The old 'Io_response_handler::io_progress_response' interface has been
replaced by the 'Vfs::Env::User::wakeup_vfs_user' (issue #4697). The
remaining 'read_ready_response' method is now hosted in the
appropriately named 'Read_ready_response_handler'.
Issue #4706
The 'blocked_handles' queue was used to notify the VFS user via the
'io_progress_response' mechanism. This is now covered by the
'wakeup_vfs_user' interface introduced in issue #4697.
Issue #4706
By adding a 'write_ready' interface following the lines of the existing
'read_ready', VFS plugins become able to propagate the (de-)saturation
of I/O buffers to the VFS user. This information is important when using
a non-blocking file descriptor for writing into a TCP socket. Once the
application observes EAGAIN, it expects a subsequent 'select' call to
return as soon as new I/O buffer space becomes available.
Before this patch, the select call would always return under this
condition, causing an unnecessarily busy write loop.
Issue #4697
The new interface is meant to replace the 'Vfs::Io_response_handler'.
In contrast to the 'Io_response_handler', which had to be called
on a 'Vfs_handle', the new interface does not require any specific
'Vfs_handle'. It is merely meant to prompt the VFS user (like the libc)
to re-attempt stalled I/O operations but it does not provide any
immediate hint, about which of the handles have become ready for
reading/writing.
Issue #4697
This patch removes the 'Insufficient_buffer' exception by returning the
WRITE_ERR_WOULD_BLOCK result value instead. It also eliminates the
superfluous WRITE_ERR_AGAIN and WRITE_ERR_INTERRUPT codes.
Issue #4697
The original size of 16K impedes the batched processing of network
packets. Changing the value to 256K reduces the number of context
switches when downloading large files and thereby improves the
throughput by more than 25% (base-hw on qemu_x86_64, using fetchurl
to download a file of 100 MiB via the NIC router from lighttpd).
Issue #4697
This patch fosters the batching of network packets transferred by the
lwIP stack over the NIC connection. It replaces the eager submission of
the packet-stream's data-flow signals by explicit wakeup notifications.
The commit also increases the NIC session's buffer size from 128 to 1024
packets.
Issue #4697
Since the lwip library is used only by the vfs_lwip plugin, this patch
moves the Genode-specific support code local to the VFS plugin, easing
further simplification and allowing for a tighter integration with the
VFS plugin.
The recipes for api/lwip and src/vfs_lwip stay separate as both parts
use different licenses.
Issue #4697
This patch facilitates the batching of I/O operations in the VFS library
by replacing the implicit wakeup of remote peer (via the traditional
packet-stream interface like 'submit_packet') by explicit wakeup
signalling.
The wakeup signalling is triggered not before the VFS user settles down.
E.g., for libc-based applications, this is the case if the libc goes
idle, waiting for external I/O.
In the case of a busy writer to a non-blocking file descriptor or socket
(e.g., lighttpd), the remote peers are woken up once a write operation
yields an out-count of 0.
The deferring of wakeup signals is accommodated by the new 'Remote_io'
mechanism (vfs/remote_io.h) that is designated to be used by all VFS
plugins that interact with asynchronous Genode services for I/O.
Issue #4697
This test reveals the patters of the batching of consecutive write
operations on a file-system session. It issues 100 writes of one byte
each, which should ideally result in large batches of operations
submitted to the file-system session at once.
The run script performs the write operations through a chain of two VFS
servers, thereby exercising the write batching of both the libc and the
intermediate VFS server.
Issue #4697
There is at least one client that performs DRM calls from multiple
pthreads, which is not supported for now. We guard the concurrent
access by grabbing a pthread_mutex before each operation.
Although we do not have the full ACPI information parsed yet, to
announce non-PCI devices derived from the ACPI tables, the device
description of the assumed devices is now integral-part of pci_decode.
Formerly, the information was gained separatedly as boot-module, whereby
we lost synchronization in between ACPI/PCI parsing, BIOS handover, and
PS/2 emulation code already acting.
This recipe combines the dummy_rtc_drv, which consumes a given time,
and provides the "Rtc" service to its clients with the simple NTP client,
which provides the current time to the RTC driver in regular intervals.
Ref genodelabs/genode#4687
Test to trigger periodically ACPI suspend and resume and periodically
trying to restart graphic driver.
Tested successfully with X201 and T420.
Issue #4669
It looks like http://www.elm-chan.org is blocking certain IP ranges,
because downloading the source archive on certain machines fails with
"403 Forbidden".
Fixesgenodelabs/genode#4623
There is one binder with one memory zone per OpenGL context. The patch was
needed before because Genode didn't have proper context support (separate page
tables) leading to binders from one or more context being allocated to the same
GPU address with the same page tables. This was clearly an error and is fixed
with context support.
issue #4664
To prevent caching side-effects of USB DMA memory taken from the packet stream
all allocations of USB packets need to be on separated cachelines at least.
Fixgenodelabs/genode#4655
The tar.gz archive of version 2.4.9 vanished from sourceforge. As the
project moves to GitHub currently and the archive of 2.4.9 is still
available there, just update the download location.
Content and hashes did not change.
If `select_from_ports` is evaluated lazily, we might miss a port during
the dependency check. A way to prevent this is to use the `:=` operator.
Fixesgenodelabs/genode#4618
The configured 'st_blksize' influences the sizing of the FILE buffer,
see '__swhatbuf()'. Increasing the blksize hint can lead to larger
'fread()' requests and in return to fewer roundtrips.
With the increase of MAXPHYS, the rump kernel requests a contiguous
allocation of 2101248 bytes, which exceeds the allocator's block size of
2 MiB.
Error: backend allocator: Unable to allocate memory (size: 2101248 align: 12)
The patch avoids this corner case by increasing the allocator's block
size to 4 MiB.
Fixes#4613
The libc kernel used to silently go on if one of the files given through
stdin/stdout/stderr or <fd path=""> was missing (with possibly vital
functionality for the component not working). A pointer to the presumably
simple configuration issue of the underlying scenario was not given to the
user.
With this commit, the libc kernel prints a descriptive warning before
proceeding with the invalid file descriptor
Fixes#4218
The libc kernel used to end up in an uncaught exception if one of the
directories in the paths given through stdin/stdout/stderr or <fd path=""> was
missing. The resulting error didn't point the user to the presumably simple
configuration issue of his scenario in any way and killed the affected
component.
With this commit, the libc kernel prints a descriptive warning instead and
simply continues to run the component with the corresponding file descriptor
missing. At least in case of stdin/stdout/stderr, the libc deals gracefully
with this approach.
Ref #4218
The `with_sub_node` method is renamed to `with_optional_sub_node` to
better reflect that the non-existence of a sub node with the desired type is
ignored.
At the same time, the new `with_sub_node` now takes a second functor that is
called when no sub node of the desired type exists.
genodelabs/genode#4600
changes. The alternative settings get reread on all interfaces as done before
"qemu-usb: fix device endpoint update" commit by introducing specific
reset_alt_settings function. The fallthrough case was intentionally before
the "qemu-usb: fix device endpoint update" case, which now is wrong.
Issue #4596
Instead of having a generic "virt_qemu" board use "virt_qemu_<arch>" in
order to have a clean distinction between boards. Current supported
boards are "virt_qemu_arm_v7a", "virt_qemu_arm_v8a", and
"virt_qemu_riscv".
issue #4034
When creating a socket, fd flags can be specified by ORing them with the
socket type. Most importantly, the flag SOCK_NONBLOCK must be propagated
to the Socket_fs::Context in order to support non-blocking recv/send.
genodelabs/genode#4550
All access to PCI devices can be handled nowadays via the platform session.
The I/O port access seems also to work not properly on newer UEFI machines.
Fixes#4532
* Instead of repeatedly spam the log with errors and
bring heavy load to the USB host controller driver
abort yourself when the device vanishs
Fixgenodelabs/genode#4515
This patch changes the libc to query the stack size from the
config attribute <libc> <stack size=""/> </libc> for regular
components, not only cloned processes.
Remove handling of ACPI (SCI) interrupt from kernel and thereby let the
handling of the ACPI irq exclusively to an Genode driver. On Genode the ACPICA
library+app handles the ACPI irq, where we can now remove the contrib patches
related to the synchronization between NOVA kernel and ACPICA library.
Fixes#4479
When executed on Linux, the test was impaired by the copy-on-write
optimisation since the source buffer was never initialised. By default,
Linux only maps a zeroed page until the first write access to the page
occurs. Since the source buffer was never written, the corresponding
page was always present in the physically-indexed data cache. In
consequence, the test merely measured write performance (similar to memset).
genodelabs/genode#4454
Send usb ctrl transfers with one second timeout as some devices (e.g.,
smartcard readers) do not response to certain control transfers.
Thanks to Peter for the investigation.
Upgrade to the well known worst cases by the GPU multiplexer. Do not
keep track of resources locally, in case resources are exceeded the
remain so anyway.
issue #4451
First, the endpoint update has to comply with the current alternate
settings of all interface, which are stored in USBDevice::altsetting[]
(one value per interface). Second, a SET_INTERFACE control request via
Packet_type::ALT_SETTING must update USBDevice::altsetting for the
interface.
Now, USB devices with multi-setting interface like the Joulescope JS110
with mixed bulk/isochronous endpoints are supported.
The run script uses core, ld.lib.so, and init from depot packages, thus
these cannot be integrated in the image explicitly from the build
directory.
Also, removed special build of report_rom, which is not subject of the
test.
Fixes#4437
* Switch from the legacy usb_host driver to the new PC version
in recipes and automated tests
* Update documentation snippets
* Remove outdated, unused usb_rndis run-script
Fixgenodelabs/genode#4416
XML allows attribute values like <node attr="\"/>. The XML parser
wrongly reflects this case as 'Invalid_syntax'. This behavior stems from
the implicit use of the 'end_of_quote' function, which considers the
sequence of '\"' as a quoted '"' rather than the end of a quoted string.
The patch solves this problem by making the 'end_of_quote' part of
the tokenizer's scanner policy.
The patch removes the 'end_of_quote' function from 'util/string.h'
because it is not universal, and to avoid the ambiguity with
'SCANNER_POLICY::end_of_quote'.
Fixes#4431
To make room for the re-newed usb_host_drv basing on Linux 5.14 and
the re-newed lx_kit/lx_emul we have to move the depot recipe and
consistently name the old drivers with a legacy_ prefix.
Ref genodelabs/genode#4416
In the error case of socket_fs_accept() the Unconfirmed utility was
incompletely applied with the result of executing the cleanup routines
in the wrong order.
Fixes#4417
The archive contents are equivalent except the creation date of the
top-level directory.
-drwxr-xr-x guivol/users 0 2021-01-07 12:54 jpeg-9d/
+drwxr-xr-x guivol/users 0 2022-01-04 12:02 jpeg-9d/
Fixes#4406
The 'read' and 'write' utilities are from a time before the VFS API
as os/vfs.h was available. They rely on the (now removed) blocking
semantics of the packet-stream interface.
The only remaining legitimate use case of the direct interaction with
the file-system session without VFS is the back end of gcov, which needs
a way to exfiltrate the statistical data using a channel that is
independent from the libc or the VFS.
Issue #4390
Do not import a freshly allocated buffer to all contexts eagerly.
Instead check buffer list in context's 'exec_buffer' call and import
only buffers needed and not present before GPU execution. This leads to
improved performance for applications that use many OpenGL contexts
(e.g., VirtualBox 6).
issue #4380
Retrieve multiple GPU sessions from VFS plugin, take advantage of buffer
import/export functionallity in order to implement gem context support.
Multiple contexts share all GPU buffers, but use different GPU sessions
and thus, differnt page tables and hardware contexts.
issue #4380
Synchronize GPU completion by calling 'read' of the vfs_gpu plugin. This
enables pthreads to wait for GPU completions instead of the main EP.
issue #4380
* let iris handle buffer management, this implies that BOs are mapped to
the PPGTT during buffer execution and unmapped by iris later, for this
to work buffers need to be unmapped when allocating cached BOs
(vma_free) which requires a patch
* support lseek (drm_lseek for now) for determining object size
issue #4380
Only parse output of the test programs one-by-one and ignore log
messages from other components not starting with '[init -> test$number]'.
Fixes sporadic failures on KERNEL=sel4 due to the following warning from
core.
Warning: flush page table entries - mapping cache full - PD: init -> test1
The socket type (in the lower bits) maybe ORed with SOCK_CLOEXEC and
SOCK_NONBLOCK options (in the higher bits). Currently, supported values
are SOCK_STREAM (1) and SOCK_DGRAM (2), so just take the lower 2 bits.
This fixes treating `SOCK_STREAM` sockets as UDP if additional flags
were set.
Fixes#4370
The path for the `nameserver` file was fixed to `/socket/nameserver`.
So, if the socket directory was configured differing from `/socket`, DNS
did not work. Now the default path for the `nameserver` takes into
account the path configured in the `socket` libc config attribute.
Fixes#4318Fixes#4343
Since the parts of the 'etnaviv' library are already part of the
'mesa.lib.so' that is normally loaded along-side remove the
duplication here.
Issue #4329.
This patch changes the 'Allocator' interface to the use of 'Attempt'
return values instead of using exceptions for propagating errors.
To largely uphold compatibility with components using the original
exception-based interface - in particluar use cases where an 'Allocator'
is passed to the 'new' operator - the traditional 'alloc' is still
supported. But it existes merely as a wrapper around the new
'try_alloc'.
Issue #4324
This patch replaces the 'Ram_allocator::alloc' RPC function by a
'try_alloc' function, which reflects errors as 'Attempt' return value
instead of an exception.
Issue #4322
Issue #3612
Always instantiating a network device with id `net0`, removes the need to call
append_qemu_nic_args in run scripts unless we want to add forwarding
rules.
genodelabs/genode#4311
In the default setting the number of root ports is set to 4. This
also limits the number of USB host devices that may be attached to
the VM. Since the USB webcam shares the slots, that leaves us with
only 3 available USB2 slots to attach USB host devices. Depending on
the use-case that might not be enough.
This commit statically increases the number of ports to 8 each and
adapts the qemu-usb glue-code accordingly.
Many thanks to Raphael for initial investigation and workaround.
Fixes#4310.
Those symbols are not satisfied by any code or dependency of libm. As
result calling cprojf function will always crash on Genode. This crash
can be turned into link time error by adding --no-undefined to LD_OPT.
This patch provides the missing symbols by including isninf.c in libm
build.
Fixes#4299
Apparently the iris driver does not make use of tiling by the kernel, so
we shortcut the 'SET_TILING' call to keep iris happy with this quickfix.
However, tiling information may get lost, if the iris driver ever calls
'MMAP_GTT' and no fence is configured for the buffer. A follow-up commit
should address this shortcoming in the future.
Issue #4284
The new 'Env::try_session' method mirrors the existing 'Env::session'
without implicitly handling exceptions of the types 'Out_of_ram',
'Out_of_caps', 'Insufficient_ram_quota', and 'Insufficient_cap_quota'.
It enables runtime environments like init to reflect those exceptions to
their children instead of paying the costs of implicit session-quota
upgrades out of the own pocket.
By changing the 'Parent_service' to use 'try_session', this patch fixes
a resource-exhaustion problem of init in Sculpt OS that occurred when
the GPU multiplexer created a large batch of IO_MEM sessions, with each
session requiring a second attempt with the session quota upgraded by
4 KiB.
Issue #3767
This commit removes the report service from the usb_webcam pkg, which
was used to enable or disable the webcam driver. The on/off state is
instead controlled by the presence of a capture client. That is, once a
capture client appears, the webcam driver is started. Vice versa, once
no capture client exists, the webcam driver is removed automatically.
Internally, the detection of presence of a capture client is based on
nitpicker's 'displays' report, which is consumed as input of the
rom_filter, which in turn generates the configuration of a dynamic sub
init.
Fixes#4287
'Out_of_ram' was so far the only exception a client had to deal with
during buffer managment. Allocating memory, however, does not only
consume RAM quota but CAP quota as well.
This commit tries to mitigate that shortcoming by reflecting the
'Out_of_caps' state back to the client. Furthermore it allows for
resource accounting on certain client allocations, e.g. buffers.
Fixes#4284.
By using the cached_fs_rom instead of fs_rom, each client obtains the
same (read-only) copy of the shared library, avoiding a RAM resource
request when running mulitple GPU applications at the same time.
With this patch, Sculpt is able to host at least 3 Mesa applications.
Issue #4263
Issue #4281
This new version of the system_clock pkg does no longer depend on the
presence of an external 'Rtc' service as previously provided by the
Sculpt base system. Instead, it hosts the rtc_drv inside the subsystem.
Because rtc_drv is board-dependent, the system_clock pkg is named
system_clock-pc now.
Issue #4281
Some guests don't handle remote wake up correctly causing devices to
stop functioning. Therefore, we disable the remote wake up bit (5) in
`bmAttributes` of the device configuration descriptor.
Thanks to Peter for the initial fix.
Fixes#4278
The various mapping methods are modelled after the requirements of
the Intel GPUs or rather the Mesa driver back end.
With upcoming support for other driver back ends, we need to
sequeeze their requirements in as well. For now hijack 'map_buffer'
to provide for specifying the kind of attributes the client needs.
For now all buffers mapped in the GGTT for Intel GPUs are treated
as RW.
Issue #4265.
This call allows for checking if the given execution buffer has been
completed and complements the completion signal. Initially the GPU
multiplexer always sent such a signal when the currently scheduled
execution buffer has been completed. During enablement of the 'iris'
driver it became necessary to properly check of sequence number.
In case of the Intel GPU multiplexer the sequence numbers are
continous, which prompted the greater-than-or-equal check in the
DRM back end. By hidding this implementation detail behind the
interface, GPU drivers are free to deal with sequence numbers any
way they like and allows for polling in the client, where the
completion signal is now more of a progress signal.
Issue #4265.
The current info implementation (as RPC) is limited in a few ways:
* The amount of data that may be transferred is constrained by the
underlying base platform
* Most information never changes during run time but is copied
nonetheless
* The information differs depending on the used GPU device and
in its current implementation only contains Intel GPU specific
details
With this commit the 'info' RPC call is replaced with the
'info_dataspace' call that transfers the capability for the dataspace
containing the information only. This is complemented by a client
local 'attached_info' call that allows for getting typed access to
the information. The layout of the information is moved to its own
and GPU-specific header file, e.g., 'gpu/info_intel.h'
Issue #4265.
Rather than using the dataspace capability directly, let the client
choose its own local identifier that is linked to the underlying
capability.
Fixes#4265.
* Wait for for completion before return from 'execbuffer2'. This makes
buffer execution synchronous.
* Because the Iris driver manages the virtual address space of the GPU
and creates one GEM context for each batch buffer we have to map/unmap
all buffer objects before and after batch buffer execution.
issue #4260
This patch adds a switch to internal poll function in libssh that
allows to force this function to immediately return without actually
polling for data and in consequence processing this data. This switch
is used to avoid calling callback functions when flushing output
streams which caused locks due to recursive access to internal
ssh_terminal sessions registry.
Issue #4258
This patch allows to replace sftp packet read and write with
completely asynchronous versions needed to properly hook in existing
ssh_terminal implementation.
Issue #4258
* Switch mesa support from DRI to gallium
Supported drivers are
- softpipe (Sebstian Sumpf)
- iris for Intel GPUs (Alexander Boetcher)
- etnaviv for Vivante GPUs (Josef Söntgen)
* Mesa's generated files are placed into 'contrib/mesa-<hash>/generated'
and are cloned per default from a separate Git repo in order to avoid
hash updates upon package build. In case you need to generate files
yourself use
! prepare_port mesa GENERATE_FILES=1
issue #4254
Support for iris and etnvaviv
* entaviv:
- libdrm on FreeBSD is not prepared for !PCI (and libc our is missing
<sys/pciio.h>
- missing <sys/types.h> include in xf86drmMode.c
- etnaviv relies on linux header files - dummy in $(INC_DIR)
- IOCTL FreeBSD ↔ Linux have swapped IO/OUT bit
- O_CLOEXEC differs between FreeBSD ↔ Linux
issue #4254
