Commit graph

10 commits

Author SHA1 Message Date
8501cb7ed8 Allow users to request a asynchronous notification
This patch adds functionality to request an asynchronous notification
if some other process has send a message to it. We currently send a
SIGALRM to avoid any changes to core infrastructure of signal handling.
Ideally we should implement a separate signal for this. Currently there
is an obvious disadvantage of not being able to use SIGALRM with alarm,
if we are using this functionality.
2016-03-22 17:51:09 +05:30
55d341bc54 Allow users to get and set attributes
Currently we only allow users to get and set number of queues. The
operation is always non blocking. There is no provision for non
blocking operation at the moment.

The queue sizes are set to MAX 256 by default. If a user changes the
attributes and sets the queue size or number of messages in a queue
to less than 256, it is the responsibility of the user to close any
queues which were opened with a descriptor greater than the value
currently set.

For example, by default queue sizes are 256. If 150 queues are opened
and then the user sets the number of queues to 32, all open descriptors
and queues from 32 to 150 must be closed. Note that descriptor number
allocation starts with zero and not one.
2016-03-22 15:57:53 +05:30
d2fb3a4b8a Allow user processes to send messages to kernel
This patch allows user processes to send messages to kernel. Without
this change, trying to do a system call for using message queues like
sys_mq_open will fail with ECALLDENIED error number -210.
2016-03-16 23:30:06 +05:30
6f36b466f3 minix: Prepare for message queue support to Minix kernel
This commit adds preparatory support for forthcoming message
queue implementation to the Minix kernel. The core implementation
will be added with the next commit.
2016-02-29 20:43:07 +05:30
David van Moolenbroek
0f8e20a12c getrusage(2): zero out ru_nsignals field
The current value was both wrong (counting spawned kernel signals
rather than delivered user signals) and returned for the calling
process even if the request was for the process's children.
For now we are better off not populating this field at all.

Change-Id: I6c660be266b5746b7c3db57ae88fa7f872961ee2
2015-09-28 14:06:58 +00:00
David van Moolenbroek
3779ed93c3 Kernel: IPC filter support for VM memory requests
The filtering also exposed the risk that a process be killed or
swapped while on the list of VM memory requests.  These cases are
now handled properly as well.

Change-Id: Ibd3897b34abdf33bce19d37b8e5f65fbd0fd9316
2015-09-17 13:46:23 +00:00
Cristiano Giuffrida
56e56d2af2 kernel: Inherit allowed memory, IRQ & I/O ranges
Inherit at live update time all the priviledges previously acquired by
the process.

Change-Id: I1fd6ab16680ea85a1f3d8d1c106a7c716bf65e8e
2015-09-16 11:02:44 +00:00
Cristiano Giuffrida
c8a9900b0c kernel: Add support for IPC filters.
Edited by David van Moolenbroek.

Change-Id: Ia0052c42a0f218d011dd2da1e3db6c5b2107adc7
2015-09-16 11:02:24 +00:00
David van Moolenbroek
1dcfbcd173 Remove support for call profiling
The entire infrastructure relied on an ACK feature, and as such, it
has been broken for years now, with no easy way to repair it.

Change-Id: I783c2a21276967af115a642199f31fef0f14a572
2014-09-30 12:39:03 +00:00
Lionel Sambuc
433d6423c3 New sources layout
Change-Id: Ic716f336b7071063997cf5b4dae6d50e0b4631e9
2014-07-31 16:00:30 +02:00
Renamed from kernel/system.c (Browse further)