minix/minix/lib/libchardriver/chardriver.c

599 lines
20 KiB
C
Raw Normal View History

Split block/character protocols and libdriver This patch separates the character and block driver communication protocols. The old character protocol remains the same, but a new block protocol is introduced. The libdriver library is replaced by two new libraries: libchardriver and libblockdriver. Their exposed API, and drivers that use them, have been updated accordingly. Together, libbdev and libblockdriver now completely abstract away the message format used by the block protocol. As the memory driver is both a character and a block device driver, it now implements its own message loop. The most important semantic change made to the block protocol is that it is no longer possible to return both partial results and an error for a single transfer. This simplifies the interaction between the caller and the driver, as the I/O vector no longer needs to be copied back. Also, drivers are now no longer supposed to decide based on the layout of the I/O vector when a transfer should be cut short. Put simply, transfers are now supposed to either succeed completely, or result in an error. After this patch, the state of the various pieces is as follows: - block protocol: stable - libbdev API: stable for synchronous communication - libblockdriver API: needs slight revision (the drvlib/partition API in particular; the threading API will also change shortly) - character protocol: needs cleanup - libchardriver API: needs cleanup accordingly - driver restarts: largely unsupported until endpoint changes are reintroduced As a side effect, this patch eliminates several bugs, hacks, and gcc -Wall and -W warnings all over the place. It probably introduces a few new ones, too. Update warning: this patch changes the protocol between MFS and disk drivers, so in order to use old/new images, the MFS from the ramdisk must be used to mount all file systems.
2011-11-22 13:27:53 +01:00
/* This file contains the device independent character driver interface.
2005-04-21 16:53:53 +02:00
*
Input infrastructure, INPUT server, PCKBD driver This commit separates the low-level keyboard driver from TTY, putting it in a separate driver (PCKBD). The commit also separates management of raw input devices from TTY, and puts it in a separate server (INPUT). All keyboard and mouse input from hardware is sent by drivers to the INPUT server, which either sends it to a process that has opened a raw input device, or otherwise forwards it to TTY for standard processing. Design by Dirk Vogt. Prototype by Uli Kastlunger. Additional changes made to the prototype: - the event communication is now based on USB HID codes; all input drivers have to use USB codes to describe events; - all TTY keymaps have been converted to USB format, with the effect that a single keymap covers all keys; there is no (static) escaped keymap anymore; - further keymap tweaks now allow remapping of literally all keys; - input device renumbering and protocol rewrite; - INPUT server rewrite, with added support for cancel and select; - PCKBD reimplementation, including PC/AT-to-USB translation; - support for manipulating keyboard LEDs has been added; - keyboard and mouse multiplexer devices have been added to INPUT, primarily so that an X server need only open two devices; - a new "libinputdriver" library abstracts away protocol details from input drivers, and should be used by all future input drivers; - both INPUT and PCKBD can be restarted; - TTY is now scheduled by KERNEL, so that it won't be punished for running a lot; without this, simply running "yes" on the console kills the system; - the KIOCBELL IOCTL has been moved to /dev/console; - support for the SCANCODES termios setting has been removed; - obsolete keymap compression has been removed; - the obsolete Olivetti M24 keymap has been removed. Change-Id: I3a672fb8c4fd566734e4b46d3994b4b7fc96d578
2013-09-28 14:46:21 +02:00
* Character drivers support the following requests. Message format m10 is
* used. Field names are prefixed with CDEV_. Separate field names are used for
* the "access", "ops", "user", and "request" fields.
2005-04-21 16:53:53 +02:00
*
* m_type MINOR GRANT COUNT FLAGS ID POS_LO POS_HI
* +-------------+-------+--------+-------+-------+------+---------+--------+
* | CDEV_OPEN | minor | access | user | | id | | |
* |-------------+-------+--------+-------+-------+------+---------+--------|
* | CDEV_CLOSE | minor | | | | id | | |
* |-------------+-------+--------+-------+-------+------+---------+--------|
* | CDEV_READ | minor | grant | bytes | flags | id | position |
* |-------------+-------+--------+-------+-------+------+---------+--------|
* | CDEV_WRITE | minor | grant | bytes | flags | id | position |
* |-------------+-------+--------+-------+-------+------+---------+--------|
* | CDEV_IOCTL | minor | grant | user | flags | id | request | |
* |-------------+-------+--------+-------+-------+------+---------+--------|
* | CDEV_CANCEL | minor | | | | id | | |
* |-------------+-------+--------+-------+-------+------+---------+--------|
* | CDEV_SELECT | minor | ops | | | | | |
* --------------------------------------------------------------------------
*
* The following reply messages are used.
*
* m_type MINOR STATUS ID
* +-----------------+-------+--------+-----+-----+------+---------+--------+
* | CDEV_REPLY | | status | | | id | | |
* |-----------------+-------+--------+-----+-----+------+---------+--------|
* | CDEV_SEL1_REPLY | minor | status | | | | | |
* |-----------------+-------+--------+-----+-----+------+---------+--------|
* | CDEV_SEL2_REPLY | minor | status | | | | | |
* --------------------------------------------------------------------------
2005-04-21 16:53:53 +02:00
*
* Changes:
* Sep 01, 2013 complete rewrite of the API (D.C. van Moolenboek)
* Aug 20, 2013 retire synchronous protocol (D.C. van Moolenbroek)
Split block/character protocols and libdriver This patch separates the character and block driver communication protocols. The old character protocol remains the same, but a new block protocol is introduced. The libdriver library is replaced by two new libraries: libchardriver and libblockdriver. Their exposed API, and drivers that use them, have been updated accordingly. Together, libbdev and libblockdriver now completely abstract away the message format used by the block protocol. As the memory driver is both a character and a block device driver, it now implements its own message loop. The most important semantic change made to the block protocol is that it is no longer possible to return both partial results and an error for a single transfer. This simplifies the interaction between the caller and the driver, as the I/O vector no longer needs to be copied back. Also, drivers are now no longer supposed to decide based on the layout of the I/O vector when a transfer should be cut short. Put simply, transfers are now supposed to either succeed completely, or result in an error. After this patch, the state of the various pieces is as follows: - block protocol: stable - libbdev API: stable for synchronous communication - libblockdriver API: needs slight revision (the drvlib/partition API in particular; the threading API will also change shortly) - character protocol: needs cleanup - libchardriver API: needs cleanup accordingly - driver restarts: largely unsupported until endpoint changes are reintroduced As a side effect, this patch eliminates several bugs, hacks, and gcc -Wall and -W warnings all over the place. It probably introduces a few new ones, too. Update warning: this patch changes the protocol between MFS and disk drivers, so in order to use old/new images, the MFS from the ramdisk must be used to mount all file systems.
2011-11-22 13:27:53 +01:00
* Oct 16, 2011 split character and block protocol (D.C. van Moolenbroek)
* Aug 27, 2011 move common functions into driver.c (A. Welzel)
* Jul 25, 2005 added SYS_SIG type for signals (Jorrit N. Herder)
* Sep 15, 2004 added SYN_ALARM type for timeouts (Jorrit N. Herder)
* Jul 23, 2004 removed kernel dependencies (Jorrit N. Herder)
* Apr 02, 1992 constructed from AT wini and floppy driver (Kees J. Bot)
2005-04-21 16:53:53 +02:00
*/
#include <assert.h>
#include <minix/drivers.h>
Split block/character protocols and libdriver This patch separates the character and block driver communication protocols. The old character protocol remains the same, but a new block protocol is introduced. The libdriver library is replaced by two new libraries: libchardriver and libblockdriver. Their exposed API, and drivers that use them, have been updated accordingly. Together, libbdev and libblockdriver now completely abstract away the message format used by the block protocol. As the memory driver is both a character and a block device driver, it now implements its own message loop. The most important semantic change made to the block protocol is that it is no longer possible to return both partial results and an error for a single transfer. This simplifies the interaction between the caller and the driver, as the I/O vector no longer needs to be copied back. Also, drivers are now no longer supposed to decide based on the layout of the I/O vector when a transfer should be cut short. Put simply, transfers are now supposed to either succeed completely, or result in an error. After this patch, the state of the various pieces is as follows: - block protocol: stable - libbdev API: stable for synchronous communication - libblockdriver API: needs slight revision (the drvlib/partition API in particular; the threading API will also change shortly) - character protocol: needs cleanup - libchardriver API: needs cleanup accordingly - driver restarts: largely unsupported until endpoint changes are reintroduced As a side effect, this patch eliminates several bugs, hacks, and gcc -Wall and -W warnings all over the place. It probably introduces a few new ones, too. Update warning: this patch changes the protocol between MFS and disk drivers, so in order to use old/new images, the MFS from the ramdisk must be used to mount all file systems.
2011-11-22 13:27:53 +01:00
#include <minix/chardriver.h>
Driver refactory for live update and crash recovery. SYSLIB CHANGES: - DS calls to publish / retrieve labels consider endpoints instead of u32_t. VFS CHANGES: - mapdriver() only adds an entry in the dmap table in VFS. - dev_up() is only executed upon reception of a driver up event. INET CHANGES: - INET no longer searches for existing drivers instances at startup. - A newtwork driver is (re)initialized upon reception of a driver up event. - Networking startup is now race-free by design. No need to waste 5 seconds at startup any more. DRIVER CHANGES: - Every driver publishes driver up events when starting for the first time or in case of restart when recovery actions must be taken in the upper layers. - Driver up events are published by drivers through DS. - For regular drivers, VFS is normally the only subscriber, but not necessarily. For instance, when the filter driver is in use, it must subscribe to driver up events to initiate recovery. - For network drivers, inet is the only subscriber for now. - Every VFS driver is statically linked with libdriver, every network driver is statically linked with libnetdriver. DRIVER LIBRARIES CHANGES: - Libdriver is extended to provide generic receive() and ds_publish() interfaces for VFS drivers. - driver_receive() is a wrapper for sef_receive() also used in driver_task() to discard spurious messages that were meant to be delivered to a previous version of the driver. - driver_receive_mq() is the same as driver_receive() but integrates support for queued messages. - driver_announce() publishes a driver up event for VFS drivers and marks the driver as initialized and expecting a DEV_OPEN message. - Libnetdriver is introduced to provide similar receive() and ds_publish() interfaces for network drivers (netdriver_announce() and netdriver_receive()). - Network drivers all support live update with no state transfer now. KERNEL CHANGES: - Added kernel call statectl for state management. Used by driver_announce() to unblock eventual callers sendrecing to the driver.
2010-04-08 15:41:35 +02:00
#include <minix/ds.h>
2005-04-21 16:53:53 +02:00
2012-03-25 20:25:53 +02:00
static int running;
/* Management data for opened devices. */
static devminor_t open_devs[MAX_NR_OPEN_DEVICES];
2012-03-25 20:25:53 +02:00
static int next_open_devs_slot = 0;
Driver refactory for live update and crash recovery. SYSLIB CHANGES: - DS calls to publish / retrieve labels consider endpoints instead of u32_t. VFS CHANGES: - mapdriver() only adds an entry in the dmap table in VFS. - dev_up() is only executed upon reception of a driver up event. INET CHANGES: - INET no longer searches for existing drivers instances at startup. - A newtwork driver is (re)initialized upon reception of a driver up event. - Networking startup is now race-free by design. No need to waste 5 seconds at startup any more. DRIVER CHANGES: - Every driver publishes driver up events when starting for the first time or in case of restart when recovery actions must be taken in the upper layers. - Driver up events are published by drivers through DS. - For regular drivers, VFS is normally the only subscriber, but not necessarily. For instance, when the filter driver is in use, it must subscribe to driver up events to initiate recovery. - For network drivers, inet is the only subscriber for now. - Every VFS driver is statically linked with libdriver, every network driver is statically linked with libnetdriver. DRIVER LIBRARIES CHANGES: - Libdriver is extended to provide generic receive() and ds_publish() interfaces for VFS drivers. - driver_receive() is a wrapper for sef_receive() also used in driver_task() to discard spurious messages that were meant to be delivered to a previous version of the driver. - driver_receive_mq() is the same as driver_receive() but integrates support for queued messages. - driver_announce() publishes a driver up event for VFS drivers and marks the driver as initialized and expecting a DEV_OPEN message. - Libnetdriver is introduced to provide similar receive() and ds_publish() interfaces for network drivers (netdriver_announce() and netdriver_receive()). - Network drivers all support live update with no state transfer now. KERNEL CHANGES: - Added kernel call statectl for state management. Used by driver_announce() to unblock eventual callers sendrecing to the driver.
2010-04-08 15:41:35 +02:00
/*===========================================================================*
* clear_open_devs *
Driver refactory for live update and crash recovery. SYSLIB CHANGES: - DS calls to publish / retrieve labels consider endpoints instead of u32_t. VFS CHANGES: - mapdriver() only adds an entry in the dmap table in VFS. - dev_up() is only executed upon reception of a driver up event. INET CHANGES: - INET no longer searches for existing drivers instances at startup. - A newtwork driver is (re)initialized upon reception of a driver up event. - Networking startup is now race-free by design. No need to waste 5 seconds at startup any more. DRIVER CHANGES: - Every driver publishes driver up events when starting for the first time or in case of restart when recovery actions must be taken in the upper layers. - Driver up events are published by drivers through DS. - For regular drivers, VFS is normally the only subscriber, but not necessarily. For instance, when the filter driver is in use, it must subscribe to driver up events to initiate recovery. - For network drivers, inet is the only subscriber for now. - Every VFS driver is statically linked with libdriver, every network driver is statically linked with libnetdriver. DRIVER LIBRARIES CHANGES: - Libdriver is extended to provide generic receive() and ds_publish() interfaces for VFS drivers. - driver_receive() is a wrapper for sef_receive() also used in driver_task() to discard spurious messages that were meant to be delivered to a previous version of the driver. - driver_receive_mq() is the same as driver_receive() but integrates support for queued messages. - driver_announce() publishes a driver up event for VFS drivers and marks the driver as initialized and expecting a DEV_OPEN message. - Libnetdriver is introduced to provide similar receive() and ds_publish() interfaces for network drivers (netdriver_announce() and netdriver_receive()). - Network drivers all support live update with no state transfer now. KERNEL CHANGES: - Added kernel call statectl for state management. Used by driver_announce() to unblock eventual callers sendrecing to the driver.
2010-04-08 15:41:35 +02:00
*===========================================================================*/
2012-03-25 20:25:53 +02:00
static void clear_open_devs(void)
Driver refactory for live update and crash recovery. SYSLIB CHANGES: - DS calls to publish / retrieve labels consider endpoints instead of u32_t. VFS CHANGES: - mapdriver() only adds an entry in the dmap table in VFS. - dev_up() is only executed upon reception of a driver up event. INET CHANGES: - INET no longer searches for existing drivers instances at startup. - A newtwork driver is (re)initialized upon reception of a driver up event. - Networking startup is now race-free by design. No need to waste 5 seconds at startup any more. DRIVER CHANGES: - Every driver publishes driver up events when starting for the first time or in case of restart when recovery actions must be taken in the upper layers. - Driver up events are published by drivers through DS. - For regular drivers, VFS is normally the only subscriber, but not necessarily. For instance, when the filter driver is in use, it must subscribe to driver up events to initiate recovery. - For network drivers, inet is the only subscriber for now. - Every VFS driver is statically linked with libdriver, every network driver is statically linked with libnetdriver. DRIVER LIBRARIES CHANGES: - Libdriver is extended to provide generic receive() and ds_publish() interfaces for VFS drivers. - driver_receive() is a wrapper for sef_receive() also used in driver_task() to discard spurious messages that were meant to be delivered to a previous version of the driver. - driver_receive_mq() is the same as driver_receive() but integrates support for queued messages. - driver_announce() publishes a driver up event for VFS drivers and marks the driver as initialized and expecting a DEV_OPEN message. - Libnetdriver is introduced to provide similar receive() and ds_publish() interfaces for network drivers (netdriver_announce() and netdriver_receive()). - Network drivers all support live update with no state transfer now. KERNEL CHANGES: - Added kernel call statectl for state management. Used by driver_announce() to unblock eventual callers sendrecing to the driver.
2010-04-08 15:41:35 +02:00
{
/* Reset the set of previously opened minor devices. */
Driver refactory for live update and crash recovery. SYSLIB CHANGES: - DS calls to publish / retrieve labels consider endpoints instead of u32_t. VFS CHANGES: - mapdriver() only adds an entry in the dmap table in VFS. - dev_up() is only executed upon reception of a driver up event. INET CHANGES: - INET no longer searches for existing drivers instances at startup. - A newtwork driver is (re)initialized upon reception of a driver up event. - Networking startup is now race-free by design. No need to waste 5 seconds at startup any more. DRIVER CHANGES: - Every driver publishes driver up events when starting for the first time or in case of restart when recovery actions must be taken in the upper layers. - Driver up events are published by drivers through DS. - For regular drivers, VFS is normally the only subscriber, but not necessarily. For instance, when the filter driver is in use, it must subscribe to driver up events to initiate recovery. - For network drivers, inet is the only subscriber for now. - Every VFS driver is statically linked with libdriver, every network driver is statically linked with libnetdriver. DRIVER LIBRARIES CHANGES: - Libdriver is extended to provide generic receive() and ds_publish() interfaces for VFS drivers. - driver_receive() is a wrapper for sef_receive() also used in driver_task() to discard spurious messages that were meant to be delivered to a previous version of the driver. - driver_receive_mq() is the same as driver_receive() but integrates support for queued messages. - driver_announce() publishes a driver up event for VFS drivers and marks the driver as initialized and expecting a DEV_OPEN message. - Libnetdriver is introduced to provide similar receive() and ds_publish() interfaces for network drivers (netdriver_announce() and netdriver_receive()). - Network drivers all support live update with no state transfer now. KERNEL CHANGES: - Added kernel call statectl for state management. Used by driver_announce() to unblock eventual callers sendrecing to the driver.
2010-04-08 15:41:35 +02:00
next_open_devs_slot = 0;
}
/*===========================================================================*
* is_open_dev *
Driver refactory for live update and crash recovery. SYSLIB CHANGES: - DS calls to publish / retrieve labels consider endpoints instead of u32_t. VFS CHANGES: - mapdriver() only adds an entry in the dmap table in VFS. - dev_up() is only executed upon reception of a driver up event. INET CHANGES: - INET no longer searches for existing drivers instances at startup. - A newtwork driver is (re)initialized upon reception of a driver up event. - Networking startup is now race-free by design. No need to waste 5 seconds at startup any more. DRIVER CHANGES: - Every driver publishes driver up events when starting for the first time or in case of restart when recovery actions must be taken in the upper layers. - Driver up events are published by drivers through DS. - For regular drivers, VFS is normally the only subscriber, but not necessarily. For instance, when the filter driver is in use, it must subscribe to driver up events to initiate recovery. - For network drivers, inet is the only subscriber for now. - Every VFS driver is statically linked with libdriver, every network driver is statically linked with libnetdriver. DRIVER LIBRARIES CHANGES: - Libdriver is extended to provide generic receive() and ds_publish() interfaces for VFS drivers. - driver_receive() is a wrapper for sef_receive() also used in driver_task() to discard spurious messages that were meant to be delivered to a previous version of the driver. - driver_receive_mq() is the same as driver_receive() but integrates support for queued messages. - driver_announce() publishes a driver up event for VFS drivers and marks the driver as initialized and expecting a DEV_OPEN message. - Libnetdriver is introduced to provide similar receive() and ds_publish() interfaces for network drivers (netdriver_announce() and netdriver_receive()). - Network drivers all support live update with no state transfer now. KERNEL CHANGES: - Added kernel call statectl for state management. Used by driver_announce() to unblock eventual callers sendrecing to the driver.
2010-04-08 15:41:35 +02:00
*===========================================================================*/
static int is_open_dev(devminor_t minor)
Driver refactory for live update and crash recovery. SYSLIB CHANGES: - DS calls to publish / retrieve labels consider endpoints instead of u32_t. VFS CHANGES: - mapdriver() only adds an entry in the dmap table in VFS. - dev_up() is only executed upon reception of a driver up event. INET CHANGES: - INET no longer searches for existing drivers instances at startup. - A newtwork driver is (re)initialized upon reception of a driver up event. - Networking startup is now race-free by design. No need to waste 5 seconds at startup any more. DRIVER CHANGES: - Every driver publishes driver up events when starting for the first time or in case of restart when recovery actions must be taken in the upper layers. - Driver up events are published by drivers through DS. - For regular drivers, VFS is normally the only subscriber, but not necessarily. For instance, when the filter driver is in use, it must subscribe to driver up events to initiate recovery. - For network drivers, inet is the only subscriber for now. - Every VFS driver is statically linked with libdriver, every network driver is statically linked with libnetdriver. DRIVER LIBRARIES CHANGES: - Libdriver is extended to provide generic receive() and ds_publish() interfaces for VFS drivers. - driver_receive() is a wrapper for sef_receive() also used in driver_task() to discard spurious messages that were meant to be delivered to a previous version of the driver. - driver_receive_mq() is the same as driver_receive() but integrates support for queued messages. - driver_announce() publishes a driver up event for VFS drivers and marks the driver as initialized and expecting a DEV_OPEN message. - Libnetdriver is introduced to provide similar receive() and ds_publish() interfaces for network drivers (netdriver_announce() and netdriver_receive()). - Network drivers all support live update with no state transfer now. KERNEL CHANGES: - Added kernel call statectl for state management. Used by driver_announce() to unblock eventual callers sendrecing to the driver.
2010-04-08 15:41:35 +02:00
{
/* Check whether the given minor device has previously been opened. */
int i;
for (i = 0; i < next_open_devs_slot; i++)
if (open_devs[i] == minor)
return TRUE;
Driver refactory for live update and crash recovery. SYSLIB CHANGES: - DS calls to publish / retrieve labels consider endpoints instead of u32_t. VFS CHANGES: - mapdriver() only adds an entry in the dmap table in VFS. - dev_up() is only executed upon reception of a driver up event. INET CHANGES: - INET no longer searches for existing drivers instances at startup. - A newtwork driver is (re)initialized upon reception of a driver up event. - Networking startup is now race-free by design. No need to waste 5 seconds at startup any more. DRIVER CHANGES: - Every driver publishes driver up events when starting for the first time or in case of restart when recovery actions must be taken in the upper layers. - Driver up events are published by drivers through DS. - For regular drivers, VFS is normally the only subscriber, but not necessarily. For instance, when the filter driver is in use, it must subscribe to driver up events to initiate recovery. - For network drivers, inet is the only subscriber for now. - Every VFS driver is statically linked with libdriver, every network driver is statically linked with libnetdriver. DRIVER LIBRARIES CHANGES: - Libdriver is extended to provide generic receive() and ds_publish() interfaces for VFS drivers. - driver_receive() is a wrapper for sef_receive() also used in driver_task() to discard spurious messages that were meant to be delivered to a previous version of the driver. - driver_receive_mq() is the same as driver_receive() but integrates support for queued messages. - driver_announce() publishes a driver up event for VFS drivers and marks the driver as initialized and expecting a DEV_OPEN message. - Libnetdriver is introduced to provide similar receive() and ds_publish() interfaces for network drivers (netdriver_announce() and netdriver_receive()). - Network drivers all support live update with no state transfer now. KERNEL CHANGES: - Added kernel call statectl for state management. Used by driver_announce() to unblock eventual callers sendrecing to the driver.
2010-04-08 15:41:35 +02:00
return FALSE;
Driver refactory for live update and crash recovery. SYSLIB CHANGES: - DS calls to publish / retrieve labels consider endpoints instead of u32_t. VFS CHANGES: - mapdriver() only adds an entry in the dmap table in VFS. - dev_up() is only executed upon reception of a driver up event. INET CHANGES: - INET no longer searches for existing drivers instances at startup. - A newtwork driver is (re)initialized upon reception of a driver up event. - Networking startup is now race-free by design. No need to waste 5 seconds at startup any more. DRIVER CHANGES: - Every driver publishes driver up events when starting for the first time or in case of restart when recovery actions must be taken in the upper layers. - Driver up events are published by drivers through DS. - For regular drivers, VFS is normally the only subscriber, but not necessarily. For instance, when the filter driver is in use, it must subscribe to driver up events to initiate recovery. - For network drivers, inet is the only subscriber for now. - Every VFS driver is statically linked with libdriver, every network driver is statically linked with libnetdriver. DRIVER LIBRARIES CHANGES: - Libdriver is extended to provide generic receive() and ds_publish() interfaces for VFS drivers. - driver_receive() is a wrapper for sef_receive() also used in driver_task() to discard spurious messages that were meant to be delivered to a previous version of the driver. - driver_receive_mq() is the same as driver_receive() but integrates support for queued messages. - driver_announce() publishes a driver up event for VFS drivers and marks the driver as initialized and expecting a DEV_OPEN message. - Libnetdriver is introduced to provide similar receive() and ds_publish() interfaces for network drivers (netdriver_announce() and netdriver_receive()). - Network drivers all support live update with no state transfer now. KERNEL CHANGES: - Added kernel call statectl for state management. Used by driver_announce() to unblock eventual callers sendrecing to the driver.
2010-04-08 15:41:35 +02:00
}
/*===========================================================================*
* set_open_dev *
Driver refactory for live update and crash recovery. SYSLIB CHANGES: - DS calls to publish / retrieve labels consider endpoints instead of u32_t. VFS CHANGES: - mapdriver() only adds an entry in the dmap table in VFS. - dev_up() is only executed upon reception of a driver up event. INET CHANGES: - INET no longer searches for existing drivers instances at startup. - A newtwork driver is (re)initialized upon reception of a driver up event. - Networking startup is now race-free by design. No need to waste 5 seconds at startup any more. DRIVER CHANGES: - Every driver publishes driver up events when starting for the first time or in case of restart when recovery actions must be taken in the upper layers. - Driver up events are published by drivers through DS. - For regular drivers, VFS is normally the only subscriber, but not necessarily. For instance, when the filter driver is in use, it must subscribe to driver up events to initiate recovery. - For network drivers, inet is the only subscriber for now. - Every VFS driver is statically linked with libdriver, every network driver is statically linked with libnetdriver. DRIVER LIBRARIES CHANGES: - Libdriver is extended to provide generic receive() and ds_publish() interfaces for VFS drivers. - driver_receive() is a wrapper for sef_receive() also used in driver_task() to discard spurious messages that were meant to be delivered to a previous version of the driver. - driver_receive_mq() is the same as driver_receive() but integrates support for queued messages. - driver_announce() publishes a driver up event for VFS drivers and marks the driver as initialized and expecting a DEV_OPEN message. - Libnetdriver is introduced to provide similar receive() and ds_publish() interfaces for network drivers (netdriver_announce() and netdriver_receive()). - Network drivers all support live update with no state transfer now. KERNEL CHANGES: - Added kernel call statectl for state management. Used by driver_announce() to unblock eventual callers sendrecing to the driver.
2010-04-08 15:41:35 +02:00
*===========================================================================*/
static void set_open_dev(devminor_t minor)
Driver refactory for live update and crash recovery. SYSLIB CHANGES: - DS calls to publish / retrieve labels consider endpoints instead of u32_t. VFS CHANGES: - mapdriver() only adds an entry in the dmap table in VFS. - dev_up() is only executed upon reception of a driver up event. INET CHANGES: - INET no longer searches for existing drivers instances at startup. - A newtwork driver is (re)initialized upon reception of a driver up event. - Networking startup is now race-free by design. No need to waste 5 seconds at startup any more. DRIVER CHANGES: - Every driver publishes driver up events when starting for the first time or in case of restart when recovery actions must be taken in the upper layers. - Driver up events are published by drivers through DS. - For regular drivers, VFS is normally the only subscriber, but not necessarily. For instance, when the filter driver is in use, it must subscribe to driver up events to initiate recovery. - For network drivers, inet is the only subscriber for now. - Every VFS driver is statically linked with libdriver, every network driver is statically linked with libnetdriver. DRIVER LIBRARIES CHANGES: - Libdriver is extended to provide generic receive() and ds_publish() interfaces for VFS drivers. - driver_receive() is a wrapper for sef_receive() also used in driver_task() to discard spurious messages that were meant to be delivered to a previous version of the driver. - driver_receive_mq() is the same as driver_receive() but integrates support for queued messages. - driver_announce() publishes a driver up event for VFS drivers and marks the driver as initialized and expecting a DEV_OPEN message. - Libnetdriver is introduced to provide similar receive() and ds_publish() interfaces for network drivers (netdriver_announce() and netdriver_receive()). - Network drivers all support live update with no state transfer now. KERNEL CHANGES: - Added kernel call statectl for state management. Used by driver_announce() to unblock eventual callers sendrecing to the driver.
2010-04-08 15:41:35 +02:00
{
/* Mark the given minor device as having been opened. */
if (next_open_devs_slot >= MAX_NR_OPEN_DEVICES)
panic("out of slots for open devices");
open_devs[next_open_devs_slot] = minor;
Driver refactory for live update and crash recovery. SYSLIB CHANGES: - DS calls to publish / retrieve labels consider endpoints instead of u32_t. VFS CHANGES: - mapdriver() only adds an entry in the dmap table in VFS. - dev_up() is only executed upon reception of a driver up event. INET CHANGES: - INET no longer searches for existing drivers instances at startup. - A newtwork driver is (re)initialized upon reception of a driver up event. - Networking startup is now race-free by design. No need to waste 5 seconds at startup any more. DRIVER CHANGES: - Every driver publishes driver up events when starting for the first time or in case of restart when recovery actions must be taken in the upper layers. - Driver up events are published by drivers through DS. - For regular drivers, VFS is normally the only subscriber, but not necessarily. For instance, when the filter driver is in use, it must subscribe to driver up events to initiate recovery. - For network drivers, inet is the only subscriber for now. - Every VFS driver is statically linked with libdriver, every network driver is statically linked with libnetdriver. DRIVER LIBRARIES CHANGES: - Libdriver is extended to provide generic receive() and ds_publish() interfaces for VFS drivers. - driver_receive() is a wrapper for sef_receive() also used in driver_task() to discard spurious messages that were meant to be delivered to a previous version of the driver. - driver_receive_mq() is the same as driver_receive() but integrates support for queued messages. - driver_announce() publishes a driver up event for VFS drivers and marks the driver as initialized and expecting a DEV_OPEN message. - Libnetdriver is introduced to provide similar receive() and ds_publish() interfaces for network drivers (netdriver_announce() and netdriver_receive()). - Network drivers all support live update with no state transfer now. KERNEL CHANGES: - Added kernel call statectl for state management. Used by driver_announce() to unblock eventual callers sendrecing to the driver.
2010-04-08 15:41:35 +02:00
next_open_devs_slot++;
}
2005-04-21 16:53:53 +02:00
/*===========================================================================*
Split block/character protocols and libdriver This patch separates the character and block driver communication protocols. The old character protocol remains the same, but a new block protocol is introduced. The libdriver library is replaced by two new libraries: libchardriver and libblockdriver. Their exposed API, and drivers that use them, have been updated accordingly. Together, libbdev and libblockdriver now completely abstract away the message format used by the block protocol. As the memory driver is both a character and a block device driver, it now implements its own message loop. The most important semantic change made to the block protocol is that it is no longer possible to return both partial results and an error for a single transfer. This simplifies the interaction between the caller and the driver, as the I/O vector no longer needs to be copied back. Also, drivers are now no longer supposed to decide based on the layout of the I/O vector when a transfer should be cut short. Put simply, transfers are now supposed to either succeed completely, or result in an error. After this patch, the state of the various pieces is as follows: - block protocol: stable - libbdev API: stable for synchronous communication - libblockdriver API: needs slight revision (the drvlib/partition API in particular; the threading API will also change shortly) - character protocol: needs cleanup - libchardriver API: needs cleanup accordingly - driver restarts: largely unsupported until endpoint changes are reintroduced As a side effect, this patch eliminates several bugs, hacks, and gcc -Wall and -W warnings all over the place. It probably introduces a few new ones, too. Update warning: this patch changes the protocol between MFS and disk drivers, so in order to use old/new images, the MFS from the ramdisk must be used to mount all file systems.
2011-11-22 13:27:53 +01:00
* chardriver_announce *
*===========================================================================*/
2012-03-25 20:25:53 +02:00
void chardriver_announce(void)
Split block/character protocols and libdriver This patch separates the character and block driver communication protocols. The old character protocol remains the same, but a new block protocol is introduced. The libdriver library is replaced by two new libraries: libchardriver and libblockdriver. Their exposed API, and drivers that use them, have been updated accordingly. Together, libbdev and libblockdriver now completely abstract away the message format used by the block protocol. As the memory driver is both a character and a block device driver, it now implements its own message loop. The most important semantic change made to the block protocol is that it is no longer possible to return both partial results and an error for a single transfer. This simplifies the interaction between the caller and the driver, as the I/O vector no longer needs to be copied back. Also, drivers are now no longer supposed to decide based on the layout of the I/O vector when a transfer should be cut short. Put simply, transfers are now supposed to either succeed completely, or result in an error. After this patch, the state of the various pieces is as follows: - block protocol: stable - libbdev API: stable for synchronous communication - libblockdriver API: needs slight revision (the drvlib/partition API in particular; the threading API will also change shortly) - character protocol: needs cleanup - libchardriver API: needs cleanup accordingly - driver restarts: largely unsupported until endpoint changes are reintroduced As a side effect, this patch eliminates several bugs, hacks, and gcc -Wall and -W warnings all over the place. It probably introduces a few new ones, too. Update warning: this patch changes the protocol between MFS and disk drivers, so in order to use old/new images, the MFS from the ramdisk must be used to mount all file systems.
2011-11-22 13:27:53 +01:00
{
/* Announce we are up after a fresh start or restart. */
int r;
char key[DS_MAX_KEYLEN];
char label[DS_MAX_KEYLEN];
const char *driver_prefix = "drv.chr.";
Split block/character protocols and libdriver This patch separates the character and block driver communication protocols. The old character protocol remains the same, but a new block protocol is introduced. The libdriver library is replaced by two new libraries: libchardriver and libblockdriver. Their exposed API, and drivers that use them, have been updated accordingly. Together, libbdev and libblockdriver now completely abstract away the message format used by the block protocol. As the memory driver is both a character and a block device driver, it now implements its own message loop. The most important semantic change made to the block protocol is that it is no longer possible to return both partial results and an error for a single transfer. This simplifies the interaction between the caller and the driver, as the I/O vector no longer needs to be copied back. Also, drivers are now no longer supposed to decide based on the layout of the I/O vector when a transfer should be cut short. Put simply, transfers are now supposed to either succeed completely, or result in an error. After this patch, the state of the various pieces is as follows: - block protocol: stable - libbdev API: stable for synchronous communication - libblockdriver API: needs slight revision (the drvlib/partition API in particular; the threading API will also change shortly) - character protocol: needs cleanup - libchardriver API: needs cleanup accordingly - driver restarts: largely unsupported until endpoint changes are reintroduced As a side effect, this patch eliminates several bugs, hacks, and gcc -Wall and -W warnings all over the place. It probably introduces a few new ones, too. Update warning: this patch changes the protocol between MFS and disk drivers, so in order to use old/new images, the MFS from the ramdisk must be used to mount all file systems.
2011-11-22 13:27:53 +01:00
/* Callers are allowed to use ipc_sendrec to communicate with drivers.
Split block/character protocols and libdriver This patch separates the character and block driver communication protocols. The old character protocol remains the same, but a new block protocol is introduced. The libdriver library is replaced by two new libraries: libchardriver and libblockdriver. Their exposed API, and drivers that use them, have been updated accordingly. Together, libbdev and libblockdriver now completely abstract away the message format used by the block protocol. As the memory driver is both a character and a block device driver, it now implements its own message loop. The most important semantic change made to the block protocol is that it is no longer possible to return both partial results and an error for a single transfer. This simplifies the interaction between the caller and the driver, as the I/O vector no longer needs to be copied back. Also, drivers are now no longer supposed to decide based on the layout of the I/O vector when a transfer should be cut short. Put simply, transfers are now supposed to either succeed completely, or result in an error. After this patch, the state of the various pieces is as follows: - block protocol: stable - libbdev API: stable for synchronous communication - libblockdriver API: needs slight revision (the drvlib/partition API in particular; the threading API will also change shortly) - character protocol: needs cleanup - libchardriver API: needs cleanup accordingly - driver restarts: largely unsupported until endpoint changes are reintroduced As a side effect, this patch eliminates several bugs, hacks, and gcc -Wall and -W warnings all over the place. It probably introduces a few new ones, too. Update warning: this patch changes the protocol between MFS and disk drivers, so in order to use old/new images, the MFS from the ramdisk must be used to mount all file systems.
2011-11-22 13:27:53 +01:00
* For this reason, there may blocked callers when a driver restarts.
* Ask the kernel to unblock them (if any).
*/
if ((r = sys_statectl(SYS_STATE_CLEAR_IPC_REFS, 0, 0)) != OK)
Input infrastructure, INPUT server, PCKBD driver This commit separates the low-level keyboard driver from TTY, putting it in a separate driver (PCKBD). The commit also separates management of raw input devices from TTY, and puts it in a separate server (INPUT). All keyboard and mouse input from hardware is sent by drivers to the INPUT server, which either sends it to a process that has opened a raw input device, or otherwise forwards it to TTY for standard processing. Design by Dirk Vogt. Prototype by Uli Kastlunger. Additional changes made to the prototype: - the event communication is now based on USB HID codes; all input drivers have to use USB codes to describe events; - all TTY keymaps have been converted to USB format, with the effect that a single keymap covers all keys; there is no (static) escaped keymap anymore; - further keymap tweaks now allow remapping of literally all keys; - input device renumbering and protocol rewrite; - INPUT server rewrite, with added support for cancel and select; - PCKBD reimplementation, including PC/AT-to-USB translation; - support for manipulating keyboard LEDs has been added; - keyboard and mouse multiplexer devices have been added to INPUT, primarily so that an X server need only open two devices; - a new "libinputdriver" library abstracts away protocol details from input drivers, and should be used by all future input drivers; - both INPUT and PCKBD can be restarted; - TTY is now scheduled by KERNEL, so that it won't be punished for running a lot; without this, simply running "yes" on the console kills the system; - the KIOCBELL IOCTL has been moved to /dev/console; - support for the SCANCODES termios setting has been removed; - obsolete keymap compression has been removed; - the obsolete Olivetti M24 keymap has been removed. Change-Id: I3a672fb8c4fd566734e4b46d3994b4b7fc96d578
2013-09-28 14:46:21 +02:00
panic("chardriver_announce: sys_statectl failed: %d", r);
Split block/character protocols and libdriver This patch separates the character and block driver communication protocols. The old character protocol remains the same, but a new block protocol is introduced. The libdriver library is replaced by two new libraries: libchardriver and libblockdriver. Their exposed API, and drivers that use them, have been updated accordingly. Together, libbdev and libblockdriver now completely abstract away the message format used by the block protocol. As the memory driver is both a character and a block device driver, it now implements its own message loop. The most important semantic change made to the block protocol is that it is no longer possible to return both partial results and an error for a single transfer. This simplifies the interaction between the caller and the driver, as the I/O vector no longer needs to be copied back. Also, drivers are now no longer supposed to decide based on the layout of the I/O vector when a transfer should be cut short. Put simply, transfers are now supposed to either succeed completely, or result in an error. After this patch, the state of the various pieces is as follows: - block protocol: stable - libbdev API: stable for synchronous communication - libblockdriver API: needs slight revision (the drvlib/partition API in particular; the threading API will also change shortly) - character protocol: needs cleanup - libchardriver API: needs cleanup accordingly - driver restarts: largely unsupported until endpoint changes are reintroduced As a side effect, this patch eliminates several bugs, hacks, and gcc -Wall and -W warnings all over the place. It probably introduces a few new ones, too. Update warning: this patch changes the protocol between MFS and disk drivers, so in order to use old/new images, the MFS from the ramdisk must be used to mount all file systems.
2011-11-22 13:27:53 +01:00
/* Publish a driver up event. */
if ((r = ds_retrieve_label_name(label, sef_self())) != OK)
Input infrastructure, INPUT server, PCKBD driver This commit separates the low-level keyboard driver from TTY, putting it in a separate driver (PCKBD). The commit also separates management of raw input devices from TTY, and puts it in a separate server (INPUT). All keyboard and mouse input from hardware is sent by drivers to the INPUT server, which either sends it to a process that has opened a raw input device, or otherwise forwards it to TTY for standard processing. Design by Dirk Vogt. Prototype by Uli Kastlunger. Additional changes made to the prototype: - the event communication is now based on USB HID codes; all input drivers have to use USB codes to describe events; - all TTY keymaps have been converted to USB format, with the effect that a single keymap covers all keys; there is no (static) escaped keymap anymore; - further keymap tweaks now allow remapping of literally all keys; - input device renumbering and protocol rewrite; - INPUT server rewrite, with added support for cancel and select; - PCKBD reimplementation, including PC/AT-to-USB translation; - support for manipulating keyboard LEDs has been added; - keyboard and mouse multiplexer devices have been added to INPUT, primarily so that an X server need only open two devices; - a new "libinputdriver" library abstracts away protocol details from input drivers, and should be used by all future input drivers; - both INPUT and PCKBD can be restarted; - TTY is now scheduled by KERNEL, so that it won't be punished for running a lot; without this, simply running "yes" on the console kills the system; - the KIOCBELL IOCTL has been moved to /dev/console; - support for the SCANCODES termios setting has been removed; - obsolete keymap compression has been removed; - the obsolete Olivetti M24 keymap has been removed. Change-Id: I3a672fb8c4fd566734e4b46d3994b4b7fc96d578
2013-09-28 14:46:21 +02:00
panic("chardriver_announce: unable to get own label: %d", r);
Split block/character protocols and libdriver This patch separates the character and block driver communication protocols. The old character protocol remains the same, but a new block protocol is introduced. The libdriver library is replaced by two new libraries: libchardriver and libblockdriver. Their exposed API, and drivers that use them, have been updated accordingly. Together, libbdev and libblockdriver now completely abstract away the message format used by the block protocol. As the memory driver is both a character and a block device driver, it now implements its own message loop. The most important semantic change made to the block protocol is that it is no longer possible to return both partial results and an error for a single transfer. This simplifies the interaction between the caller and the driver, as the I/O vector no longer needs to be copied back. Also, drivers are now no longer supposed to decide based on the layout of the I/O vector when a transfer should be cut short. Put simply, transfers are now supposed to either succeed completely, or result in an error. After this patch, the state of the various pieces is as follows: - block protocol: stable - libbdev API: stable for synchronous communication - libblockdriver API: needs slight revision (the drvlib/partition API in particular; the threading API will also change shortly) - character protocol: needs cleanup - libchardriver API: needs cleanup accordingly - driver restarts: largely unsupported until endpoint changes are reintroduced As a side effect, this patch eliminates several bugs, hacks, and gcc -Wall and -W warnings all over the place. It probably introduces a few new ones, too. Update warning: this patch changes the protocol between MFS and disk drivers, so in order to use old/new images, the MFS from the ramdisk must be used to mount all file systems.
2011-11-22 13:27:53 +01:00
snprintf(key, DS_MAX_KEYLEN, "%s%s", driver_prefix, label);
if ((r = ds_publish_u32(key, DS_DRIVER_UP, DSF_OVERWRITE)) != OK)
Input infrastructure, INPUT server, PCKBD driver This commit separates the low-level keyboard driver from TTY, putting it in a separate driver (PCKBD). The commit also separates management of raw input devices from TTY, and puts it in a separate server (INPUT). All keyboard and mouse input from hardware is sent by drivers to the INPUT server, which either sends it to a process that has opened a raw input device, or otherwise forwards it to TTY for standard processing. Design by Dirk Vogt. Prototype by Uli Kastlunger. Additional changes made to the prototype: - the event communication is now based on USB HID codes; all input drivers have to use USB codes to describe events; - all TTY keymaps have been converted to USB format, with the effect that a single keymap covers all keys; there is no (static) escaped keymap anymore; - further keymap tweaks now allow remapping of literally all keys; - input device renumbering and protocol rewrite; - INPUT server rewrite, with added support for cancel and select; - PCKBD reimplementation, including PC/AT-to-USB translation; - support for manipulating keyboard LEDs has been added; - keyboard and mouse multiplexer devices have been added to INPUT, primarily so that an X server need only open two devices; - a new "libinputdriver" library abstracts away protocol details from input drivers, and should be used by all future input drivers; - both INPUT and PCKBD can be restarted; - TTY is now scheduled by KERNEL, so that it won't be punished for running a lot; without this, simply running "yes" on the console kills the system; - the KIOCBELL IOCTL has been moved to /dev/console; - support for the SCANCODES termios setting has been removed; - obsolete keymap compression has been removed; - the obsolete Olivetti M24 keymap has been removed. Change-Id: I3a672fb8c4fd566734e4b46d3994b4b7fc96d578
2013-09-28 14:46:21 +02:00
panic("chardriver_announce: unable to publish driver up event: %d", r);
Split block/character protocols and libdriver This patch separates the character and block driver communication protocols. The old character protocol remains the same, but a new block protocol is introduced. The libdriver library is replaced by two new libraries: libchardriver and libblockdriver. Their exposed API, and drivers that use them, have been updated accordingly. Together, libbdev and libblockdriver now completely abstract away the message format used by the block protocol. As the memory driver is both a character and a block device driver, it now implements its own message loop. The most important semantic change made to the block protocol is that it is no longer possible to return both partial results and an error for a single transfer. This simplifies the interaction between the caller and the driver, as the I/O vector no longer needs to be copied back. Also, drivers are now no longer supposed to decide based on the layout of the I/O vector when a transfer should be cut short. Put simply, transfers are now supposed to either succeed completely, or result in an error. After this patch, the state of the various pieces is as follows: - block protocol: stable - libbdev API: stable for synchronous communication - libblockdriver API: needs slight revision (the drvlib/partition API in particular; the threading API will also change shortly) - character protocol: needs cleanup - libchardriver API: needs cleanup accordingly - driver restarts: largely unsupported until endpoint changes are reintroduced As a side effect, this patch eliminates several bugs, hacks, and gcc -Wall and -W warnings all over the place. It probably introduces a few new ones, too. Update warning: this patch changes the protocol between MFS and disk drivers, so in order to use old/new images, the MFS from the ramdisk must be used to mount all file systems.
2011-11-22 13:27:53 +01:00
/* Expect an open for any device before serving regular driver requests. */
Split block/character protocols and libdriver This patch separates the character and block driver communication protocols. The old character protocol remains the same, but a new block protocol is introduced. The libdriver library is replaced by two new libraries: libchardriver and libblockdriver. Their exposed API, and drivers that use them, have been updated accordingly. Together, libbdev and libblockdriver now completely abstract away the message format used by the block protocol. As the memory driver is both a character and a block device driver, it now implements its own message loop. The most important semantic change made to the block protocol is that it is no longer possible to return both partial results and an error for a single transfer. This simplifies the interaction between the caller and the driver, as the I/O vector no longer needs to be copied back. Also, drivers are now no longer supposed to decide based on the layout of the I/O vector when a transfer should be cut short. Put simply, transfers are now supposed to either succeed completely, or result in an error. After this patch, the state of the various pieces is as follows: - block protocol: stable - libbdev API: stable for synchronous communication - libblockdriver API: needs slight revision (the drvlib/partition API in particular; the threading API will also change shortly) - character protocol: needs cleanup - libchardriver API: needs cleanup accordingly - driver restarts: largely unsupported until endpoint changes are reintroduced As a side effect, this patch eliminates several bugs, hacks, and gcc -Wall and -W warnings all over the place. It probably introduces a few new ones, too. Update warning: this patch changes the protocol between MFS and disk drivers, so in order to use old/new images, the MFS from the ramdisk must be used to mount all file systems.
2011-11-22 13:27:53 +01:00
clear_open_devs();
}
/*===========================================================================*
* chardriver_reply_task *
*===========================================================================*/
void chardriver_reply_task(endpoint_t endpt, cdev_id_t id, int r)
{
/* Reply to a (read, write, ioctl) task request that was suspended earlier.
* Not-so-well-written drivers may use this function to send a reply to a
* request that is being processed right now, and then return EDONTREPLY later.
*/
message m_reply;
if (r == EDONTREPLY || r == SUSPEND)
panic("chardriver: bad task reply: %d", r);
memset(&m_reply, 0, sizeof(m_reply));
m_reply.m_type = CDEV_REPLY;
m_reply.m_lchardriver_vfs_reply.status = r;
m_reply.m_lchardriver_vfs_reply.id = id;
if ((r = asynsend3(endpt, &m_reply, AMF_NOREPLY)) != OK)
printf("chardriver_reply_task: send to %d failed: %d\n", endpt, r);
}
/*===========================================================================*
* chardriver_reply_select *
*===========================================================================*/
void chardriver_reply_select(endpoint_t endpt, devminor_t minor, int r)
{
/* Reply to a select request with a status update. This must not be used to
* reply to a select request that is being processed right now.
*/
message m_reply;
/* Replying with an error is allowed (if unusual). */
if (r == EDONTREPLY || r == SUSPEND)
panic("chardriver: bad select reply: %d", r);
memset(&m_reply, 0, sizeof(m_reply));
m_reply.m_type = CDEV_SEL2_REPLY;
m_reply.m_lchardriver_vfs_sel2.minor = minor;
m_reply.m_lchardriver_vfs_sel2.status = r;
if ((r = asynsend3(endpt, &m_reply, AMF_NOREPLY)) != OK)
printf("chardriver_reply_select: send to %d failed: %d\n", endpt, r);
}
Split block/character protocols and libdriver This patch separates the character and block driver communication protocols. The old character protocol remains the same, but a new block protocol is introduced. The libdriver library is replaced by two new libraries: libchardriver and libblockdriver. Their exposed API, and drivers that use them, have been updated accordingly. Together, libbdev and libblockdriver now completely abstract away the message format used by the block protocol. As the memory driver is both a character and a block device driver, it now implements its own message loop. The most important semantic change made to the block protocol is that it is no longer possible to return both partial results and an error for a single transfer. This simplifies the interaction between the caller and the driver, as the I/O vector no longer needs to be copied back. Also, drivers are now no longer supposed to decide based on the layout of the I/O vector when a transfer should be cut short. Put simply, transfers are now supposed to either succeed completely, or result in an error. After this patch, the state of the various pieces is as follows: - block protocol: stable - libbdev API: stable for synchronous communication - libblockdriver API: needs slight revision (the drvlib/partition API in particular; the threading API will also change shortly) - character protocol: needs cleanup - libchardriver API: needs cleanup accordingly - driver restarts: largely unsupported until endpoint changes are reintroduced As a side effect, this patch eliminates several bugs, hacks, and gcc -Wall and -W warnings all over the place. It probably introduces a few new ones, too. Update warning: this patch changes the protocol between MFS and disk drivers, so in order to use old/new images, the MFS from the ramdisk must be used to mount all file systems.
2011-11-22 13:27:53 +01:00
/*===========================================================================*
* send_reply *
*===========================================================================*/
static void send_reply(endpoint_t endpt, message *m_ptr, int ipc_status)
{
/* Send a reply message to a request. */
int r;
/* If we would block sending the message, send it asynchronously. */
if (IPC_STATUS_CALL(ipc_status) == SENDREC)
r = ipc_sendnb(endpt, m_ptr);
else
r = asynsend3(endpt, m_ptr, AMF_NOREPLY);
if (r != OK)
printf("chardriver: unable to send reply to %d: %d\n", endpt, r);
}
/*===========================================================================*
* chardriver_reply *
*===========================================================================*/
static void chardriver_reply(message *mess, int ipc_status, int r)
{
/* Prepare and send a reply message. */
message reply_mess;
/* If the EDONTREPLY pseudo-reply is given, we do not reply. This is however
* allowed only for blocking task calls. Perform a sanity check.
*/
if (r == EDONTREPLY) {
switch (mess->m_type) {
case CDEV_READ:
case CDEV_WRITE:
case CDEV_IOCTL:
/* FIXME: we should be able to check FLAGS against
* CDEV_NONBLOCK here, but in practice, several drivers do not
* send a reply through this path (eg TTY) or simply do not
* implement nonblocking calls properly (eg audio, LWIP).
*/
#if 0
if (mess->m_vfs_lchardriver_readwrite.flags & CDEV_NONBLOCK)
panic("chardriver: cannot suspend nonblocking I/O");
#endif
/*fall-through*/
case CDEV_CANCEL:
return; /* alright */
default:
panic("chardriver: cannot suspend request %d", mess->m_type);
}
}
if (r == SUSPEND)
panic("chardriver: SUSPEND should not be used anymore");
/* Do not reply with ERESTART. The only possible caller, VFS, will find out
* through other means when we have restarted, and is not (fully) ready to
* deal with ERESTART errors.
*/
if (r == ERESTART)
return;
memset(&reply_mess, 0, sizeof(reply_mess));
switch (mess->m_type) {
case CDEV_OPEN:
case CDEV_CLOSE:
reply_mess.m_type = CDEV_REPLY;
reply_mess.m_lchardriver_vfs_reply.status = r;
reply_mess.m_lchardriver_vfs_reply.id =
mess->m_vfs_lchardriver_openclose.id;
break;
case CDEV_READ:
case CDEV_WRITE:
case CDEV_IOCTL:
reply_mess.m_type = CDEV_REPLY;
reply_mess.m_lchardriver_vfs_reply.status = r;
reply_mess.m_lchardriver_vfs_reply.id =
mess->m_vfs_lchardriver_readwrite.id;
break;
case CDEV_CANCEL: /* For cancel, this is a reply to the original request! */
reply_mess.m_type = CDEV_REPLY;
reply_mess.m_lchardriver_vfs_reply.status = r;
reply_mess.m_lchardriver_vfs_reply.id =
mess->m_vfs_lchardriver_cancel.id;
break;
case CDEV_SELECT:
reply_mess.m_type = CDEV_SEL1_REPLY;
reply_mess.m_lchardriver_vfs_sel1.status = r;
reply_mess.m_lchardriver_vfs_sel1.minor =
mess->m_vfs_lchardriver_select.minor;
break;
default:
panic("chardriver: unknown request %d", mess->m_type);
}
send_reply(mess->m_source, &reply_mess, ipc_status);
}
/*===========================================================================*
* do_open *
*===========================================================================*/
static int do_open(struct chardriver *cdp, message *m_ptr)
{
/* Open a minor device. */
endpoint_t user_endpt;
devminor_t minor;
int r, bits;
/* Default action if no open hook is in place. */
if (cdp->cdr_open == NULL)
return OK;
/* Call the open hook. */
minor = m_ptr->m_vfs_lchardriver_openclose.minor;
bits = m_ptr->m_vfs_lchardriver_openclose.access;
user_endpt = m_ptr->m_vfs_lchardriver_openclose.user;
r = cdp->cdr_open(minor, bits, user_endpt);
/* If the device has been cloned, mark the new minor as open too. */
if (r >= 0 && (r & CDEV_CLONED)) {
minor = r & ~(CDEV_CLONED | CDEV_CTTY);
if (!is_open_dev(minor))
set_open_dev(minor);
}
return r;
2005-04-21 16:53:53 +02:00
}
/*===========================================================================*
* do_close *
2005-04-21 16:53:53 +02:00
*===========================================================================*/
static int do_close(struct chardriver *cdp, message *m_ptr)
2005-04-21 16:53:53 +02:00
{
/* Close a minor device. */
devminor_t minor;
2005-04-21 16:53:53 +02:00
/* Default action if no close hook is in place. */
if (cdp->cdr_close == NULL)
return OK;
2005-04-21 16:53:53 +02:00
/* Call the close hook. */
minor = m_ptr->m_vfs_lchardriver_openclose.minor;
2005-04-21 16:53:53 +02:00
return cdp->cdr_close(minor);
}
/*===========================================================================*
* do_trasnfer *
*===========================================================================*/
static int do_transfer(struct chardriver *cdp, message *m_ptr, int do_write)
{
/* Carry out a read or write task request. */
devminor_t minor;
off_t position;
endpoint_t endpt;
cp_grant_id_t grant;
size_t size;
int flags;
cdev_id_t id;
ssize_t r;
minor = m_ptr->m_vfs_lchardriver_readwrite.minor;
position = m_ptr->m_vfs_lchardriver_readwrite.pos;
endpt = m_ptr->m_source;
grant = m_ptr->m_vfs_lchardriver_readwrite.grant;
size = m_ptr->m_vfs_lchardriver_readwrite.count;
flags = m_ptr->m_vfs_lchardriver_readwrite.flags;
id = m_ptr->m_vfs_lchardriver_readwrite.id;
/* Call the read/write hook, if the appropriate one is in place. */
if (!do_write && cdp->cdr_read != NULL)
r = cdp->cdr_read(minor, position, endpt, grant, size, flags, id);
else if (do_write && cdp->cdr_write != NULL)
r = cdp->cdr_write(minor, position, endpt, grant, size, flags, id);
else
r = EIO; /* Default action if no read/write hook is in place. */
2005-04-21 16:53:53 +02:00
return r;
}
2005-04-21 16:53:53 +02:00
/*===========================================================================*
* do_ioctl *
*===========================================================================*/
static int do_ioctl(struct chardriver *cdp, message *m_ptr)
{
/* Carry out an I/O control task request. */
devminor_t minor;
unsigned long request;
cp_grant_id_t grant;
endpoint_t endpt, user_endpt;
int flags;
cdev_id_t id;
/* Default action if no ioctl hook is in place. */
if (cdp->cdr_ioctl == NULL)
return ENOTTY;
/* Call the ioctl hook. */
minor = m_ptr->m_vfs_lchardriver_readwrite.minor;
request = m_ptr->m_vfs_lchardriver_readwrite.request;
endpt = m_ptr->m_source;
grant = m_ptr->m_vfs_lchardriver_readwrite.grant;
flags = m_ptr->m_vfs_lchardriver_readwrite.flags;
user_endpt = m_ptr->m_vfs_lchardriver_readwrite.user;
id = m_ptr->m_vfs_lchardriver_readwrite.id;
return cdp->cdr_ioctl(minor, request, endpt, grant, flags, user_endpt, id);
2005-04-21 16:53:53 +02:00
}
/*===========================================================================*
* do_cancel *
*===========================================================================*/
static int do_cancel(struct chardriver *cdp, message *m_ptr)
2005-04-21 16:53:53 +02:00
{
/* Cancel a suspended (read, write, ioctl) task request. The original request
* may already have finished, in which case no reply should be sent.
2005-04-21 16:53:53 +02:00
*/
devminor_t minor;
endpoint_t endpt;
cdev_id_t id;
2005-04-21 16:53:53 +02:00
/* Default action if no cancel hook is in place: let the request finish. */
if (cdp->cdr_cancel == NULL)
return EDONTREPLY;
2005-04-21 16:53:53 +02:00
/* Call the cancel hook. */
minor = m_ptr->m_vfs_lchardriver_cancel.minor;
endpt = m_ptr->m_source;
id = m_ptr->m_vfs_lchardriver_cancel.id;
return cdp->cdr_cancel(minor, endpt, id);
}
2005-04-21 16:53:53 +02:00
/*===========================================================================*
* do_select *
*===========================================================================*/
static int do_select(struct chardriver *cdp, message *m_ptr)
{
/* Perform a select query on a minor device. */
devminor_t minor;
unsigned int ops;
endpoint_t endpt;
2005-04-21 16:53:53 +02:00
/* Default action if no select hook is in place. */
if (cdp->cdr_select == NULL)
return EBADF;
2005-04-21 16:53:53 +02:00
/* Call the select hook. */
minor = m_ptr->m_vfs_lchardriver_select.minor;
ops = m_ptr->m_vfs_lchardriver_select.ops;
endpt = m_ptr->m_source;
return cdp->cdr_select(minor, ops, endpt);
2005-04-21 16:53:53 +02:00
}
/*===========================================================================*
* do_block_open *
*===========================================================================*/
static void do_block_open(message *m_ptr, int ipc_status)
{
/* Reply to a block driver open request stating there is no such device. */
message m_reply;
memset(&m_reply, 0, sizeof(m_reply));
m_reply.m_type = BDEV_REPLY;
m_reply.m_lblockdriver_lbdev_reply.status = ENXIO;
m_reply.m_lblockdriver_lbdev_reply.id = m_ptr->m_lbdev_lblockdriver_msg.id;
send_reply(m_ptr->m_source, &m_reply, ipc_status);
}
/*===========================================================================*
* chardriver_process *
*===========================================================================*/
void chardriver_process(struct chardriver *cdp, message *m_ptr, int ipc_status)
{
/* Call the appropiate driver function, based on the type of request. Send a
* reply to the caller if necessary.
*/
int r;
/* Check for notifications first. We never reply to notifications. */
if (is_ipc_notify(ipc_status)) {
switch (_ENDPOINT_P(m_ptr->m_source)) {
case HARDWARE:
if (cdp->cdr_intr)
cdp->cdr_intr(m_ptr->m_notify.interrupts);
break;
case CLOCK:
if (cdp->cdr_alarm)
cdp->cdr_alarm(m_ptr->m_notify.timestamp);
break;
default:
if (cdp->cdr_other)
cdp->cdr_other(m_ptr, ipc_status);
}
return; /* do not send a reply */
}
/* Reply to block driver open requests with an error code. Otherwise, if
* someone creates a block device node for a character driver, opening that
* device node will cause the corresponding VFS thread to block forever.
*/
if (m_ptr->m_type == BDEV_OPEN) {
do_block_open(m_ptr, ipc_status);
return;
}
if (IS_CDEV_RQ(m_ptr->m_type)) {
int minor;
/* Try to retrieve minor device number */
r = chardriver_get_minor(m_ptr, &minor);
if (OK != r)
return;
/* We might get spurious requests if the driver has been restarted.
* Deny any requests on devices that have not previously been opened.
*/
if (!is_open_dev(minor)) {
/* Ignore spurious requests for unopened devices. */
if (m_ptr->m_type != CDEV_OPEN)
return; /* do not send a reply */
/* Mark the device as opened otherwise. */
set_open_dev(minor);
}
}
/* Call the appropriate function(s) for this request. */
switch (m_ptr->m_type) {
case CDEV_OPEN: r = do_open(cdp, m_ptr); break;
case CDEV_CLOSE: r = do_close(cdp, m_ptr); break;
case CDEV_READ: r = do_transfer(cdp, m_ptr, FALSE); break;
case CDEV_WRITE: r = do_transfer(cdp, m_ptr, TRUE); break;
case CDEV_IOCTL: r = do_ioctl(cdp, m_ptr); break;
case CDEV_CANCEL: r = do_cancel(cdp, m_ptr); break;
case CDEV_SELECT: r = do_select(cdp, m_ptr); break;
default:
Split block/character protocols and libdriver This patch separates the character and block driver communication protocols. The old character protocol remains the same, but a new block protocol is introduced. The libdriver library is replaced by two new libraries: libchardriver and libblockdriver. Their exposed API, and drivers that use them, have been updated accordingly. Together, libbdev and libblockdriver now completely abstract away the message format used by the block protocol. As the memory driver is both a character and a block device driver, it now implements its own message loop. The most important semantic change made to the block protocol is that it is no longer possible to return both partial results and an error for a single transfer. This simplifies the interaction between the caller and the driver, as the I/O vector no longer needs to be copied back. Also, drivers are now no longer supposed to decide based on the layout of the I/O vector when a transfer should be cut short. Put simply, transfers are now supposed to either succeed completely, or result in an error. After this patch, the state of the various pieces is as follows: - block protocol: stable - libbdev API: stable for synchronous communication - libblockdriver API: needs slight revision (the drvlib/partition API in particular; the threading API will also change shortly) - character protocol: needs cleanup - libchardriver API: needs cleanup accordingly - driver restarts: largely unsupported until endpoint changes are reintroduced As a side effect, this patch eliminates several bugs, hacks, and gcc -Wall and -W warnings all over the place. It probably introduces a few new ones, too. Update warning: this patch changes the protocol between MFS and disk drivers, so in order to use old/new images, the MFS from the ramdisk must be used to mount all file systems.
2011-11-22 13:27:53 +01:00
if (cdp->cdr_other)
cdp->cdr_other(m_ptr, ipc_status);
return; /* do not send a reply */
}
chardriver_reply(m_ptr, ipc_status, r);
}
2005-04-21 16:53:53 +02:00
/*===========================================================================*
* chardriver_terminate *
2005-04-21 16:53:53 +02:00
*===========================================================================*/
void chardriver_terminate(void)
2005-04-21 16:53:53 +02:00
{
/* Break out of the main loop after finishing the current request. */
Split block/character protocols and libdriver This patch separates the character and block driver communication protocols. The old character protocol remains the same, but a new block protocol is introduced. The libdriver library is replaced by two new libraries: libchardriver and libblockdriver. Their exposed API, and drivers that use them, have been updated accordingly. Together, libbdev and libblockdriver now completely abstract away the message format used by the block protocol. As the memory driver is both a character and a block device driver, it now implements its own message loop. The most important semantic change made to the block protocol is that it is no longer possible to return both partial results and an error for a single transfer. This simplifies the interaction between the caller and the driver, as the I/O vector no longer needs to be copied back. Also, drivers are now no longer supposed to decide based on the layout of the I/O vector when a transfer should be cut short. Put simply, transfers are now supposed to either succeed completely, or result in an error. After this patch, the state of the various pieces is as follows: - block protocol: stable - libbdev API: stable for synchronous communication - libblockdriver API: needs slight revision (the drvlib/partition API in particular; the threading API will also change shortly) - character protocol: needs cleanup - libchardriver API: needs cleanup accordingly - driver restarts: largely unsupported until endpoint changes are reintroduced As a side effect, this patch eliminates several bugs, hacks, and gcc -Wall and -W warnings all over the place. It probably introduces a few new ones, too. Update warning: this patch changes the protocol between MFS and disk drivers, so in order to use old/new images, the MFS from the ramdisk must be used to mount all file systems.
2011-11-22 13:27:53 +01:00
running = FALSE;
Split block/character protocols and libdriver This patch separates the character and block driver communication protocols. The old character protocol remains the same, but a new block protocol is introduced. The libdriver library is replaced by two new libraries: libchardriver and libblockdriver. Their exposed API, and drivers that use them, have been updated accordingly. Together, libbdev and libblockdriver now completely abstract away the message format used by the block protocol. As the memory driver is both a character and a block device driver, it now implements its own message loop. The most important semantic change made to the block protocol is that it is no longer possible to return both partial results and an error for a single transfer. This simplifies the interaction between the caller and the driver, as the I/O vector no longer needs to be copied back. Also, drivers are now no longer supposed to decide based on the layout of the I/O vector when a transfer should be cut short. Put simply, transfers are now supposed to either succeed completely, or result in an error. After this patch, the state of the various pieces is as follows: - block protocol: stable - libbdev API: stable for synchronous communication - libblockdriver API: needs slight revision (the drvlib/partition API in particular; the threading API will also change shortly) - character protocol: needs cleanup - libchardriver API: needs cleanup accordingly - driver restarts: largely unsupported until endpoint changes are reintroduced As a side effect, this patch eliminates several bugs, hacks, and gcc -Wall and -W warnings all over the place. It probably introduces a few new ones, too. Update warning: this patch changes the protocol between MFS and disk drivers, so in order to use old/new images, the MFS from the ramdisk must be used to mount all file systems.
2011-11-22 13:27:53 +01:00
sef_cancel();
2005-04-21 16:53:53 +02:00
}
/*===========================================================================*
Split block/character protocols and libdriver This patch separates the character and block driver communication protocols. The old character protocol remains the same, but a new block protocol is introduced. The libdriver library is replaced by two new libraries: libchardriver and libblockdriver. Their exposed API, and drivers that use them, have been updated accordingly. Together, libbdev and libblockdriver now completely abstract away the message format used by the block protocol. As the memory driver is both a character and a block device driver, it now implements its own message loop. The most important semantic change made to the block protocol is that it is no longer possible to return both partial results and an error for a single transfer. This simplifies the interaction between the caller and the driver, as the I/O vector no longer needs to be copied back. Also, drivers are now no longer supposed to decide based on the layout of the I/O vector when a transfer should be cut short. Put simply, transfers are now supposed to either succeed completely, or result in an error. After this patch, the state of the various pieces is as follows: - block protocol: stable - libbdev API: stable for synchronous communication - libblockdriver API: needs slight revision (the drvlib/partition API in particular; the threading API will also change shortly) - character protocol: needs cleanup - libchardriver API: needs cleanup accordingly - driver restarts: largely unsupported until endpoint changes are reintroduced As a side effect, this patch eliminates several bugs, hacks, and gcc -Wall and -W warnings all over the place. It probably introduces a few new ones, too. Update warning: this patch changes the protocol between MFS and disk drivers, so in order to use old/new images, the MFS from the ramdisk must be used to mount all file systems.
2011-11-22 13:27:53 +01:00
* chardriver_task *
*===========================================================================*/
void chardriver_task(struct chardriver *cdp)
2005-04-21 16:53:53 +02:00
{
Input infrastructure, INPUT server, PCKBD driver This commit separates the low-level keyboard driver from TTY, putting it in a separate driver (PCKBD). The commit also separates management of raw input devices from TTY, and puts it in a separate server (INPUT). All keyboard and mouse input from hardware is sent by drivers to the INPUT server, which either sends it to a process that has opened a raw input device, or otherwise forwards it to TTY for standard processing. Design by Dirk Vogt. Prototype by Uli Kastlunger. Additional changes made to the prototype: - the event communication is now based on USB HID codes; all input drivers have to use USB codes to describe events; - all TTY keymaps have been converted to USB format, with the effect that a single keymap covers all keys; there is no (static) escaped keymap anymore; - further keymap tweaks now allow remapping of literally all keys; - input device renumbering and protocol rewrite; - INPUT server rewrite, with added support for cancel and select; - PCKBD reimplementation, including PC/AT-to-USB translation; - support for manipulating keyboard LEDs has been added; - keyboard and mouse multiplexer devices have been added to INPUT, primarily so that an X server need only open two devices; - a new "libinputdriver" library abstracts away protocol details from input drivers, and should be used by all future input drivers; - both INPUT and PCKBD can be restarted; - TTY is now scheduled by KERNEL, so that it won't be punished for running a lot; without this, simply running "yes" on the console kills the system; - the KIOCBELL IOCTL has been moved to /dev/console; - support for the SCANCODES termios setting has been removed; - obsolete keymap compression has been removed; - the obsolete Olivetti M24 keymap has been removed. Change-Id: I3a672fb8c4fd566734e4b46d3994b4b7fc96d578
2013-09-28 14:46:21 +02:00
/* Main program of any character device driver task. */
Split block/character protocols and libdriver This patch separates the character and block driver communication protocols. The old character protocol remains the same, but a new block protocol is introduced. The libdriver library is replaced by two new libraries: libchardriver and libblockdriver. Their exposed API, and drivers that use them, have been updated accordingly. Together, libbdev and libblockdriver now completely abstract away the message format used by the block protocol. As the memory driver is both a character and a block device driver, it now implements its own message loop. The most important semantic change made to the block protocol is that it is no longer possible to return both partial results and an error for a single transfer. This simplifies the interaction between the caller and the driver, as the I/O vector no longer needs to be copied back. Also, drivers are now no longer supposed to decide based on the layout of the I/O vector when a transfer should be cut short. Put simply, transfers are now supposed to either succeed completely, or result in an error. After this patch, the state of the various pieces is as follows: - block protocol: stable - libbdev API: stable for synchronous communication - libblockdriver API: needs slight revision (the drvlib/partition API in particular; the threading API will also change shortly) - character protocol: needs cleanup - libchardriver API: needs cleanup accordingly - driver restarts: largely unsupported until endpoint changes are reintroduced As a side effect, this patch eliminates several bugs, hacks, and gcc -Wall and -W warnings all over the place. It probably introduces a few new ones, too. Update warning: this patch changes the protocol between MFS and disk drivers, so in order to use old/new images, the MFS from the ramdisk must be used to mount all file systems.
2011-11-22 13:27:53 +01:00
int r, ipc_status;
message mess;
running = TRUE;
Input infrastructure, INPUT server, PCKBD driver This commit separates the low-level keyboard driver from TTY, putting it in a separate driver (PCKBD). The commit also separates management of raw input devices from TTY, and puts it in a separate server (INPUT). All keyboard and mouse input from hardware is sent by drivers to the INPUT server, which either sends it to a process that has opened a raw input device, or otherwise forwards it to TTY for standard processing. Design by Dirk Vogt. Prototype by Uli Kastlunger. Additional changes made to the prototype: - the event communication is now based on USB HID codes; all input drivers have to use USB codes to describe events; - all TTY keymaps have been converted to USB format, with the effect that a single keymap covers all keys; there is no (static) escaped keymap anymore; - further keymap tweaks now allow remapping of literally all keys; - input device renumbering and protocol rewrite; - INPUT server rewrite, with added support for cancel and select; - PCKBD reimplementation, including PC/AT-to-USB translation; - support for manipulating keyboard LEDs has been added; - keyboard and mouse multiplexer devices have been added to INPUT, primarily so that an X server need only open two devices; - a new "libinputdriver" library abstracts away protocol details from input drivers, and should be used by all future input drivers; - both INPUT and PCKBD can be restarted; - TTY is now scheduled by KERNEL, so that it won't be punished for running a lot; without this, simply running "yes" on the console kills the system; - the KIOCBELL IOCTL has been moved to /dev/console; - support for the SCANCODES termios setting has been removed; - obsolete keymap compression has been removed; - the obsolete Olivetti M24 keymap has been removed. Change-Id: I3a672fb8c4fd566734e4b46d3994b4b7fc96d578
2013-09-28 14:46:21 +02:00
/* Here is the main loop of the character driver task. It waits for a
* message, carries it out, and sends a reply.
Split block/character protocols and libdriver This patch separates the character and block driver communication protocols. The old character protocol remains the same, but a new block protocol is introduced. The libdriver library is replaced by two new libraries: libchardriver and libblockdriver. Their exposed API, and drivers that use them, have been updated accordingly. Together, libbdev and libblockdriver now completely abstract away the message format used by the block protocol. As the memory driver is both a character and a block device driver, it now implements its own message loop. The most important semantic change made to the block protocol is that it is no longer possible to return both partial results and an error for a single transfer. This simplifies the interaction between the caller and the driver, as the I/O vector no longer needs to be copied back. Also, drivers are now no longer supposed to decide based on the layout of the I/O vector when a transfer should be cut short. Put simply, transfers are now supposed to either succeed completely, or result in an error. After this patch, the state of the various pieces is as follows: - block protocol: stable - libbdev API: stable for synchronous communication - libblockdriver API: needs slight revision (the drvlib/partition API in particular; the threading API will also change shortly) - character protocol: needs cleanup - libchardriver API: needs cleanup accordingly - driver restarts: largely unsupported until endpoint changes are reintroduced As a side effect, this patch eliminates several bugs, hacks, and gcc -Wall and -W warnings all over the place. It probably introduces a few new ones, too. Update warning: this patch changes the protocol between MFS and disk drivers, so in order to use old/new images, the MFS from the ramdisk must be used to mount all file systems.
2011-11-22 13:27:53 +01:00
*/
while (running) {
if ((r = sef_receive_status(ANY, &mess, &ipc_status)) != OK) {
if (r == EINTR && !running)
break;
Input infrastructure, INPUT server, PCKBD driver This commit separates the low-level keyboard driver from TTY, putting it in a separate driver (PCKBD). The commit also separates management of raw input devices from TTY, and puts it in a separate server (INPUT). All keyboard and mouse input from hardware is sent by drivers to the INPUT server, which either sends it to a process that has opened a raw input device, or otherwise forwards it to TTY for standard processing. Design by Dirk Vogt. Prototype by Uli Kastlunger. Additional changes made to the prototype: - the event communication is now based on USB HID codes; all input drivers have to use USB codes to describe events; - all TTY keymaps have been converted to USB format, with the effect that a single keymap covers all keys; there is no (static) escaped keymap anymore; - further keymap tweaks now allow remapping of literally all keys; - input device renumbering and protocol rewrite; - INPUT server rewrite, with added support for cancel and select; - PCKBD reimplementation, including PC/AT-to-USB translation; - support for manipulating keyboard LEDs has been added; - keyboard and mouse multiplexer devices have been added to INPUT, primarily so that an X server need only open two devices; - a new "libinputdriver" library abstracts away protocol details from input drivers, and should be used by all future input drivers; - both INPUT and PCKBD can be restarted; - TTY is now scheduled by KERNEL, so that it won't be punished for running a lot; without this, simply running "yes" on the console kills the system; - the KIOCBELL IOCTL has been moved to /dev/console; - support for the SCANCODES termios setting has been removed; - obsolete keymap compression has been removed; - the obsolete Olivetti M24 keymap has been removed. Change-Id: I3a672fb8c4fd566734e4b46d3994b4b7fc96d578
2013-09-28 14:46:21 +02:00
panic("chardriver: sef_receive_status failed: %d", r);
}
Split block/character protocols and libdriver This patch separates the character and block driver communication protocols. The old character protocol remains the same, but a new block protocol is introduced. The libdriver library is replaced by two new libraries: libchardriver and libblockdriver. Their exposed API, and drivers that use them, have been updated accordingly. Together, libbdev and libblockdriver now completely abstract away the message format used by the block protocol. As the memory driver is both a character and a block device driver, it now implements its own message loop. The most important semantic change made to the block protocol is that it is no longer possible to return both partial results and an error for a single transfer. This simplifies the interaction between the caller and the driver, as the I/O vector no longer needs to be copied back. Also, drivers are now no longer supposed to decide based on the layout of the I/O vector when a transfer should be cut short. Put simply, transfers are now supposed to either succeed completely, or result in an error. After this patch, the state of the various pieces is as follows: - block protocol: stable - libbdev API: stable for synchronous communication - libblockdriver API: needs slight revision (the drvlib/partition API in particular; the threading API will also change shortly) - character protocol: needs cleanup - libchardriver API: needs cleanup accordingly - driver restarts: largely unsupported until endpoint changes are reintroduced As a side effect, this patch eliminates several bugs, hacks, and gcc -Wall and -W warnings all over the place. It probably introduces a few new ones, too. Update warning: this patch changes the protocol between MFS and disk drivers, so in order to use old/new images, the MFS from the ramdisk must be used to mount all file systems.
2011-11-22 13:27:53 +01:00
chardriver_process(cdp, &mess, ipc_status);
2005-04-21 16:53:53 +02:00
}
}
/*===========================================================================*
* chardriver_get_minor *
*===========================================================================*/
int chardriver_get_minor(message *m, devminor_t *minor)
{
assert(NULL != m);
assert(NULL != minor);
switch(m->m_type)
{
case CDEV_OPEN:
case CDEV_CLOSE:
*minor = m->m_vfs_lchardriver_openclose.minor;
return OK;
case CDEV_CANCEL:
*minor = m->m_vfs_lchardriver_cancel.minor;
return OK;
case CDEV_SELECT:
*minor = m->m_vfs_lchardriver_select.minor;
return OK;
case CDEV_READ:
case CDEV_WRITE:
case CDEV_IOCTL:
*minor = m->m_vfs_lchardriver_readwrite.minor;
return OK;
default:
return EINVAL;
}
}