Previously the CCB would never complete when a timeout occured, leading
to all further IO stopping.
Abort the request by setting the CCB status to aborted and handing the
CCB back to the SCSI layer. That one will then handle the error (and
possibly retrying). This is similar to how such errors are handled in
AHCI and the ATA stack.
Since it is now possible that we receive a completion interrupt for an
already aborted request, we need to keep track of what request the
interrupts belong to and only notify when the current one completes. As
there currently can't be multiple requests in flight, a simple counter
is used.
Change-Id: Ib80e146605efd2f81123803f424cc7f66f52a6c8
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1815
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
It previously returned the cookie directly, which made it impossible
to distinguish between a NULL cookie and the function not having
anything to dequeue. This lead to some code setting a cookie that was
not actually used.
Return the dequeue status as a boolean and provide the cookie with an
optionally handed in pointer instead and adjust all users.
Change-Id: Iaac1726ac4bc7ae42bb96b8f0915852b6def5822
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1814
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
As soon as the lock is released, fCCB may change due to a new Start().
With the current use of the request class this could not happen, so
this is purely theoretical.
Change-Id: I6caee7f904f1864621aeef088e2bd611eb3b0a1f
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1813
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
The value is never actually checked. The use of a condition variable
makes it redundant since it will already either have a waiting entry,
or not.
Change-Id: Iafaecb7f9e56a2cf8932e05b82aad2b350fbce1f
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1812
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
This change will allow the user to view the user
usage conditions from the HDS system in a
HaikuDepot window. The display of the text is not
currently well formatted in that the Markdown is
not yet properly parsed, but the display of the
data is working.
Relates to 15209
Change-Id: Ia6ad4ef995f5fe3c29c40221964e44d4554a033d
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1750
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
When dispatching a message to a BView, pass the message on to the view's
MessageReceived() method rather than invoking one of its handlers (e.g.
KeyDown() or MouseDown()) directly.
This commit moves the existing handler-invoking logic from
BWindow::DispatchMessage() into BView::MessageReceived(), preserving the
existing functionality.
Fixes#15254.
Change-Id: I230c0781df13e54e08573a5a31d7550520c060d5
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1723
Reviewed-by: Stephan Aßmus <superstippi@gmx.de>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Reviewed-by: Ryan Leavengood <leavengood@gmail.com>
So many Jamfiles to search through...runs also, but there
are lots of graphical glitches
Change-Id: Ibf9e64566a5b8c5742792ac9b1b0f9ccc6693c8d
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1753
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
This makes ARM64 target compile more files. This patch is one of
series of patches to support new architecture, as fixes in many
places are required just to compile the code.
Signed-off-by: Jaroslaw Pelczar <jarek@jpelczar.com>
Change-Id: Ia060612733cd3a0fcb781fec449da164ed635b8e
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1807
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Now it is possible to run applications, do Git checkouts, etc.
on a ramfs (and those seem to work just fine -- a git checkout
followed by a git fsck both succeeded.)
This adds one (private) VFS function, and checks in all usages of
the vnode->cache as a VMVnodeCache that it really is one. (Generic
usages, for the moment just the ReleaseRef() calls in vnode
destruction, are intentionally not touched.)
This will be used by ramfs to set the cache from its own,
so that map_file() calls on a ramfs can work.
This only works because DataCollector stores its data in blocks
which are mapped into the kernel's address space. After the
next series of commits, it won't, so we can't depend on that.
This required some changes to the indexes to keep a copy of the
keys.
Now that Attributes don't use a table, we can replace the generic
system with a specific DirectoryEntryTable, upgrading to BOpenHashTable
in the process.
This is what packagefs does; so if it's performant enough for packagefs,
it's performant enough for us. Greatly simplifies this code (and will
allow for further simplification of the NodeChildHash.)
The "exec" tool can only handle one command with environs set at
the beginning of the line, so now we set the ADD_BUILD_COMPAT...
in this format. This also seems to be a general performance
improvement to builds using real shells, too.
Change-Id: If4b3117651b5475039d5e8116cd3de398582290a
Refactored out of execvpe. Originally I did this for my attempted
change to posix_spawn, but that change turned out to be wrong and
actually not that beneficial. This bit seems potentially useful,
though, so here it is.
- It appears that, in this case, port->open_fd is NULL.
Maybe, this would require some investigation too.
But in this case, i think it is better to remove port
from the device list anyway because remove_port will freed
the port at the end of the loop.
And adding 0xdeadbeef in the device list is not a good idea.
Change-Id: Id6b55f5e19ebbbbb48f0fa0b7d9ac65b996dbaf2
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1746
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Having valid superblocks from multiple filesystems on the same disk is
confusing, let's try to avoid that.
Change-Id: I1a58843d45ea52193a77faaf8dcc3ca6a049406e
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1802
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
If multipole filesystems think they can handle a volume, it makes sense
to assume the native one is to be used.
Makes my partitions with a BFS superblock and a leftover ext2 superblock
on them mountable again without manually specifying the filesystem type.
Fixes#15045.
Change-Id: Icb694472aa3de0a79e5f7d3c54976038e1e2590a
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1801
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
