minix/servers/inet/mnx_eth.c

948 lines
22 KiB
C
Raw Normal View History

2005-04-21 16:53:53 +02:00
/*
inet/mnx_eth.c
Created: Jan 2, 1992 by Philip Homburg
Copyright 1995 Philip Homburg
*/
#include "inet.h"
#include <minix/ds.h>
2006-07-10 14:55:33 +02:00
#include <minix/safecopies.h>
2005-04-21 16:53:53 +02:00
#include "proto.h"
#include "osdep_eth.h"
#include "generic/type.h"
#include "generic/assert.h"
#include "generic/buf.h"
#include "generic/clock.h"
#include "generic/eth.h"
#include "generic/eth_int.h"
#include "generic/sr.h"
THIS_FILE
static int recv_debug= 0;
2005-04-21 16:53:53 +02:00
FORWARD _PROTOTYPE( void setup_read, (eth_port_t *eth_port) );
FORWARD _PROTOTYPE( void read_int, (eth_port_t *eth_port, int count) );
FORWARD _PROTOTYPE( void eth_issue_send, (eth_port_t *eth_port) );
2005-04-21 16:53:53 +02:00
FORWARD _PROTOTYPE( void write_int, (eth_port_t *eth_port) );
FORWARD _PROTOTYPE( void eth_recvev, (event_t *ev, ev_arg_t ev_arg) );
FORWARD _PROTOTYPE( void eth_sendev, (event_t *ev, ev_arg_t ev_arg) );
FORWARD _PROTOTYPE( eth_port_t *find_port, (message *m) );
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
FORWARD _PROTOTYPE( void eth_restart, (eth_port_t *eth_port,
endpoint_t endpoint) );
FORWARD _PROTOTYPE( void send_getstat, (eth_port_t *eth_port) );
2005-04-21 16:53:53 +02:00
PUBLIC void osdep_eth_init()
{
int i, j, r, rport;
2005-04-21 16:53:53 +02:00
struct eth_conf *ecp;
eth_port_t *eth_port, *rep;
2006-07-10 14:55:33 +02:00
cp_grant_id_t gid;
2005-04-21 16:53:53 +02:00
/* First initialize normal ethernet interfaces */
for (i= 0, ecp= eth_conf, eth_port= eth_port_table;
i<eth_conf_nr; i++, ecp++, eth_port++)
2005-04-21 16:53:53 +02:00
{
2006-07-10 14:55:33 +02:00
/* Set all grants to invalid */
for (j= 0; j<IOVEC_NR; j++)
eth_port->etp_osdep.etp_wr_iovec[j].iov_grant= -1;
eth_port->etp_osdep.etp_wr_vec_grant= -1;
for (j= 0; j<RD_IOVEC; j++)
eth_port->etp_osdep.etp_rd_iovec[j].iov_grant= -1;
eth_port->etp_osdep.etp_rd_vec_grant= -1;
eth_port->etp_osdep.etp_state= OEPS_INIT;
eth_port->etp_osdep.etp_flags= OEPF_EMPTY;
eth_port->etp_osdep.etp_stat_gid= -1;
eth_port->etp_osdep.etp_stat_buf= NULL;
if (eth_is_vlan(ecp))
continue;
2006-07-10 14:55:33 +02:00
/* Allocate grants */
for (j= 0; j<IOVEC_NR; j++)
{
if (cpf_getgrants(&gid, 1) != 1)
{
ip_panic((
"osdep_eth_init: cpf_getgrants failed: %d\n",
errno));
}
eth_port->etp_osdep.etp_wr_iovec[j].iov_grant= gid;
}
if (cpf_getgrants(&gid, 1) != 1)
{
ip_panic((
"osdep_eth_init: cpf_getgrants failed: %d\n",
errno));
}
eth_port->etp_osdep.etp_wr_vec_grant= gid;
for (j= 0; j<RD_IOVEC; j++)
{
if (cpf_getgrants(&gid, 1) != 1)
{
ip_panic((
"osdep_eth_init: cpf_getgrants failed: %d\n",
errno));
}
eth_port->etp_osdep.etp_rd_iovec[j].iov_grant= gid;
}
if (cpf_getgrants(&gid, 1) != 1)
{
ip_panic((
"osdep_eth_init: cpf_getgrants failed: %d\n",
errno));
}
eth_port->etp_osdep.etp_rd_vec_grant= gid;
eth_port->etp_osdep.etp_port= ecp->ec_port;
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
eth_port->etp_osdep.etp_task= ANY;
eth_port->etp_osdep.etp_recvconf= 0;
eth_port->etp_osdep.etp_send_ev= 0;
2005-04-21 16:53:53 +02:00
ev_init(&eth_port->etp_osdep.etp_recvev);
sr_add_minor(if2minor(ecp->ec_ifno, ETH_DEV_OFF),
i, eth_open, eth_close, eth_read,
eth_write, eth_ioctl, eth_cancel, eth_select);
2005-04-21 16:53:53 +02:00
eth_port->etp_flags |= EPF_ENABLED;
eth_port->etp_vlan= 0;
eth_port->etp_vlan_port= NULL;
2005-04-21 16:53:53 +02:00
eth_port->etp_wr_pack= 0;
eth_port->etp_rd_pack= 0;
}
/* And now come the VLANs */
for (i= 0, ecp= eth_conf, eth_port= eth_port_table;
i<eth_conf_nr; i++, ecp++, eth_port++)
{
if (!eth_is_vlan(ecp))
continue;
eth_port->etp_osdep.etp_port= ecp->ec_port;
eth_port->etp_osdep.etp_task= ANY;
ev_init(&eth_port->etp_osdep.etp_recvev);
rport= eth_port->etp_osdep.etp_port;
assert(rport >= 0 && rport < eth_conf_nr);
rep= &eth_port_table[rport];
if (!(rep->etp_flags & EPF_ENABLED))
{
printf(
"eth%d: underlying ethernet device %d not enabled",
i, rport);
continue;
}
if (rep->etp_vlan != 0)
{
printf(
"eth%d: underlying ethernet device %d is a VLAN",
i, rport);
continue;
}
if (rep->etp_flags & EPF_GOT_ADDR)
{
eth_port->etp_ethaddr= rep->etp_ethaddr;
printf("osdep_eth_init: setting EPF_GOT_ADDR\n");
eth_port->etp_flags |= EPF_GOT_ADDR;
}
sr_add_minor(if2minor(ecp->ec_ifno, ETH_DEV_OFF),
i, eth_open, eth_close, eth_read,
eth_write, eth_ioctl, eth_cancel, eth_select);
eth_port->etp_flags |= EPF_ENABLED;
eth_port->etp_vlan= ecp->ec_vlan;
eth_port->etp_vlan_port= rep;
assert(eth_port->etp_vlan != 0);
eth_port->etp_wr_pack= 0;
eth_port->etp_rd_pack= 0;
eth_reg_vlan(rep, eth_port);
2005-04-21 16:53:53 +02:00
}
}
PUBLIC void eth_write_port(eth_port, pack)
eth_port_t *eth_port;
acc_t *pack;
{
assert(!no_ethWritePort);
assert(!eth_port->etp_vlan);
2005-04-21 16:53:53 +02:00
assert(eth_port->etp_wr_pack == NULL);
eth_port->etp_wr_pack= pack;
if (eth_port->etp_osdep.etp_state != OEPS_IDLE)
2005-04-21 16:53:53 +02:00
{
eth_port->etp_osdep.etp_flags |= OEPF_NEED_SEND;
return;
2005-04-21 16:53:53 +02:00
}
eth_issue_send(eth_port);
}
#if 0
PRIVATE int notification_count;
#endif
PUBLIC void eth_rec(message *m)
{
int i, r, m_type, stat;
eth_port_t *loc_port, *vlan_port;
char *drivername;
struct eth_conf *ecp;
m_type= m->m_type;
if (m_type == DL_NAME_REPLY)
2005-04-21 16:53:53 +02:00
{
drivername= m->m3_ca1;
#if 0
printf("eth_rec: got name: %s\n", drivername);
notification_count= 0;
#endif
/* Re-init ethernet interfaces */
for (i= 0, ecp= eth_conf, loc_port= eth_port_table;
i<eth_conf_nr; i++, ecp++, loc_port++)
2005-04-21 16:53:53 +02:00
{
if (eth_is_vlan(ecp))
continue;
if (strcmp(ecp->ec_task, drivername) != 0)
2006-07-10 14:55:33 +02:00
{
/* Wrong driver */
continue;
2006-07-10 14:55:33 +02:00
}
eth_restart(loc_port, m->m_source);
2005-04-21 16:53:53 +02:00
}
return;
2005-04-21 16:53:53 +02:00
}
assert(m_type == DL_CONF_REPLY || m_type == DL_TASK_REPLY ||
m_type == DL_STAT_REPLY);
2006-07-10 14:55:33 +02:00
for (i=0, loc_port= eth_port_table; i<eth_conf_nr; i++, loc_port++)
2005-04-21 16:53:53 +02:00
{
if (loc_port->etp_osdep.etp_port == m->DL_PORT &&
loc_port->etp_osdep.etp_task == m->m_source)
break;
}
if (i >= eth_conf_nr)
{
printf("eth_rec: bad port %d in message type 0x%x from %d\n",
m->DL_PORT, m_type, m->m_source);
return;
2005-04-21 16:53:53 +02:00
}
if (loc_port->etp_osdep.etp_state == OEPS_CONF_SENT)
2005-04-21 16:53:53 +02:00
{
if (m_type == DL_TASK_REPLY)
{
stat= m->DL_STAT & 0xffff;
2005-04-21 16:53:53 +02:00
if (stat & DL_PACK_SEND)
write_int(loc_port);
if (stat & DL_PACK_RECV)
read_int(loc_port, m->DL_COUNT);
return;
}
2005-04-21 16:53:53 +02:00
if (m_type != DL_CONF_REPLY)
2005-04-21 16:53:53 +02:00
{
printf(
"eth_rec: got bad message type 0x%x from %d in CONF state\n",
m_type, m->m_source);
return;
2005-04-21 16:53:53 +02:00
}
r= m->m3_i1;
if (r == ENXIO)
{
printf(
"eth_rec(conf_reply): no ethernet device at task=%d,port=%d\n",
loc_port->etp_osdep.etp_task,
loc_port->etp_osdep.etp_port);
return;
}
if (r < 0)
2005-04-21 16:53:53 +02:00
{
ip_panic(("eth_rec: DL_INIT returned error %d\n", r));
return;
2005-04-21 16:53:53 +02:00
}
loc_port->etp_osdep.etp_flags &= ~OEPF_NEED_CONF;
loc_port->etp_osdep.etp_state= OEPS_IDLE;
loc_port->etp_flags |= EPF_ENABLED;
loc_port->etp_ethaddr= *(ether_addr_t *)m->m3_ca1;
if (!(loc_port->etp_flags & EPF_GOT_ADDR))
2005-04-21 16:53:53 +02:00
{
loc_port->etp_flags |= EPF_GOT_ADDR;
#if 0
printf("eth_rec: calling eth_restart_ioctl\n");
#endif
eth_restart_ioctl(loc_port);
/* Also update any VLANs on this device */
for (i=0, vlan_port= eth_port_table; i<eth_conf_nr;
i++, vlan_port++)
{
if (!(vlan_port->etp_flags & EPF_ENABLED))
continue;
if (vlan_port->etp_vlan_port != loc_port)
continue;
vlan_port->etp_ethaddr= loc_port->etp_ethaddr;
vlan_port->etp_flags |= EPF_GOT_ADDR;
eth_restart_ioctl(vlan_port);
}
2005-04-21 16:53:53 +02:00
}
if (!(loc_port->etp_flags & EPF_READ_IP))
setup_read (loc_port);
2005-04-21 16:53:53 +02:00
#if 0
if (loc_port->etp_osdep.etp_flags & OEPF_NEED_SEND)
{
printf("eth_rec(conf): OEPF_NEED_SEND is set\n");
}
if (loc_port->etp_osdep.etp_flags & OEPF_NEED_RECV)
{
printf("eth_rec(conf): OEPF_NEED_RECV is set\n");
}
if (loc_port->etp_osdep.etp_flags & OEPF_NEED_STAT)
{
printf("eth_rec(conf): OEPF_NEED_STAT is set\n");
}
#endif
2005-04-21 16:53:53 +02:00
return;
}
if (loc_port->etp_osdep.etp_state == OEPS_GETSTAT_SENT)
2005-04-21 16:53:53 +02:00
{
if (m_type != DL_STAT_REPLY)
{
printf(
"eth_rec: got bad message type 0x%x from %d in GETSTAT state\n",
m_type, m->m_source);
return;
}
2005-04-21 16:53:53 +02:00
r= m->DL_STAT;
if (r != OK)
{
ip_warning(("eth_rec: DL_STAT returned error %d\n",
r));
return;
}
loc_port->etp_osdep.etp_state= OEPS_IDLE;
loc_port->etp_osdep.etp_flags &= ~OEPF_NEED_STAT;
2005-04-21 16:53:53 +02:00
assert(loc_port->etp_osdep.etp_stat_gid != -1);
cpf_revoke(loc_port->etp_osdep.etp_stat_gid);
loc_port->etp_osdep.etp_stat_gid= -1;
loc_port->etp_osdep.etp_stat_buf= NULL;
/* Finish ioctl */
assert(loc_port->etp_flags & EPF_GOT_ADDR);
eth_restart_ioctl(loc_port);
#if 0
if (loc_port->etp_osdep.etp_flags & OEPF_NEED_SEND)
{
printf("eth_rec(stat): OEPF_NEED_SEND is set\n");
}
if (loc_port->etp_osdep.etp_flags & OEPF_NEED_RECV)
{
printf("eth_rec(stat): OEPF_NEED_RECV is set\n");
}
if (loc_port->etp_osdep.etp_flags & OEPF_NEED_CONF)
{
printf("eth_rec(stat): OEPF_NEED_CONF is set\n");
}
#endif
#if 0
if (loc_port->etp_osdep.etp_state == OEPS_IDLE &&
(loc_port->etp_osdep.etp_flags & OEPF_NEED_CONF))
{
eth_set_rec_conf(loc_port,
loc_port->etp_osdep.etp_recvconf);
}
#endif
2005-04-21 16:53:53 +02:00
return;
}
assert(loc_port->etp_osdep.etp_state == OEPS_IDLE ||
loc_port->etp_osdep.etp_state == OEPS_RECV_SENT ||
loc_port->etp_osdep.etp_state == OEPS_SEND_SENT ||
(printf("etp_state = %d\n", loc_port->etp_osdep.etp_state), 0));
loc_port->etp_osdep.etp_state= OEPS_IDLE;
2005-04-21 16:53:53 +02:00
#if 0 /* Ethernet driver is not trusted */
2005-04-21 16:53:53 +02:00
set_time (m->DL_CLCK);
#endif
2005-04-21 16:53:53 +02:00
stat= m->DL_STAT & 0xffff;
#if 0
if (!(stat & (DL_PACK_SEND|DL_PACK_RECV)))
printf("eth_rec: neither DL_PACK_SEND nor DL_PACK_RECV\n");
#endif
2005-04-21 16:53:53 +02:00
if (stat & DL_PACK_SEND)
write_int(loc_port);
if (stat & DL_PACK_RECV)
{
if (recv_debug)
{
printf("eth_rec: eth%d got DL_PACK_RECV\n",
m->DL_PORT);
}
2005-04-21 16:53:53 +02:00
read_int(loc_port, m->DL_COUNT);
}
if (loc_port->etp_osdep.etp_state == OEPS_IDLE &&
loc_port->etp_osdep.etp_flags & OEPF_NEED_SEND)
{
loc_port->etp_osdep.etp_flags &= ~OEPF_NEED_SEND;
if (loc_port->etp_wr_pack)
eth_issue_send(loc_port);
}
if (loc_port->etp_osdep.etp_state == OEPS_IDLE &&
(loc_port->etp_osdep.etp_flags & OEPF_NEED_RECV))
{
loc_port->etp_osdep.etp_flags &= ~OEPF_NEED_RECV;
if (!(loc_port->etp_flags & EPF_READ_IP))
setup_read (loc_port);
}
if (loc_port->etp_osdep.etp_flags & OEPF_NEED_CONF)
{
New RS and new signal handling for system processes. UPDATING INFO: 20100317: /usr/src/etc/system.conf updated to ignore default kernel calls: copy it (or merge it) to /etc/system.conf. The hello driver (/dev/hello) added to the distribution: # cd /usr/src/commands/scripts && make clean install # cd /dev && MAKEDEV hello KERNEL CHANGES: - Generic signal handling support. The kernel no longer assumes PM as a signal manager for every process. The signal manager of a given process can now be specified in its privilege slot. When a signal has to be delivered, the kernel performs the lookup and forwards the signal to the appropriate signal manager. PM is the default signal manager for user processes, RS is the default signal manager for system processes. To enable ptrace()ing for system processes, it is sufficient to change the default signal manager to PM. This will temporarily disable crash recovery, though. - sys_exit() is now split into sys_exit() (i.e. exit() for system processes, which generates a self-termination signal), and sys_clear() (i.e. used by PM to ask the kernel to clear a process slot when a process exits). - Added a new kernel call (i.e. sys_update()) to swap two process slots and implement live update. PM CHANGES: - Posix signal handling is no longer allowed for system processes. System signals are split into two fixed categories: termination and non-termination signals. When a non-termination signaled is processed, PM transforms the signal into an IPC message and delivers the message to the system process. When a termination signal is processed, PM terminates the process. - PM no longer assumes itself as the signal manager for system processes. It now makes sure that every system signal goes through the kernel before being actually processes. The kernel will then dispatch the signal to the appropriate signal manager which may or may not be PM. SYSLIB CHANGES: - Simplified SEF init and LU callbacks. - Added additional predefined SEF callbacks to debug crash recovery and live update. - Fixed a temporary ack in the SEF init protocol. SEF init reply is now completely synchronous. - Added SEF signal event type to provide a uniform interface for system processes to deal with signals. A sef_cb_signal_handler() callback is available for system processes to handle every received signal. A sef_cb_signal_manager() callback is used by signal managers to process system signals on behalf of the kernel. - Fixed a few bugs with memory mapping and DS. VM CHANGES: - Page faults and memory requests coming from the kernel are now implemented using signals. - Added a new VM call to swap two process slots and implement live update. - The call is used by RS at update time and in turn invokes the kernel call sys_update(). RS CHANGES: - RS has been reworked with a better functional decomposition. - Better kernel call masks. com.h now defines the set of very basic kernel calls every system service is allowed to use. This makes system.conf simpler and easier to maintain. In addition, this guarantees a higher level of isolation for system libraries that use one or more kernel calls internally (e.g. printf). - RS is the default signal manager for system processes. By default, RS intercepts every signal delivered to every system process. This makes crash recovery possible before bringing PM and friends in the loop. - RS now supports fast rollback when something goes wrong while initializing the new version during a live update. - Live update is now implemented by keeping the two versions side-by-side and swapping the process slots when the old version is ready to update. - Crash recovery is now implemented by keeping the two versions side-by-side and cleaning up the old version only when the recovery process is complete. DS CHANGES: - Fixed a bug when the process doing ds_publish() or ds_delete() is not known by DS. - Fixed the completely broken support for strings. String publishing is now implemented in the system library and simply wraps publishing of memory ranges. Ideally, we should adopt a similar approach for other data types as well. - Test suite fixed. DRIVER CHANGES: - The hello driver has been added to the Minix distribution to demonstrate basic live update and crash recovery functionalities. - Other drivers have been adapted to conform the new SEF interface.
2010-03-17 02:15:29 +01:00
#if 0
printf("eth_rec: OEPF_NEED_CONF is set\n");
New RS and new signal handling for system processes. UPDATING INFO: 20100317: /usr/src/etc/system.conf updated to ignore default kernel calls: copy it (or merge it) to /etc/system.conf. The hello driver (/dev/hello) added to the distribution: # cd /usr/src/commands/scripts && make clean install # cd /dev && MAKEDEV hello KERNEL CHANGES: - Generic signal handling support. The kernel no longer assumes PM as a signal manager for every process. The signal manager of a given process can now be specified in its privilege slot. When a signal has to be delivered, the kernel performs the lookup and forwards the signal to the appropriate signal manager. PM is the default signal manager for user processes, RS is the default signal manager for system processes. To enable ptrace()ing for system processes, it is sufficient to change the default signal manager to PM. This will temporarily disable crash recovery, though. - sys_exit() is now split into sys_exit() (i.e. exit() for system processes, which generates a self-termination signal), and sys_clear() (i.e. used by PM to ask the kernel to clear a process slot when a process exits). - Added a new kernel call (i.e. sys_update()) to swap two process slots and implement live update. PM CHANGES: - Posix signal handling is no longer allowed for system processes. System signals are split into two fixed categories: termination and non-termination signals. When a non-termination signaled is processed, PM transforms the signal into an IPC message and delivers the message to the system process. When a termination signal is processed, PM terminates the process. - PM no longer assumes itself as the signal manager for system processes. It now makes sure that every system signal goes through the kernel before being actually processes. The kernel will then dispatch the signal to the appropriate signal manager which may or may not be PM. SYSLIB CHANGES: - Simplified SEF init and LU callbacks. - Added additional predefined SEF callbacks to debug crash recovery and live update. - Fixed a temporary ack in the SEF init protocol. SEF init reply is now completely synchronous. - Added SEF signal event type to provide a uniform interface for system processes to deal with signals. A sef_cb_signal_handler() callback is available for system processes to handle every received signal. A sef_cb_signal_manager() callback is used by signal managers to process system signals on behalf of the kernel. - Fixed a few bugs with memory mapping and DS. VM CHANGES: - Page faults and memory requests coming from the kernel are now implemented using signals. - Added a new VM call to swap two process slots and implement live update. - The call is used by RS at update time and in turn invokes the kernel call sys_update(). RS CHANGES: - RS has been reworked with a better functional decomposition. - Better kernel call masks. com.h now defines the set of very basic kernel calls every system service is allowed to use. This makes system.conf simpler and easier to maintain. In addition, this guarantees a higher level of isolation for system libraries that use one or more kernel calls internally (e.g. printf). - RS is the default signal manager for system processes. By default, RS intercepts every signal delivered to every system process. This makes crash recovery possible before bringing PM and friends in the loop. - RS now supports fast rollback when something goes wrong while initializing the new version during a live update. - Live update is now implemented by keeping the two versions side-by-side and swapping the process slots when the old version is ready to update. - Crash recovery is now implemented by keeping the two versions side-by-side and cleaning up the old version only when the recovery process is complete. DS CHANGES: - Fixed a bug when the process doing ds_publish() or ds_delete() is not known by DS. - Fixed the completely broken support for strings. String publishing is now implemented in the system library and simply wraps publishing of memory ranges. Ideally, we should adopt a similar approach for other data types as well. - Test suite fixed. DRIVER CHANGES: - The hello driver has been added to the Minix distribution to demonstrate basic live update and crash recovery functionalities. - Other drivers have been adapted to conform the new SEF interface.
2010-03-17 02:15:29 +01:00
#endif
}
if (loc_port->etp_osdep.etp_state == OEPS_IDLE &&
(loc_port->etp_osdep.etp_flags & OEPF_NEED_STAT))
{
send_getstat(loc_port);
}
2005-04-21 16:53:53 +02:00
}
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
PUBLIC void eth_check_driver(endpoint_t endpoint)
{
int r;
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
message m;
m.m_type = DL_GETNAME;
r= asynsend(endpoint, &m);
if (r != OK)
{
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
printf("eth_check_driver: asynsend to %d failed: %d\n",
endpoint, r);
return;
}
}
2005-04-21 16:53:53 +02:00
PUBLIC int eth_get_stat(eth_port, eth_stat)
eth_port_t *eth_port;
eth_stat_t *eth_stat;
{
2006-07-10 14:55:33 +02:00
cp_grant_id_t gid;
2005-04-21 16:53:53 +02:00
assert(!eth_port->etp_vlan);
if (eth_port->etp_osdep.etp_flags & OEPF_NEED_STAT)
ip_panic(( "eth_get_stat: getstat already in progress" ));
2006-07-10 14:55:33 +02:00
gid= cpf_grant_direct(eth_port->etp_osdep.etp_task,
(vir_bytes)eth_stat, sizeof(*eth_stat), CPF_WRITE);
if (gid == -1)
{
ip_panic(( "eth_get_stat: cpf_grant_direct failed: %d\n",
errno));
}
assert(eth_port->etp_osdep.etp_stat_gid == -1);
eth_port->etp_osdep.etp_stat_gid= gid;
eth_port->etp_osdep.etp_stat_buf= eth_stat;
2006-07-10 14:55:33 +02:00
if (eth_port->etp_osdep.etp_state != OEPS_IDLE)
2005-10-21 18:51:21 +02:00
{
eth_port->etp_osdep.etp_flags |= OEPF_NEED_STAT;
return SUSPEND;
2005-10-21 18:51:21 +02:00
}
send_getstat(eth_port);
2005-04-21 16:53:53 +02:00
return SUSPEND;
2005-04-21 16:53:53 +02:00
}
PUBLIC void eth_set_rec_conf (eth_port, flags)
eth_port_t *eth_port;
u32_t flags;
{
2005-10-21 18:51:21 +02:00
int r;
unsigned dl_flags, mask;
message mess;
2005-04-21 16:53:53 +02:00
assert(!eth_port->etp_vlan);
if (!(eth_port->etp_flags & EPF_GOT_ADDR))
{
/* We have never seen the device. */
#if 0
printf("eth_set_rec_conf: waiting for device to appear\n");
#endif
return;
}
if (eth_port->etp_osdep.etp_state != OEPS_IDLE)
{
printf(
"eth_set_rec_conf: setting OEPF_NEED_CONF, state = %d\n",
eth_port->etp_osdep.etp_state);
eth_port->etp_osdep.etp_flags |= OEPF_NEED_CONF;
return;
}
mask = NWEO_EN_BROAD | NWEO_EN_MULTI | NWEO_EN_PROMISC;
if ((eth_port->etp_osdep.etp_recvconf & mask) == (flags & mask))
{
/* No change for the driver, so don't send an update */
return;
}
2005-10-21 18:51:21 +02:00
eth_port->etp_osdep.etp_recvconf= flags;
2005-04-21 16:53:53 +02:00
dl_flags= DL_NOMODE;
if (flags & NWEO_EN_BROAD)
dl_flags |= DL_BROAD_REQ;
if (flags & NWEO_EN_MULTI)
dl_flags |= DL_MULTI_REQ;
if (flags & NWEO_EN_PROMISC)
dl_flags |= DL_PROMISC_REQ;
2006-07-10 14:55:33 +02:00
mess.m_type= DL_CONF;
2005-04-21 16:53:53 +02:00
mess.DL_PORT= eth_port->etp_osdep.etp_port;
mess.DL_PROC= this_proc;
mess.DL_MODE= dl_flags;
assert(eth_port->etp_osdep.etp_state == OEPS_IDLE);
r= asynsend(eth_port->etp_osdep.etp_task, &mess);
eth_port->etp_osdep.etp_state= OEPS_CONF_SENT;
if (r < 0)
2005-04-21 16:53:53 +02:00
{
printf("eth_set_rec_conf: asynsend to %d failed: %d\n",
eth_port->etp_osdep.etp_task, r);
return;
}
}
PRIVATE void eth_issue_send(eth_port)
eth_port_t *eth_port;
{
int i, r, pack_size;
acc_t *pack, *pack_ptr;
iovec_s_t *iovec;
message m;
iovec= eth_port->etp_osdep.etp_wr_iovec;
pack= eth_port->etp_wr_pack;
pack_size= 0;
for (i=0, pack_ptr= pack; i<IOVEC_NR && pack_ptr; i++,
pack_ptr= pack_ptr->acc_next)
{
r= cpf_setgrant_direct(iovec[i].iov_grant,
eth_port->etp_osdep.etp_task,
(vir_bytes)ptr2acc_data(pack_ptr),
(vir_bytes)pack_ptr->acc_length,
CPF_READ);
if (r != 0)
2005-04-21 16:53:53 +02:00
{
ip_panic((
"eth_write_port: cpf_setgrant_direct failed: %d\n",
errno));
}
pack_size += iovec[i].iov_size= pack_ptr->acc_length;
}
if (i>= IOVEC_NR)
{
pack= bf_pack(pack); /* packet is too fragmented */
eth_port->etp_wr_pack= pack;
pack_size= 0;
for (i=0, pack_ptr= pack; i<IOVEC_NR && pack_ptr;
i++, pack_ptr= pack_ptr->acc_next)
{
r= cpf_setgrant_direct(iovec[i].iov_grant,
eth_port->etp_osdep.etp_task,
(vir_bytes)ptr2acc_data(pack_ptr),
(vir_bytes)pack_ptr->acc_length,
CPF_READ);
if (r != 0)
2005-04-21 16:53:53 +02:00
{
ip_panic((
"eth_write_port: cpf_setgrant_direct failed: %d\n",
errno));
2005-04-21 16:53:53 +02:00
}
pack_size += iovec[i].iov_size= pack_ptr->acc_length;
2005-04-21 16:53:53 +02:00
}
}
assert (i< IOVEC_NR);
assert (pack_size >= ETH_MIN_PACK_SIZE);
r= cpf_setgrant_direct(eth_port->etp_osdep.etp_wr_vec_grant,
eth_port->etp_osdep.etp_task,
(vir_bytes)iovec,
(vir_bytes)(i * sizeof(iovec[0])),
CPF_READ);
if (r != 0)
2005-10-21 18:51:21 +02:00
{
ip_panic((
"eth_write_port: cpf_setgrant_direct failed: %d\n",
errno));
2005-10-21 18:51:21 +02:00
}
m.DL_COUNT= i;
m.DL_GRANT= eth_port->etp_osdep.etp_wr_vec_grant;
m.m_type= DL_WRITEV_S;
m.DL_PORT= eth_port->etp_osdep.etp_port;
m.DL_PROC= this_proc;
m.DL_MODE= DL_NOMODE;
assert(eth_port->etp_osdep.etp_state == OEPS_IDLE);
r= asynsend(eth_port->etp_osdep.etp_task, &m);
2005-04-21 16:53:53 +02:00
if (r < 0)
2005-04-21 16:53:53 +02:00
{
printf("eth_issue_send: send to %d failed: %d\n",
eth_port->etp_osdep.etp_task, r);
return;
2005-04-21 16:53:53 +02:00
}
eth_port->etp_osdep.etp_state= OEPS_SEND_SENT;
2005-04-21 16:53:53 +02:00
}
PRIVATE void write_int(eth_port_t *eth_port)
2005-04-21 16:53:53 +02:00
{
acc_t *pack;
int multicast;
u8_t *eth_dst_ptr;
pack= eth_port->etp_wr_pack;
if (pack == NULL)
{
printf("write_int: strange no packet on eth port %d\n",
eth_port-eth_port_table);
eth_restart_write(eth_port);
return;
}
2005-04-21 16:53:53 +02:00
eth_port->etp_wr_pack= NULL;
eth_dst_ptr= (u8_t *)ptr2acc_data(pack);
multicast= (*eth_dst_ptr & 1); /* low order bit indicates multicast */
if (multicast || (eth_port->etp_osdep.etp_recvconf & NWEO_EN_PROMISC))
{
assert(!no_ethWritePort);
no_ethWritePort= 1;
2005-04-21 16:53:53 +02:00
eth_arrive(eth_port, pack, bf_bufsize(pack));
assert(no_ethWritePort);
no_ethWritePort= 0;
}
2005-04-21 16:53:53 +02:00
else
bf_afree(pack);
eth_restart_write(eth_port);
}
PRIVATE void read_int(eth_port, count)
eth_port_t *eth_port;
int count;
{
acc_t *pack, *cut_pack;
pack= eth_port->etp_rd_pack;
eth_port->etp_rd_pack= NULL;
if (count < ETH_MIN_PACK_SIZE)
{
printf("mnx_eth`read_int: packet size too small (%d)\n",
count);
bf_afree(pack);
}
else if (count > ETH_MAX_PACK_SIZE_TAGGED)
{
printf("mnx_eth`read_int: packet size too big (%d)\n",
count);
bf_afree(pack);
}
else
{
cut_pack= bf_cut(pack, 0, count);
bf_afree(pack);
2005-04-21 16:53:53 +02:00
assert(!no_ethWritePort);
no_ethWritePort= 1;
eth_arrive(eth_port, cut_pack, count);
assert(no_ethWritePort);
no_ethWritePort= 0;
}
2005-04-21 16:53:53 +02:00
eth_port->etp_flags &= ~(EPF_READ_IP|EPF_READ_SP);
setup_read(eth_port);
}
PRIVATE void setup_read(eth_port)
eth_port_t *eth_port;
{
acc_t *pack, *pack_ptr;
message mess1;
2006-07-10 14:55:33 +02:00
iovec_s_t *iovec;
2005-04-21 16:53:53 +02:00
int i, r;
assert(!eth_port->etp_vlan);
2005-04-21 16:53:53 +02:00
assert(!(eth_port->etp_flags & (EPF_READ_IP|EPF_READ_SP)));
if (eth_port->etp_osdep.etp_state != OEPS_IDLE)
2005-04-21 16:53:53 +02:00
{
eth_port->etp_osdep.etp_flags |= OEPF_NEED_RECV;
2005-04-21 16:53:53 +02:00
return;
}
2005-04-21 16:53:53 +02:00
assert (!eth_port->etp_rd_pack);
2005-04-21 16:53:53 +02:00
iovec= eth_port->etp_osdep.etp_rd_iovec;
pack= bf_memreq (ETH_MAX_PACK_SIZE_TAGGED);
for (i=0, pack_ptr= pack; i<RD_IOVEC && pack_ptr;
i++, pack_ptr= pack_ptr->acc_next)
{
r= cpf_setgrant_direct(iovec[i].iov_grant,
2006-07-10 14:55:33 +02:00
eth_port->etp_osdep.etp_task,
(vir_bytes)ptr2acc_data(pack_ptr),
(vir_bytes)pack_ptr->acc_length,
CPF_WRITE);
2006-07-10 14:55:33 +02:00
if (r != 0)
{
ip_panic((
"mnx_eth`setup_read: cpf_setgrant_direct failed: %d\n",
errno));
}
iovec[i].iov_size= (vir_bytes)pack_ptr->acc_length;
}
assert (!pack_ptr);
2006-07-10 14:55:33 +02:00
r= cpf_setgrant_direct(eth_port->etp_osdep.etp_rd_vec_grant,
eth_port->etp_osdep.etp_task,
(vir_bytes)iovec,
(vir_bytes)(i * sizeof(iovec[0])),
CPF_READ);
if (r != 0)
{
ip_panic((
"mnx_eth`setup_read: cpf_setgrant_direct failed: %d\n",
errno));
}
mess1.m_type= DL_READV_S;
mess1.DL_PORT= eth_port->etp_osdep.etp_port;
mess1.DL_PROC= this_proc;
mess1.DL_COUNT= i;
mess1.DL_GRANT= eth_port->etp_osdep.etp_rd_vec_grant;
2005-04-21 16:53:53 +02:00
assert(eth_port->etp_osdep.etp_state == OEPS_IDLE);
r= asynsend(eth_port->etp_osdep.etp_task, &mess1);
eth_port->etp_osdep.etp_state= OEPS_RECV_SENT;
2005-04-21 16:53:53 +02:00
if (r < 0)
{
printf(
"mnx_eth`setup_read: asynsend to %d failed: %d\n",
eth_port->etp_osdep.etp_task, r);
}
eth_port->etp_rd_pack= pack;
eth_port->etp_flags |= EPF_READ_IP;
2005-04-21 16:53:53 +02:00
eth_port->etp_flags |= EPF_READ_SP;
}
PRIVATE void eth_recvev(ev, ev_arg)
event_t *ev;
ev_arg_t ev_arg;
{
eth_port_t *eth_port;
message *m_ptr;
eth_port= ev_arg.ev_ptr;
assert(ev == &eth_port->etp_osdep.etp_recvev);
m_ptr= &eth_port->etp_osdep.etp_recvrepl;
assert(m_ptr->m_type == DL_TASK_REPLY);
assert(eth_port->etp_osdep.etp_port == m_ptr->DL_PORT &&
eth_port->etp_osdep.etp_task == m_ptr->m_source);
2005-04-21 16:53:53 +02:00
assert(m_ptr->DL_STAT & DL_PACK_RECV);
m_ptr->DL_STAT &= ~DL_PACK_RECV;
if (recv_debug)
{
printf("eth_recvev: eth%d got DL_PACK_RECV\n", m_ptr->DL_PORT);
}
2005-04-21 16:53:53 +02:00
read_int(eth_port, m_ptr->DL_COUNT);
}
PRIVATE void eth_sendev(ev, ev_arg)
event_t *ev;
ev_arg_t ev_arg;
{
eth_port_t *eth_port;
message *m_ptr;
eth_port= ev_arg.ev_ptr;
assert(ev == &eth_port->etp_sendev);
m_ptr= &eth_port->etp_osdep.etp_sendrepl;
assert (m_ptr->m_type == DL_TASK_REPLY);
assert(eth_port->etp_osdep.etp_port == m_ptr->DL_PORT &&
eth_port->etp_osdep.etp_task == m_ptr->m_source);
2005-04-21 16:53:53 +02:00
assert(m_ptr->DL_STAT & DL_PACK_SEND);
m_ptr->DL_STAT &= ~DL_PACK_SEND;
assert(eth_port->etp_osdep.etp_send_ev);
eth_port->etp_osdep.etp_send_ev= 0;
2005-04-21 16:53:53 +02:00
/* packet is sent */
write_int(eth_port);
}
PRIVATE eth_port_t *find_port(m)
message *m;
{
eth_port_t *loc_port;
int i;
for (i=0, loc_port= eth_port_table; i<eth_conf_nr; i++, loc_port++)
{
if (loc_port->etp_osdep.etp_port == m->DL_PORT &&
loc_port->etp_osdep.etp_task == m->m_source)
2005-04-21 16:53:53 +02:00
break;
}
assert (i<eth_conf_nr);
return loc_port;
}
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 eth_restart(eth_port_t *eth_port, endpoint_t endpoint)
{
int r;
2005-10-21 18:51:21 +02:00
unsigned flags, dl_flags;
cp_grant_id_t gid;
message mess;
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
eth_port->etp_osdep.etp_task= endpoint;
switch(eth_port->etp_osdep.etp_state)
{
case OEPS_INIT:
case OEPS_CONF_SENT:
case OEPS_RECV_SENT:
case OEPS_SEND_SENT:
/* We can safely ignore the pending CONF, RECV, and SEND
* requests. If this is the first time that we see this
* driver at all, that's fine too.
*/
eth_port->etp_osdep.etp_state= OEPS_IDLE;
break;
case OEPS_GETSTAT_SENT:
/* Set the OEPF_NEED_STAT to trigger a new request */
eth_port->etp_osdep.etp_flags |= OEPF_NEED_STAT;
eth_port->etp_osdep.etp_state= OEPS_IDLE;
break;
}
/* If there is a pending GETSTAT request then we have to create a
* new grant.
*/
if (eth_port->etp_osdep.etp_flags & OEPF_NEED_STAT)
{
assert(eth_port->etp_osdep.etp_stat_gid != -1);
cpf_revoke(eth_port->etp_osdep.etp_stat_gid);
gid= cpf_grant_direct(eth_port->etp_osdep.etp_task,
(vir_bytes)eth_port->etp_osdep.etp_stat_buf,
sizeof(*eth_port->etp_osdep.etp_stat_buf), CPF_WRITE);
if (gid == -1)
{
ip_panic((
"eth_restart: cpf_grant_direct failed: %d\n",
errno));
}
eth_port->etp_osdep.etp_stat_gid= gid;
}
2005-10-21 18:51:21 +02:00
flags= eth_port->etp_osdep.etp_recvconf;
dl_flags= DL_NOMODE;
if (flags & NWEO_EN_BROAD)
dl_flags |= DL_BROAD_REQ;
if (flags & NWEO_EN_MULTI)
dl_flags |= DL_MULTI_REQ;
if (flags & NWEO_EN_PROMISC)
dl_flags |= DL_PROMISC_REQ;
2006-07-10 14:55:33 +02:00
mess.m_type= DL_CONF;
mess.DL_PORT= eth_port->etp_osdep.etp_port;
mess.DL_PROC= this_proc;
2005-10-21 18:51:21 +02:00
mess.DL_MODE= dl_flags;
compare(eth_port->etp_osdep.etp_state, ==, OEPS_IDLE);
r= asynsend(eth_port->etp_osdep.etp_task, &mess);
if (r<0)
{
printf(
"eth_restart: send to ethernet task %d failed: %d\n",
2005-10-21 18:51:21 +02:00
eth_port->etp_osdep.etp_task, r);
return;
}
eth_port->etp_osdep.etp_state= OEPS_CONF_SENT;
if (eth_port->etp_wr_pack)
2005-10-21 18:51:21 +02:00
{
bf_afree(eth_port->etp_wr_pack);
eth_port->etp_wr_pack= NULL;
eth_restart_write(eth_port);
}
if (eth_port->etp_rd_pack)
2005-10-21 18:51:21 +02:00
{
bf_afree(eth_port->etp_rd_pack);
eth_port->etp_rd_pack= NULL;
eth_port->etp_flags &= ~(EPF_READ_IP|EPF_READ_SP);
}
}
PRIVATE void send_getstat(eth_port)
eth_port_t *eth_port;
{
int r;
message mess;
mess.m_type= DL_GETSTAT_S;
mess.DL_PORT= eth_port->etp_osdep.etp_port;
mess.DL_PROC= this_proc;
mess.DL_GRANT= eth_port->etp_osdep.etp_stat_gid;
assert(eth_port->etp_osdep.etp_state == OEPS_IDLE);
r= asynsend(eth_port->etp_osdep.etp_task, &mess);
eth_port->etp_osdep.etp_state= OEPS_GETSTAT_SENT;
if (r != OK)
ip_panic(( "eth_get_stat: asynsend failed: %d", r));
}
2005-04-21 16:53:53 +02:00
/*
* $PchId: mnx_eth.c,v 1.16 2005/06/28 14:24:37 philip Exp $
2005-04-21 16:53:53 +02:00
*/