- 09 Nov, 2012 2 commits
-
-
Alessandro Rubini authored
Only useful to diagnose buf-vmalloc, but _there_ it is useful. Signed-off-by:
Alessandro Rubini <rubini@gnudd.com> Acked-by:
Federico Vaga <federico.vaga@gmail.com>
-
Alessandro Rubini authored
Signed-off-by:
-
- 08 Nov, 2012 5 commits
-
-
Alessandro Rubini authored
The driver itself is allowed to change nsamples, because it may have less samples than requested when the trigger fires. This is done by zio-mini, for example, that will be merged soon. Therefore, the number in current_ctrl must be updated, or the previous shorter value will stick. Also, move seq_num incrementing earlier, so even if allocation fails the sequence number is updated (and we need to do alarms too for that). Signed-off-by:
-
Alessandro Rubini authored
Cset 4 of zio-loop returns a combined control and data stream to /dev/zio-loop-4-r-ctrldata . The same will soon happen with PF_ZIO sockets, so add "-c" to zio-dump to support combined mode (i.e. control is followed by data). Signed-off-by:
-
Alessandro Rubini authored
The driver supports an output cset that loops to an input cset, and two output csets of 1 channel each, whose output is returned to two char devices. Intended for demonstration and testing. Signed-off-by:
-
Alessandro Rubini authored
This is either one-shot or periodic. For output it uses the timestamp structure in the control block, for input it doesn't yet. Output may have some buglets; I'll fix when zio-loop is in place for diagnostics. Signed-off-by:
-
Alessandro Rubini authored
Just like drivers should update current_ctrl and not block->ctrl, the same happens to buffers. This only applies to vmalloc, though. Bug introduced by my commit 6a7378a5: zio-sys: delay copying current_ctrl to block->ctrl Signed-off-by:
-
- 05 Nov, 2012 16 commits
-
-
Alessandro Rubini authored
This removes the annyoing message about the missing attribute. Simple driver can choose to not have attributes at all, which simplifies source code, so the core should better not complain about them. Signed-off-by:
-
Federico Vaga authored
This fixes all public names to use zio_ or ZIO_ as a prefix. This introduces incompatibilities for the users, but was a bad design choice that stuck for too long. We are going to tag 1.0 soon, so this was unavoidable.
-
Federico Vaga authored
This patch removes the union field priv in the zio_attribute and it introduces the id field. From our experience, the priv.ptr pointer is useless; the priv.addr field can be used for every purpose, also if you have your own private structure. The new field id replaces the role of priv.addr. It can represent a register address, a register offset, an index of a vector of private structure which describe how to gain access to a register, or whatever can describe uniquely a register (or register field) Signed-off-by:
-
Federico Vaga authored
Signed-off-by:
-
Federico Vaga authored
Signed-off-by:
-
Federico Vaga authored
Signed-off-by:
-
Federico Vaga authored
Signed-off-by:
-
Federico Vaga authored
Signed-off-by:
-
Federico Vaga authored
Signed-off-by:
-
Federico Vaga authored
Signed-off-by:
-
Federico Vaga authored
Signed-off-by:
-
Federico Vaga authored
Signed-off-by:
-
Federico Vaga authored
Signed-off-by:
-
Federico Vaga authored
Signed-off-by:
-
Federico Vaga authored
This function can be used outside the ZIO core. At the moment there are not driver which use this function, but it can be useful in the future. Signed-off-by:
-
Federico Vaga authored
This function can be used outside the ZIO core (e.g. AD788x driver). A driver can be interested to know how many channel are active during an acquisition to allocate the proper buffer/structure Signed-off-by:
-
- 04 Nov, 2012 3 commits
-
-
Federico Vaga authored
Signed-off-by:
-
Federico Vaga authored
Signed-off-by:
-
Federico Vaga authored
This is more in line with current kernel practices. We do not use zio_ here because the context is always clear when this macro is used. Signed-off-by:
-
- 03 Nov, 2012 2 commits
-
-
Federico Vaga authored
Signed-off-by:
-
Federico Vaga authored
Signed-off-by:
-
- 01 Nov, 2012 4 commits
-
-
Alessandro Rubini authored
Changes attribute print-out to only show attributes that are defined (using the mask). Show them as both decimal and hex. All lines are grep-friendly, as the used to be. You can grep for "Ctrl:" or "Data:" or specific attributes. Example (with "-a"): Ctrl: version 0.7, trigger user, dev zloop-0000, cset 1, chan 1 Ctrl: seq 18, n 16, size 1, bits 8, flags 01000001 (little-endian) Ctrl: stamp 1351773544.171337188 (0) Ctrl: trigger-std-mask: 0x0002 Ctrl: trigger-std-1 0x00000010 16 Data: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 For those who, sometimes, want to see all attributes, even when the mask is zero, there is the "-A" option. Signed-off-by:
-
Alessandro Rubini authored
Signed-off-by: -
Alessandro Rubini authored
Signed-off-by:
-
Alessandro Rubini authored
So the inline that followed is in the right place Signed-off-by:
-
- 31 Oct, 2012 2 commits
-
-
Alessandro Rubini authored
Signed-off-by: -
Alessandro Rubini authored
This is a potentially incompatible change: if your driver fills attributes in the input block at raw_io() time, it won't work any more. You should fill chan->current_ctrl instead, which will be visible as a binary blob in sysfs soon. Moving the memcpy() to a later point doesn't change overhead. Signed-off-by:
-
- 29 Oct, 2012 4 commits
-
-
Federico Vaga authored
If a driver module registerscreate its own preferred trigger or buffer, we can't increment its own use count, or we won't be able to unload the module. This patch resolves the issue. Now the ZIO core doesn't increment the use counter of the trigger or buffer, if it's owned by the device. Signed-off-by:
-
Alessandro Rubini authored
By copying ti->tstamp to the control at a later time, we support devices that use an internal time stamp (e.g: white rabbit). Signed-off-by:
-
Alessandro Rubini authored
If the user changes the buffer size, this might make more space for a sleeping writer. Similarly, the buffer may be fuller than the new value after a change, so we check for ">= limit" and not "== limit". This doesn't apply to the vmalloc buffer, that currently doesn't allow a size change. Signed-off-by:
-
Alessandro Rubini authored
buffer->store_block calls trigger->push_block when it receives the first block. The trigger user fires directly , and if the device accepts the data immediately, our data_done() will retr_block from the buffer. This happens with zio-zero, cset 1, and leads to a deadlock. To allow immediate success of push_block down to device level, release the lock before calling push_block. The code here also handles pathologic situations like two processes storing at the same time. Signed-off-by:
-
- 26 Oct, 2012 2 commits
-
-
Federico Vaga authored
Signed-off-by:
-
Federico Vaga authored
It is difficult to support a board which has pre-samples and post-samples registers and synchronize these value with the nsamples zio_attribute and control field. But, it is really simple for board that hasn't pre and post samples to use these two attributes because they can use only one attribute for their purpose (only pre-sample or only post-sample). Signed-off-by:
-
