57e9c96
mapped page.
* debug_{mem,strl}cpy():
- Added "team" parameter for specifying the address space the address are
to be interpreted in.
- When the standard memcpy() (with fault handler) fails, fall back to
vm_debug_copy_page_memory().
* Added debug_is_debugged_team(): Predicate returning true, if the supplied
team_id refers to the same team debug_get_debugged_thread() belongs to.
* Added DebuggedThreadSetter class for scope-based debug_set_debugged_thread().
Made use of it in several debugger functions.
* print_demangled_call() (x86): Fixed unsafe memory access.
Allows KDL stack traces to work correctly again, even if the page daemon has
already unmapped the concerned pages.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36230 a95241bf-73f2-0310-859d-f6bbb57e9c96
interrupts (MSI).
* Add the remaining IDT entries and redirection functions in the interrupt code.
* Make the PIC end_of_interrupt() return a result to indicate whether the vector
was handled by this PIC. If it isn't we now issue a apic_end_of_interrupt()
in the assumption of apic local interrupt, MSI or IPI. This also removes
the need for the gUsingIOAPIC global and doing manual apic_end_of_interrupt()
calls in the SMP and timer code.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36221 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Removed rounding up the end address to page alignment. It's not necessary
and could cause an overflow.
* Fixed possible infinite loop triggered by a rare race condition: When two
threads of a team were accessing the same unmapped page at the same time
each would trigger a page fault. One thread would map the page again, the
second would wait until the first one was done and update the page
protection (unnecessarily but harmlessly). If the first thread accessed the
page again at an unfortunate time, it would implicitly change the
accessed/dirty flags of the page's PTE, which was a situation the loop in
Protect() didn't consider and thus run forever.
Seen the problem twice today in form of an app server freeze.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36197 a95241bf-73f2-0310-859d-f6bbb57e9c96
pass stricter allocation flags to VMAddressSpace::{Remove,Delete}Area() or
could deadlock otherwise.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36188 a95241bf-73f2-0310-859d-f6bbb57e9c96
other places where previously the same functionality was duplicated. Also
seperated the header which was originally arch_smp.h into apic.h and arch_smp.h
again as some of it is MP and not actually APIC.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36182 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Implemented a tiny bit more sophisticated version of
estimate_max_scheduling_latency() that uses a syscall that lets the scheduler
decide.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36170 a95241bf-73f2-0310-859d-f6bbb57e9c96
locks.
* Added syscalls for a new kind of mutex. A mutex consists only of an int32 and
doesn't require any kernel resources. So it's initialization cannot fail
(it consists only of setting the mutex value to 0). An uncontended lock or
unlock operation can basically consist of an atomic_*() in userland. The
syscalls (when the mutex is contended) are a bit more expensive than semaphore
operations, though.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36158 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Added new thread flag THREAD_FLAGS_ALWAYS_RESTART_SYSCALL. If set, it forces
syscall restart even when a signal handler without SA_RESTART was invoked.
* Fixed sigwait(): If one of requested signals wasn't already pending it would
never wake up. Also, the syscall always needs to be restarted, if interrupted
by another signal.
* Renamed a bunch of the POSIX signal function implementations which did return
an error code directly (instead via errno). Added correct POSIX functions
where needed.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36054 a95241bf-73f2-0310-859d-f6bbb57e9c96
doesn't mention the expected set_alarm() behavior. Tests under BeOS R5 Bone
crash the system in either case, though, so I guess we're free to go with the
expected POSIX behavior.
This make detection of hanging tests in the Open POSIX Test Suite work.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36046 a95241bf-73f2-0310-859d-f6bbb57e9c96
return it.
* lock_memory_etc(): On error the VMAreaWiredRange object could be leaked.
* [un]lock_memory_etc(): Call VMArea::Unwire() with the cache locked and
explicitly delete the range object after unlocking the cache to avoid
potential deadlocks.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36035 a95241bf-73f2-0310-859d-f6bbb57e9c96
Since the requirement is that the area's top cache is locked, allocating
memory isn't allowed.
* lock_memory_etc(): Create the VMAreaWiredRange object explicitly before
locking the area's top cache.
Fixes#5680 (deadlocks when using the slab as malloc() backend).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36033 a95241bf-73f2-0310-859d-f6bbb57e9c96
* _user_map_file(), _user_unmap_memory(): Verify that the address (if given) is
page aligned.
* Reworked memory locking (wiring):
- VMArea does now have a list of wired memory ranges and supports waiting for
a range to be removed.
- vm_soft_fault():
- Added "wirePage" parameter that, if given, makes the function wire the
page and return it.
- Added "wiredRange" parameter (for calls from lock_memory_etc()) and made
sure we never unmap wired pages. This could e.g. happen when a page from a
lower cache was read-mapped and a write fault occurred. Now in such a
situation the function waits for the page to be unwired and restarts.
- All functions that manipulate areas in a way that could affect wired ranges
do now either require the caller to make sure there are no wired ranges in
the way or do that themselves. Added a few wait_if_*_is_wired() helper
functions for that purpose.
- lock_memory_etc():
- Does now also work correctly when the range spans more than one area.
- Adds VMAreaWiredRanges to the affected VMAreas and retains an address
space reference (so that the address space won't be deleted as long as a
wired range exists).
- Resolved TODO: The area's caches are now locked when
increment_page_wired_count() is called.
- Resolved TODO: The race condition due to missing locking after looking up
the page mapping is now prevented. We hold the cache locks (in case the
page is already mapped) and the new vm_soft_fault() parameter allows us
to get the page wired.
- unlock_memory_etc(): Changes symmetrical to those in lock_memory_etc() and
resolved all TODOs.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@36030 a95241bf-73f2-0310-859d-f6bbb57e9c96
syslog in the current KDL session. Added option "-k" for the former behavior.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35976 a95241bf-73f2-0310-859d-f6bbb57e9c96
Added fields for temporary storage of the debug registers dr6 and dr7 to the
arch_cpu_info structure. The actual registers are stored at the beginning of
x86_exit_user_debug_at_kernel_entry() and read in
x86_handle_debug_exception().
The problem was that x86_exit_user_debug_at_kernel_entry() itself overwrote
dr7 and, if kernel breakpoints were enabled, dr6 could be overwritten anytime
after. So x86_handle_debug_exception() would find incorrect values in the
registers (definitely in dr7) and thus interpret the detected debug condition
incorrectly. Usually watchpoints were recognized as breakpoints.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35951 a95241bf-73f2-0310-859d-f6bbb57e9c96
occurred. This fixes the automatic syscall restart case (would wait forever)
and aligns the behavior with poll().
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35948 a95241bf-73f2-0310-859d-f6bbb57e9c96
arch_debug_stack_trace() through debug_call_with_fault_handler(), since the
one in the CPU structure can only be used for debug_{strl,mem}cpy(), which do
not potentially have nested debug_call_with_fault_handler() calls.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35920 a95241bf-73f2-0310-859d-f6bbb57e9c96
block was freshly allocated. A block returned by _GetUnusedBlock() already
has current_data and we would leak it before.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35918 a95241bf-73f2-0310-859d-f6bbb57e9c96
arch_debug_registers instead.
* Call arch_debug_save_registers() on all CPUs when entering the kernel
debugger.
* Added debug_get_debug_registers() to return a specified CPU's saved
registers.
* x86:
- Replaced the previous arch_debug_save_registers() implementation. Disabled
getting the registers via the gdb interface for the time being.
- Fixed the "sc", "call", and "calling" commands to also work for threads
running on another CPU.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35907 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Print the initial stack trace safely, i.e. in a setjmp() + fault handler
environment.
* Disable pagination while executing the executing the panic() commands. This
way even when it is not possible to use the keyboard we get the full output.
* Added "panic_commands" kernel debugger command, to execute the panic()
commands again (with pagination enabled).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35902 a95241bf-73f2-0310-859d-f6bbb57e9c96
is passed to some v*printf(), we clone it first. This way we can use it
repeatedly. It worked only on certain architectures correctly before.
* Changed the panic() semantics a bit: When the format contains the
(hopefully unusual enough) string "@!" only the part before it is
considered to belong to the panic message. The part after it is interpreted
as a command line to be executed directly after entering KDL. This way it
is possible to provide helpful command invocations only developers somewhat
familiar with the subsystem might come up with. E.g.:
panic("dec_vnode_ref_count: called on busy vnode %p"
"@! vnode -p %p; mount _mount", vnode, vnode);
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35896 a95241bf-73f2-0310-859d-f6bbb57e9c96
line, separated by ";".
* Fixed incorrect parsing when "|" was not separated by a space from the
previous argument.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35895 a95241bf-73f2-0310-859d-f6bbb57e9c96
the default setting for the "debug syslog" feature (can still be overridden
in the boot loader). Per default enabled for kdebug level >= 1.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35884 a95241bf-73f2-0310-859d-f6bbb57e9c96
for people who cannot capture serial output easier.
* The syslog sender thread waits with timeout now, so output added to the
syslog buffer without explicit notification is still written after a few
seconds at the latest.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35883 a95241bf-73f2-0310-859d-f6bbb57e9c96
support file creation.
* Extended open() and open_from() to support O_CREAT to create files.
open_from() has got an optional "permissions" parameter for that purpose.
* Fixed errno. It would crash when being used. Also changed the POSIX functions
to return their error code via errno as expected.
* Added writev().
* FAT file system:
- Added support for reading long file names.
- Added support for creating files (8.3 name only) and writing to them.
- Enabled scanning partitions with it.
* Boot loader menu:
- Enabled the "Reboot" menu item unconditionally.
- Added "Save syslog from previous session" menu item to the debug menu.
Currently saving the syslog to FAT32 volumes is supported.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35882 a95241bf-73f2-0310-859d-f6bbb57e9c96
received, but if the signal was in the thread's signal block mask, it would
not be handled. Added thread::sig_temp_enabled, an additional mask of not
blocked signals, which is set by sigsuspend() and evaluated and reset by
handle_signals(). Fixes#5567.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35836 a95241bf-73f2-0310-859d-f6bbb57e9c96
contribute significantly to the usable memory, but possibly make the MTRR
setup impossible. Might improve #5550.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35825 a95241bf-73f2-0310-859d-f6bbb57e9c96
machine the previous location was overwritten, probably by GRUB.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35817 a95241bf-73f2-0310-859d-f6bbb57e9c96
* The kernel syslog ring buffer is no longer emptied by the syslog sender
thread. Instead we only drop the oldest data from the buffer when we're
writing to it and there's not enough free space in it.
Advantages: We drop old data rather than the most recent data when the buffer
is full. The "syslog" KDL command has more data available now. So the odds
are that kernel syslog messages not written to disk yet are at least still
in the kernel buffer.
* Changed dprintf_no_syslog() semantics: Now it writes to the syslog, but
doesn't notify the syslog sender thread.
boot loader:
* Added the ring_buffer implementation and a dummy user_memcpy().
* bios_x86: Moved the syslog stuff from serial.{cpp,h} to debug.{cpp.h}.
* Moved the debug options from the "Select safe mode options" menu to a new
"Select debug options" menu.
* Added option "Enable debug syslog" to the new menu (ATM available on x86
only). It allocates a 1 MB in-memory buffer for the syslog for this session
in such a way that it can be accessed by the boot loader after a reset.
* Added item "Display syslog from previous session" to the new menu, doing
what its name suggests.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35816 a95241bf-73f2-0310-859d-f6bbb57e9c96
a given flat buffer.
* Added ring_buffer_peek() for random position reading from the ring buffer
without changing its state.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35813 a95241bf-73f2-0310-859d-f6bbb57e9c96
gKernelArgs.physical_allocated_range array. This way it is always up to date
and we don't need to fix it in mmu_init_for_kernel().
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35802 a95241bf-73f2-0310-859d-f6bbb57e9c96
in the unused list if it got the block via block_cache::_GetUnusedBlock().
* block_cache::_GetUnusedBlock() leaked the compare data block if the
BLOCK_CACHE_DEBUG_CHANGED feature had been enabled.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35799 a95241bf-73f2-0310-859d-f6bbb57e9c96
2 size we might need to use two registers (for the start and the end of the
range).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35753 a95241bf-73f2-0310-859d-f6bbb57e9c96
the maximum magazine count wasn't reached yet. With object_depot_store() not
resetting its local variable, a magazine could thus be emptied and freed
twice. Fixes#5489 and #5497.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35752 a95241bf-73f2-0310-859d-f6bbb57e9c96
* More output.
* Actually increment errorCount when encountering an error. The loop
condition would never be false this way.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35750 a95241bf-73f2-0310-859d-f6bbb57e9c96
from inserting a clean page, if a fault happens. VMNullCaches are used by the
slab's memory manager -- all page faults in slab areas are serious bugs and
we want to panic() immediately.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35748 a95241bf-73f2-0310-859d-f6bbb57e9c96
providing incorrect memory types. It doesn't cost us anything, since the
kernel reserves all but the unusable range later anyway. Should fix#1925.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35736 a95241bf-73f2-0310-859d-f6bbb57e9c96
(the time of the last access), as what we really want is a frequency/last
access time scoring, and accessed alone is useless for that.
* put_cached_block() no longer frees any unused blocks.
* The low memory handler will now only lock the cache if there is something
to do. Also, it did not take address space warnings into account.
* Even when memory is critical, we don't free all unused blocks anymore - if
the number of blocks we free now (10000) is not sufficient to get out of the
critical condition, chances are good that we will be called again :-)
* block_notifier_and_writer() now tries to make sure that the total block cache
memory consumption grows not much larger than half of the available RAM.
* This should all help to limit the block cache usage a bit better. Hopefully,
a checkfs run will no longer run out of memory here (couldn't test yet).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35733 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Made the page table allocation more flexible. Got rid of sMaxVirtualAddress
and added new virtual_end address to the architecture specific kernel args.
* Increased the virtual space we reserve for the kernel to 16 MB. That
should suffice for quite a while. The previous 2 MB were too tight when
building the kernel with debug info.
* mmu_init(): The way we were translating the BIOS' extended memory map to
our physical ranges arrays was broken. Small gaps between usable memory
ranges would be ignored and instead marked allocated. This worked fine for
the boot loader and during the early kernel initialization, but after the
VM has been fully set up it frees all physical ranges that have not been
claimed otherwise. So those ranges could be entered into the free pages
list and would be used later. This could possibly cause all kinds of weird
problems, probably including ACPI issues. Now we add only the actually
usable ranges to our list.
Kernel:
* vm_page_init(): The pages of the ranges between the usable physical memory
ranges are now marked PAGE_STATE_UNUSED, the allocated ranges
PAGE_STATE_WIRED.
* unmap_and_free_physical_pages(): Don't free pages marked as unused.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35726 a95241bf-73f2-0310-859d-f6bbb57e9c96
object_cache_alloc(): the ObjectCache::total_objects count was increased in
ObjectCache::InitSlab(), but the slab was really only added at a later point
between the cache could be unlocked.
* If a second object_cache_reserve_internal() managed to be called while the
lock was unlocked, it would see that there has to be space available, and
will then return -- however, since the other thread could not yet place the
slab into the cache, object_cache_alloc() cannot find it.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35702 a95241bf-73f2-0310-859d-f6bbb57e9c96
remove the block from an unlocked cache and the cache->lock would be unlocked
twice when the calling function unlocked. It panics in that case anyway, but
this should make it continuable and is more correct.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35648 a95241bf-73f2-0310-859d-f6bbb57e9c96
it is fully initialized.
* arch_vm_translation_map_is_kernel_page_accessible(): Check whether
sPhysicalPageMapper has already been initialized.
If a panic() during or before the initialization of the physical page mapper
occurred, we no longer access a partially initialized object or a NULL pointer.
This should fix the triple fault part of #1925.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35644 a95241bf-73f2-0310-859d-f6bbb57e9c96
happen on syscalls or "int" instructions. The debug exception handler sets
the thread debug flags B_THREAD_DEBUG_STOP and
B_THREAD_DEBUG_NOTIFY_SINGLE_STEP (new) and lets the thread continue. Before
leaving the kernel the thread is stopped and a single-step notification is
sent. Fixes#3487.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35620 a95241bf-73f2-0310-859d-f6bbb57e9c96
notably:
- in hpet_clear_hardware_timer() we cleared the HPET_CONF_TIMER_INT_ENABLE
flag, and never set it back anymore (now we do, in hpet_set_hardware_timer())
- the incorrect interrupt configuration (although it shouldn't matter in legacy mode).
Also test the HPET counter before trying to use it.
Improved debug information.
HPET timers are still at priority 0, since I coulnd't yet test on other
machines (besides mine, QEMU and XenServer 5.5), and doen't work for
some reason with SMP enabled.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35611 a95241bf-73f2-0310-859d-f6bbb57e9c96
* On context switches we do now set the IA32_MSR_SYSENTER_ESP. This saves us
setting esp at the beginning of x86_sysenter(). More importantly when
entering it in single-step mode, this no longer causes a double fault
(cf. #3487). We still don't handle the resulting debug exception correctly,
so that we still get a (continuable) panic().
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35609 a95241bf-73f2-0310-859d-f6bbb57e9c96
over ownership of the object. Fixes double free introduced in r35605.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35608 a95241bf-73f2-0310-859d-f6bbb57e9c96
they were never freed unless the cache was destroyed (I just wondered why
my system would bury >1G in the magazines).
* Made the magazine capacity variable per cache, ie. for larger objects, it's
not a good idea to have 64*CPU buffers lying around in the worst case.
* Furthermore, the create_object_cache_etc()/object_depot_init() now have
arguments for the magazine capacity as well as the maximum number of full
unused magazines.
* By default, you might want to initialize both to zero, as then some hopefully
usable defaults are computed. Otherwise (the only current example is the
vm_page_mapping cache) you can just put in the values you'd want there.
The page mapping cache uses larger values, as its objects are usually
allocated and deleted in larger chunks.
* Beware, though, I couldn't test these changes yet as Qemu didn't like to run
today. I'll test these changes on another machine now.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35601 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Since the same block can be in up to two transactions, it's very well possible
that one shouldn't write all transactions in a single run.
* Forgot to pass on the iterator from BlockWriter::Add(transaction) to
Add(block).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35565 a95241bf-73f2-0310-859d-f6bbb57e9c96
lines later anyway, after the cache has been locked and it has been verified
that the page is still a candidate. Fixes#5432.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35539 a95241bf-73f2-0310-859d-f6bbb57e9c96
the message doesn't get truncated.
* No longer print the "PANIC:" prefix when entering the kernel debugger via
kernel_debugger().
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35535 a95241bf-73f2-0310-859d-f6bbb57e9c96
whether the pages where mapped. Was introduced already in r22731.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35532 a95241bf-73f2-0310-859d-f6bbb57e9c96
needs to be or'ed to the address specification), "uncached" is assumed.
* Set the memory type for the "BIOS" and "DMA" areas to write-back. Not sure, if
that's correct, but that's what was effectively used on my machines before.
* Changed x86_set_mtrrs() and the CPU module hook to also set the default memory
type.
* Rewrote the MTRR computation once more:
- Now we know all used memory ranges, so we are free to extend used ranges
into unused ones in order to simplify them for MTRR setup.
- Leverage the subtractive properties of uncached and write-through ranges to
simplify ranges of any other respectively write-back type.
- Set the default memory type to write-back, so we don't need MTRRs for the
RAM ranges.
- If a new range intersects with an existing one, we no longer just fail.
Instead we use the strictest requirements implied by the ranges. This fixes
#5383.
Overall the new algorithm should be sufficient with far less MTRRs than before
(on my desktop machine 4 are used at maximum, while 8 didn't quite suffice
before). A drawback of the current implementation is that it doesn't deal with
the case of running out of MTRRs at all, which might result in some ranges
having weaker caching/memory ordering properties than requested.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35515 a95241bf-73f2-0310-859d-f6bbb57e9c96
therefore does not unlock anymore while iterating over the transactions.
This gave other threads the opportunity to finish a transaction, causing
bug #5412. Also, the BlockWriter will now always close transactions on its
own, and you need to pass the transaction hash iterator to Add().
* Also, transactions that contain blocks that are currently written back will
be ignored by the block writer, as well as cache_sync_transaction(). This
fixes bug #5415.
* Improved error handling if BlockWriter fails to write back blocks. Most
notably, they are no longer left busy_writing, and the functions calling
it do proper error reporting (besides block_cache_discard() that does not
return any erro code; I've added a TODO note there for now).
* The BlockWriter now starts with a larger array once it has to allocate one.
* One can now limit the number of blocks that go into a BlockWriter. This is
used by the block writer thread, that shouldn't always write back everything
every two seconds.
* Also, the fixed array is larger now (leaving enough space such that the
block writer/notifier does not need to allocate anything).
* And finally, if allocating the array fails, the BlockWriter falls back to the
synchronous write back used previously. IOW it will never write back less
blocks than you ask for.
* Added static BlockWriter::WriteBlock() method replacing write_cached_block().
* Forgot to rename block_cache::busy_count to busy_reading_count.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35510 a95241bf-73f2-0310-859d-f6bbb57e9c96
arch_timer_set_hardware_timer(). This was harmless, at least with our
current x86 timers implementation, since they checked for minimum timeouts.
Very small cleanup (now that the file is compiled as C++).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35505 a95241bf-73f2-0310-859d-f6bbb57e9c96
particular it wouldn't set the flag when writing something to a page, but
only move it to the modified queue. Since mapping the page would move it to
another queue, the information that the page was modified would be lost and
it would never be written to disk. Was well reproducible with a Haiku image
build and limited amount of memory.
Fixes the hopefully last remaining cause for #5374.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35500 a95241bf-73f2-0310-859d-f6bbb57e9c96
* full_scan_inactive_pages(): Fixed syntactical glitch (missing "else").
Affected only the active paging mode and was relatively harmless. The worst
case would be that an inactive page would be moved to the cached queue
although its usage_count hadn't dropped to 0 yet, thus freeing it before
pages that deserved it more.
* move_page_to_active_or_inactive_queue():
- Was ignoring the page's modified flag, thus potentially moving a modified
page to the cached queue. That could happen only in rare cases though,
like when the page was still mmap()ped while being written and modified and
unmapped before being done.
- No longer move the page to the inactive queue, even if its usage count is 0.
In idle mode the page daemon doesn't look at inactive pages, so the page's
stats wouldn't be updated anymore.
- Renamed to move_page_to_appropriate_queue().
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35499 a95241bf-73f2-0310-859d-f6bbb57e9c96
inactive pages to the active queue. This has the advantage that the page
daemon will keep track of those pages even in idle mode (where it only
processes the active queue).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35498 a95241bf-73f2-0310-859d-f6bbb57e9c96
block_cache_discard() use the BlockWriter directly as well.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35496 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Added new KDL command "cached_block" that dumps a cached block.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35493 a95241bf-73f2-0310-859d-f6bbb57e9c96
to make the guard page inaccessible. Thanks Ingo for the pointer!
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35488 a95241bf-73f2-0310-859d-f6bbb57e9c96
* vm_page_write_modified_page_range(): Need to DEBUG_PAGE_ACCESS_END() a bit
later.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35487 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Use the same criterion when to write back temporary pages as the page daemon.
* Move wired and temporary pages that shall not be written back to the active or
inactive queue so they don't get stuck in the modified queue and potentially
cause the page writer to run permanently without actually making progress
(#5382).
page writing implementation:
* If writing back a temporary page failed, move it back to the inactive or
active queue, so it doesn't get stuck in the modified queue. If active paging
continues, it might find its way back to the modified queue in the next
iteration, but that improves the situation a bit at least. Particularly with
the port heap pages not really being swappable ATM.
* Never dequeue pages from the modified queue. We mark them busy, so the page
writer will skip them anyway. This allows others to play with the page to some
extend (at least allowing to move it between the queues). This fixes#5404.
* Removed PageWriteWrapper::ClearModifiedFlag(). We clear the modified flag in
PageWriteWrapper::SetTo(), now, so that the page writer doesn't need to do
that explicitly either.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35485 a95241bf-73f2-0310-859d-f6bbb57e9c96
the heap debug panics. Instead syslog output is generated if turned off.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35481 a95241bf-73f2-0310-859d-f6bbb57e9c96
well as the thread allocating it. Can for example be used to verify that an
object or buffer is as large as expected.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35480 a95241bf-73f2-0310-859d-f6bbb57e9c96
keeping all returned heap memory in the 0xdeadbeef state (including the
first sizeof(void *) bytes otherwise for the free list). While wasting a lot
of memory it allows you to rely on 0xdeadbeef being always present as no
future allocation will reuse the freed memory block.
* Also added heap_debug_malloc_with_guard_page() which is intended to allocate
a memory block so it is aligned that the start of invalid memory past the
allocation is in an unmapped guard page. However the kernel backend that would
guarantee this is not yet implemented, so right now this works only by chance
if no other area happens to be allocated exactly past the created one. With a
very specifc suspicion you can put that one allocation you get to good use
though. It causes a crash when accessing memory past the allocation size so
you actually get a backtrace from where the access happened instead of only
after freeing/wall checking.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35478 a95241bf-73f2-0310-859d-f6bbb57e9c96
it has been unmapped. This way modified pages could end up in the "cached"
queue without having been written back. That would be a good explanation for
#5374 (partially wrong file contents) -- as soon as such a page was freed,
the invalid on-disk contents would become visible.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35477 a95241bf-73f2-0310-859d-f6bbb57e9c96
