Unfortunately, the implementation was not thread safe and was not
sticking to the thread local main context.
In addition to this, it is not entirely clear to me how this API should
behave. The current approach is to simply cancel the transition with the
state machine halting in its current state. Instead, it could also make
sense for cancellation to cause the state machine to return a
G_IO_ERROR_CANCELLED.
As such, simply remove the feature for now. If anyone actually has a
good use-case then we can add it again.
In some situations one may want to guarantee that the last steps of an
SSM are run even when the SSM is completed early or failed.
This can easily be done by making fpi_ssm_mark_completed jump to the
next cleanup stage when called (this also includes mark_failed). Due to
the mechanism, it is still possible to explicitly jump cleanup states by
using fpi_ssm_jump_to_state, including a jump to the final state in
order to skip all cleanup states.
We need more than 1 line for assembling, but in general, we should
have a reasonable amount of lines. So use the width in the hope we'll
get an image that is about square at least.
Closes: #135
Add a GCancellable parameter to fpi_ssm_nex_state_delayed and
fpi_ssm_jump_to_state_delayed() so that it's possible to cancel an action
from the caller and in case the driver wants to cancel a delayed operation
when a device action has been cancelled.
Since GSource data can be automatically cleaned up on source destruction, we
can mimic this for the devices timeout easily as well.
Add an extra parameter, and let's use this cocci file to adapt all the
drivers like magic:
@@
expression e1, e2, e3, e4;
@@
fpi_device_add_timeout (e1, e2, e3, e4
+ , NULL
)
When a transfer is completed, we automatically unref it since we can't
consider it valid anymore since this point.
Update the drivers not to free the transfer after submitting anymore.
When a machine is completed, we automatically free it since we can't
consider it valid anymore since this point.
Update the drivers not to free the SSM on completion callback anymore.
Use the same approach of GTask, making possible to set the data from a
function. Givent the fact that a SSM has now a device parameter, it's
generally not needed to pass an extra data value.
In such case make it possible to set it and to define a destroy-notify
function to handle its destruction when freeing the SSM.
The state was always AWAIT_FINGER and it was never used by any driver
(except for error checking). So remove it, in particular as a correct
state change will be done after activation anyway.
The only driver with code that actually did anything based on this was
the URU4000 driver. However, all it did was an explicit state change
execution. This is not necessary, as the state_change handler is called
anyway (i.e. we now only write the AWAIT_FINGER register once rather
than twice).
Manual changes plus:
@ init @
identifier driver_name;
identifier activate_func;
@@
struct fp_img_driver driver_name = {
...,
.activate = activate_func,
...,
};
@ remove_arg @
identifier dev;
identifier state;
identifier init.activate_func;
@@
activate_func (
struct fp_img_dev *dev
- , enum fp_imgdev_state state
)
{
<...
- if (state != IMGDEV_STATE_AWAIT_FINGER_ON) { ... }
...>
}
Pass the struct fp_dev and user_data to fpi_ssm callbacks, so that we
might be able to get rid of the fpi_ssm_get_user_data(), and
fpi_ssm_get_dev() as most drivers just get those from the ssm anyway
in their callbacks.
Which avoids passing zero lines to fpi_assemble_lines()
"gmem.c:130: failed to allocate 18446744073709551612 bytes"
#3 0x00007fe4f6ef428f in g_log (log_domain=log_domain@entry=0x7fe4f6f3506e "GLib", log_level=log_level@entry=G_LOG_LEVEL_ERROR, format=format@entry=0x7fe4f6f3e610 "%s: failed to allocate %lu bytes") at gmessages.c:1398
#4 0x00007fe4f6ef2ac4 in g_malloc0 (n_bytes=n_bytes@entry=18446744073709551612) at gmem.c:129
#5 0x00007fe4f8052020 in median_filter (filtersize=25, size=-1, data=0x0) at assembling.c:309
#6 fpi_assemble_lines (ctx=ctx@entry=0x7fe4f82ac3c0 <assembling_ctx>, lines=0x0, lines_len=0) at assembling.c:389
#7 0x00007fe4f805f3db in submit_image (ssm=ssm@entry=0x16c3cba360, data=data@entry=0x16c3cb9cc0) at drivers/vfs5011.c:412
See https://bugzilla.redhat.com/show_bug.cgi?id=1484812Closes: #42
Remove all the headers already included through "fp_internal.h" such as
<libusb.h> and <errno.h>, include "assembling.h" and "driver_ids.h" there
as well to avoid doing it in (almost) every driver.
It seems that on faster devices, the driver can get more duplicate lines.
Without this change scan looks too stretched.
Suggested by: Thomas Rinsma <thomasrinsma@gmail.com>
New driver for VFS5011 138a:0011 and 138a:0018
https://bugs.freedesktop.org/show_bug.cgi?id=61692
[vasilykh]:
- use g_get_real_time() instead of non-portable time()
- use g_free() instead of free()
- comment out "RECV(VFS5011_IN_ENDPOINT_CTRL2, 8)"
