Removed unused minipc calls:
-check_running
-reset_port
-calibration_cmd

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Change waiting for shm lock. Until now, after read attempt there was
usleep/sleep then check if data was read correctly. Now wait in sleep/usleep
only on not consistent data read.

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Using not compatible versions of shm data can produce unpredictable results.
To be save always check version of shm after opening it.

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

After moving from mimipc to shm, function halexp_client_init is not used.

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Remove opening hal's shm in hal_client.c since it is not used further.
Only caller of function halexp_client_try_connect (wr_phytool) opens shm
anyway.

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Read locks needed to have consistent data reads.

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Needed to keep consistent reads from shmem.

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

In case there are all buckets full, rtu_fd_create_entry function returns error,
but also should unload mutex.

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Remove get_fd_list and get_vd_list from rtud's minipc since they're not used
anymore. Rtud's clients use now shmem to get the same information.
Remove helper functions from rtu_fd.c as not needed:
-rtu_fd_lookup_htab_entry
-rtu_vlan_entry_get
-vlan_entry_rd

Additionally following structures were removed as not needed:
-rtudexp_fd_entry_t
-rtudexp_fd_list_t
-rtudexp_vd_entry_t
-rtudexp_vd_list_t

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Part of a process to use shmem instead of minipc

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

All global data and most functions was made static.
Moved functions' declarations from wrs_vlans.h to wrs_vlans.c.

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

From now, rtu_stat will use rtud's shm to get vlan and htab entries instead of minipc

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Print error when wrs_shm_follow is not successful in wrs_dump_shmem.
Avoid segfaults.

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Adam Wujek authored 10 years ago and Alessandro Rubini committed 10 years ago

avoid including rtu.h by rtud clients

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Adam Wujek authored 10 years ago and Alessandro Rubini committed 10 years ago

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Adam Wujek authored 10 years ago and Alessandro Rubini committed 10 years ago

WRS_SHM_MAX_SIZE has to be increased because rtud data is bigger than 256KB
vlan_tab is 81920
rtu_htab is 180224
rtud total data: 262,144+shm_header

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Adam Wujek authored 10 years ago and Alessandro Rubini committed 10 years ago

Add rtu_shmem_header to keep track of all rtu's data in shmem

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Adam Wujek authored 10 years ago and Alessandro Rubini committed 10 years ago

Add "rtu_" to structure names in rtu_shmem.h and change all users.

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Move definitions of structures used in rtud's shm to rtu_shm.h in libwr.
Move few more definitions to rtu_shm.h and update .h files relaying on moved
structures/definitions.

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Remove unused minipc call get_timing_state.
Remove functions functions halexp_get_timing_state, export_get_timing_state
and struct hexp_timing_state_t.
Update developer manual

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Update developer manual after removal of get_port_state.

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

hal_port_lookup was redundant with hal_lookup_port.
File hal_shmem.c is not needed anymore.
add const qualifier to name parameter of hal_lookup_port function

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

also removed typedefs
hexp_port_state_t
hexp_port_list_t

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Minipc call to get port state will not be used anymore, since HAL export port
state through shmem.

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Function hal_port_get_exported_state, fills structure hal_port_state with data
that are not available in SHM. It just assigns defines.
These defines are private to HAL, and assignment should be done in HAL.

Similar problem is for wrpc. There is a wrapper function that pretends to be
HAL. All HAL info should be filled in that function.

Fields added:
--uint32_t clock_period;
--uint32_t t2_phase_transition;
--uint32_t t4_phase_transition;

Additionally updated wrs_dump_shmem.c to dump new fields and the version number
of HAL SHM.

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

No technical change

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Alessandro Rubini authored 10 years ago and Adam Wujek committed 10 years ago

It can be regenerated if needed (and added to .gitignore I hope).
At this point, it's not useful. It actually hits us in "git grep"
when we look for code.

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Alessandro Rubini authored 10 years ago and Adam Wujek committed 10 years ago

-DDEBUG in libwr forces LOG_DEBUG level in libwr/wrs-msg.c. This is
the desired behaviour, but we want it off by default now.

All other DEBUG symbols in Makefiles have no effect, so they are
removed to avoid tricking our developers in setting/unsetting them
and see no effect at all.

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Alessandro Rubini authored 10 years ago and Adam Wujek committed 10 years ago

This was a build-time conditional. It is now a run-time conditional
using wrs-msg.  As a result, we have a verbose rtud. Fixed in next
commit.

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Alessandro Rubini authored 10 years ago and Adam Wujek committed 10 years ago

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Previous commits made a number of local variables and cast assignment
into identity operations. So this removes them.

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

typedefs are bad; and typedefs to void pointers more so (there is
no checking by the compiler any more).  Use an opaque structure instead.

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

There is no need to use packed structures unless they match hardware
and or stored-structures of legacy processors.  etherpacket here may
look like an issue, but it's naturally aligned by itself. Nobody does it
for a reason.

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Meanwhile, remove one of the typedefs and avoid "my_socket":
"wr_socket" is more correct.

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

meanwhile, turn "6" to ETH_ALEN. It doesn't make sense to typedef
if it doesn't hide details; so let's use uint8_t and it's defined size.

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>