We don't need to create a BNode from a BDirectory because BDirectory derives
from BNode! So use the BDirectory we already have as a BNode and get the
node_ref from it so we can pass it into WatchNode().
Sorry for the noise.
When an ARMv7 CPU is detected, immediately turn on the FPU. This allows
us to use vsnprintf in the TRACE call in that function, as our libc is
compiled with floating point support and will trigger a fault if the FPU
is not available.
This lets the boot go further, and crash in mmu_init. Next steps:
* Find why mmu_init is crashing
* Setup some fault handlers, otherwise we call uboot ones, and they are
not very helpful. They will also probably not work once the mmu is
enabledvery helpful. They will also probably not work once the mmu is
enabledvery helpful. They will also probably not work once the mmu is
enabled...
Loading of haiku_loader from an uImage is a 2-step process:
* First, the uImage is loaded (in our case from SD card using fatload)
to RAM at a temporary address.
* Then (using bootm), it is unpacked. The uImage is a container format
and can hold several files, with a load and execution address. The files
are copied from the uImage to their final location, and it's better if
that doesn't overlap with the uImage content
When this loading is done, bootm jumps to the entry point found in the
uImage.
We now actually execute our code from haiku_loader. This crashes with
the following call stack:
* vsnprintf
* dprintf
* boot_arch_cpu_init
* cpu_init
It seems vsnprintf is trying to use VFP instructions (probably from the
libgcc) but that triggers some kind of fault, and the handler (setup by
uboot?) ends up crashing the system by jumping to unmapped memory at 0.
* Cleanup the SD card image building to allow jam -q @bootstrap-mmc to
work.
There are a few remaining tricks before you can safely build an image:
* This uses a non-POSIX du option, and is only tested with Linux du
only (Linux is the only supported system to run bootstrap builds,
anyway)
* The Python recipe in haikuports.cross is known to not build on
Debian/Ubuntu, but work fine on OpenSuse. There is a patch available in
haikuports bugtracker to allow the reverse.
* You need to populate the haikuports repo package list with some
packages (which don't exist yet) to make the build system happy. But our
git hook to generate the repositories is preventnig me to share this
hack.
Once built, the image currently crashes early in the kernel execution.
On to debug that!
* Previously, BPath::Append() would return an empty path if the
path didn't exist (and an error); in a nightly image, the
B_USER_ADDONS_DIRECTORY didn't exist, triggering the bug.
* Replaced the code with calls to BPathFinder instead, which takes
care of the empty path automatically.
You can't increment the list item counter as you delete because the count
decreases as you go. Instead delete the first item until there are no more
items, then delete the list.
The logic was reversed accidentally in hrev47355
Rewrite the method to make more sense. If data is NULL return
B_BAD_VALUE right away. Otherwise set the status based on the first
operation, if that succeeds perform a second one, and return the
result.
Fixes CID 1222666
No functional change intended.
Focused on documented classes only.
* Update copyright information.
* whitespace fixes.
* pointer style
* Rename some variables, msg => message, form => what
* Need consistent variable names to make documentation easier,
allows us to use \copydoc or \copydetails instead of repeating
ourselves over and over again.
* Each BHttpAuthentication object is locked on all field accesses,
* They are owned by the BUrlContext and never deleted, so there is no
need for reference-counting them,
* The BUrlContext itself is now reference counted, and all BUrlRequests
hold a reference to it.
This makes sure using the BHttpAuthentication objects from requests is
thread-safe.
* Change the semantics of the iterators copy constructor and assignment
operator: they now return a new iterator for the same cookie jar (and
same url for the UrlIterator). They don't try to point to the same
position as the copied iterator. The only purpose of these is to write
code such as:
Iterator it = jar.GetIterator();
so having a full copy isn't that useful.
* The per-domain cookie lists are now protected with a read-write lock.
The iterators retain a read lock while they are handling cookies from
that list. They get a write lock when doing Remove. Adding a cookie to
the jar also gets the write lock for the matching list
* Fix a memory leak when adding a new domain-list to the jar failed
* Simplify the declaration of the PrivateHashMap type (it would be
even simpler if HashMap was a public API)
* The domain hashmap is now a SynchronizedHashMap. It is locked as long
as an Iterator or UrlIterator exists, which may be a problem as these
are public APIs. Writing safe iterators for an hashmap with concurrent
accesses is not easy, so the API could be modified to return a list of
domains and a list of cookies for a given domain or URL instead. This
would suit the intended uses just as well.
* The jar now store const cookies, so there is no need to lock them for
access/modification. Updating a cookie is done by replacing it with
another one in the jar (with the same domain and value). There is still
the problem of deleting a cookie while other threads may still access
it, this will be fixed by making cookies BReferenceable.
* Word 105 of ATA IDENTIFY should contain the maximum number of block
ranges the device supports, but it always supports at least a single
one.
* This change lets the trim command do something on Virtual Box.
* This tests requires some private APIs which have been changed, it
needs to be rewritten.
* I'm also not sure if it's still possible to work on a copy of the MIME
* database to avoid messing with the system one.
If you have the Desktop window open when you switch to spatial
mode this will now close the window. You aren't allowed to have
the Desktop open in 2 places symultaneously in spatial mode.
... in a few more cases.
If you are in spatial mode and you double-click on the Desktop folder
it activates the Desktop putting all other Tracker windows in the
background.
If you are in browser mode and you double-click on the Desktop folder
it opens the Desktop folder in the current window.
If you issue 'open ~/Desktop' from Terminal or if you have Expander set
to open the expanded folder and you expand something on the Desktop
it won't open the Desktop folder in its own window in either spatial or
browser mode.
This patch alters Trackers behavior to open the Desktop folder in its own
window in these cases when you are in browser mode.
One might argue (as a few have on IRC) that this behavior should also
happen in spatial mode, but, this patch does not alter the behavior of
spatial mode. In spatial mode opening the Desktop folder with
'open ~/Desktop' or extracting a zip file with Expander on the Desktop
simply activates the Desktop pushing all other Tracker windows into the
background state.
Fixes#10929
* thanks to Ingo for suggesting the idea, quoting him:
"by holding sVnodeLock read-locked, get_mount() ensures that fs_unmount() can't
process the nodes. If it is already past that point, the root node check
(not NULL, not busy, ref count > 0) is supposed to detect that. But it doesn't
look like this can work. fs_unmount() doesn't set the root node to NULL (the
root node field is NULL only during a short period in fs_mount()), but it just
frees the nodes after releasing sVnodeLock. So the not busy and ref count > 0
checks could already access freed memory".
* tested OK, this fixes#10522.
* replaced mount->root_vnode by the local variable with the same value.
* There is a SimpleTest in AlertTest.cpp, and an UnitTest in
balert/AlertTest.cpp. They ended up with the same grist as they are
built from the same jamfile. Rename one of them to avoid this.
* The Udf namespace was removed from the headers, but not all the
implementation files
* kernel_cpp.h is not actually used
* TRACE_ERROR can't be set to dprintf when building as part of the test
(and not as a kernel add-on)
First step on the way to get jam -q unittests to work again...
* Cookies sometimes come with the UTC timezone, or no tz info at all.
* No other timezone seems to be used
This allows better matching of cookies which would otherwise be kept
only as session cookies.
... try 2 without a spurious \
sorry for the noise.
From the commits list (Ingo) <<-EOF
> Terminal crashes because fView is not connected to App Server when
> this is called so calling fView->GetMouse() is not allowed.
The interesting questions therefore are:
1) Why is the view not attached?
2) Why is it in the hyper link state?
1) is because the BTabView removes the non-selected tabs instead of just
hiding them.
The reason for 2) is that the tab was opened with Cmd+T (the bug is not
reproducible when the tab is opened via menu item). Pressing Cmd causes
hyper link state to be entered and switching to the new tab will leave
the view in that state due to 1).
Possible solutions:
* TermView::_VisibleTextBufferChanged(): Call state hook only when attached
to window. All other occurrences are safe as they are in BView hooks.
* Leave the hyper link state when the view is detached from the window. A
new dummy state could be active as long as the view is not attached, though
using DefaultState would be harmless as well.
EOF
I chose solution 2 using DefaultState rather than creating a new dummy state.
Thanks Ingo.
Fixes#10902
From the commits list (Ingo) <<-EOF
> Terminal crashes because fView is not connected to App Server when
> this is called so calling fView->GetMouse() is not allowed.
The interesting questions therefore are:
1) Why is the view not attached?
2) Why is it in the hyper link state?
1) is because the BTabView removes the non-selected tabs instead of just
hiding them.
The reason for 2) is that the tab was opened with Cmd+T (the bug is not
reproducible when the tab is opened via menu item). Pressing Cmd causes
hyper link state to be entered and switching to the new tab will leave
the view in that state due to 1).
Possible solutions:
* TermView::_VisibleTextBufferChanged(): Call state hook only when attached
to window. All other occurrences are safe as they are in BView hooks.
* Leave the hyper link state when the view is detached from the window. A
new dummy state could be active as long as the view is not attached, though
using DefaultState would be harmless as well.
EOF
I chose solution 2 using DefaultState rather than creating a new dummy state.
Thanks Ingo.
Fixes#10902
Also derived from Opera test suite.
* Fix nbsp instead of space in some tests (I shouldn't have copypasted
these...)
* Add most of the advanced tests from Opera. This exposed several bugs
in the cookie code (all fixed) as well as strptime and strtol (tickets
opened)
* All the cookie2 tests are skipped. I didn't find any use of it on the
internets, so it doesn't seem useful to implement it.
* The last two tests sets are still missing. I'll add them next week.
* An empty "expires" field results in a session cookie, rather than
rejecting the cookie altogether
* A page can set a cookie it is not allowed to access (for example in a
subdirectory of where the page is located). Separate IsValidForUrl and
_CanBeSetFromUrl to perform the appropriate checks in each case.
* Limit cookie path to 4096 characters. As a result of the previous
change, a page would be allowed to set a cookie with an aribrarily long
subpath, wasting disk space and RAM by growing hte cookie jar.
* Don't allow path with . or .. elements. These are a source of
confusion and are not needed.
* Reset the cookie fields when parsing failed. This does not matter when
using the cookie jar, but is useful when working directly with
BNetworkCookie.
strptime can return non-NULL values even if it only parsed part of the
string. This was sometimes making us use the wrong format. Now we try
all formats and checks how much of the string strptime managed to parse.
We stop when it has parsed a big enough part of it.
Terminal crashes because fView is not connected to App Server when
this is called so calling fView->GetMouse() is not allowed. We could
perhaps check to make sure that fView is connected to App Server by
checking that fView->Window() != NULL, but, a better way to handle this
is to grab the already set fLastClickPoint variable from fView
which eliminates the need to grab the current mouse position again.
Worst case scenario is that fLastClickPoint hasn't been set in which
case Terminal won't find anything to highlight or unhighlight.
Also check to make sure that fView != NULL here for good measure.
Fixes#10902
* BString::CopyInfo() takes length as second
parameter, not the end-offset. This bug didn't
have any effect, since BString clamps the length.
* ',' is a comma, ':' is a colon.
* When parsing "additionalData" in the loop for
name=value pairs, an empty name is not useful,
continue the loop early.
* "value" may have the length 0, accessing it
with value[0] would not lead to an access violation,
since it just reads the terminating zero (I think),
but it's nicer if the code makes it clear that
value being empty is considered.
* _H() was using a static buffer, in a heavily
multi-threaded situation. I don't think this was
healthy.
* Implement the proposed optimization of using
BString::LockBuffer(). Appending one char at a time
is really bad for peformance. The Base64 encoding/decoding
should really be rewritten as well for similar reasons.
These were getting out of sync and causing trouble, and they are easy to
compute from existing information.
Fixes some problems detected by the testsuite where the user/password or
the host would sometime disappear from the URL.
No functional changes intended.
* Updated copyright information.
* Reduced doxygen documentation down to a helpful summary
in a regular comment, the documentation has been moved into
the Haiku Book.
* Some parameter renaming for consistency and clarity.
* A few other style fixes.
This shows there are some problems with BUrl, still.
CppUnit stops at the first failed assert in each test, which doesn't
give a good overview of the number of failures in this case (the first
subtest fails). A possible solution is splitting this into smaller
subsets.
The way we handle paging is very wasteful and relies heavily on virtual
funcions even if there is absolutely no reason to do so. The proper
solution would be to do a major rework of paging code (including
arch-independent parts).
On x86_64 physical page mapper is very simple what makes the overhead
resulting from the desing of paging interface very expensive. This
patch attempts to make things a bit better by helping GCC with
devirtualization and allowing inlining physical page mapper impementation
(well, only when it is devirtualized).
This fixes the problem with find_unique_check_sums() taking a very long
time to complete when one or more drives report read errors.
Fixes#10880.
[Paweł: minor style issue]
Signed-off-by: Paweł Dziepak <pdziepak@quarnos.org>
* A problem with our gcc requires adding casts for gcc4 when
the __builtin_bswap functions are used with a format string
* Unlike gcc2, the __builtin_bswap functions do not get disabled
despite using -fno-builtins, hence added compiler check in
runtime_loader/utility.cpp
It enables just a few printf()s in PackageItem.cpp.
I went with a quick solution using a global variable
instead of using/writing some nice logging framework which
could even generate install logs... but it helps in tracking
down some package problems without first having to uncomment
some printf()s...
... and perhaps more packages build for pre-PM Haiku.
* Restore the original path rewriting of "/boot/system"
alongside rewriting "/boot/beos/system".
* Rewrite "/boot/apps" in install scripts. This one fits
the existing TODO that the rewriting should not use
hard-coded replacement paths.
* Fixes ticket #10875
* These aren't tested, but since we go off of DCE
versions for a lot of stuff, they may work.
* AMD doens't include market names in their drivers
anymore, so if we want to label them it will take
additional work.
* pptable.h now 100% stock
* atombios.h now 99% stock
* Resolves issues with radeon_hd on x86_64 by
making atombios structs match OS typedefs
* Resolves#10614
...when the transformation moves the text into the view bounds. The
rect returned from Painter::BoundingBox() is already transformed, as
should be expected.
Since the bookmark bar can now be resized smaller than the space needed
for the items, move items that don't fit in it to a menu, which is
always the last item in the BMenuBar.
* Add test for a translation that makes things that are out of view
bounds to become visible
* Fix the "rectangles" test to handle scrolling of the view. It seems
that in some cases some of the rectangles are not drawn.
The gradient and bitmap tests currently have problem with view
scrolling, but I'm not sure why.
The DrawingEngine didn't properly make a distinction between the
rectangle being filled and the damaged region on screen. This led to
unexpected results when using BAffineTransform.
- Destroy the background worker before going after all the managers. Otherwise,
there might potentially still be jobs in flight that might touch them, leading
to crashes.
The ScreenSaver window is now a BDirectWindow, but, _DrawFractals() makes
the assumption that a direct window means that the screen saver is running full
screen, so, we need to make sure we draw the preview using the
DrawBitmapAsync() method.
Fixes#10832
* data files are still in the source tree.
* gutenprint headers contain a image.h header file which collides
with ours. This is solved by forcing include search first on
os/kernel directory.
Due to introduction of user-defined suffixes C++11 requires that there
is a space between literal and identifier to avoid ambiguity.
This patch makes the whole kernel build successfully with C++11.
This patch makes it possible to inline rdmsr and wrmsr instruction. The
performance impact shouldn't be significant since they are used relatively
rarely and wrmsr is usually a serializing instruction, but there is no reason
not to do so.
The goal of this patch is to amortize the cost of context switch by making
the compiler aware that context switch clobbers all registers. Because all
register need to be saved anyway there is no additional cost of using
callee saved register in the function that does the context switch.
Similarly to previous patch regarding GDT this is mostly a rewrite of
IDT handling code from C to C++. Thanks to constexpr IDT is now entirely
generated at compile-time.
Virtually no functional change, just rewriting the code from
"C in *.cpp files" to C++. Use of constexpr may be advantageous but
that code is not performance critical anyway.
