hi Philippe,
attached patch fixes spelling in adeos r9c6 patch;
prioritary -> priority.
I resisted urge to tinker with sentences around them;
I didnt want to subtly alter the meaning of stuff I dont
understand.
hth, and thanks for your work.
jimc
--- adeos-linux-2.6.10-i386-r9c6.patch.orig 2005-03-13 22:24:51.000000000
-0700
+++ adeos-linux-2.6.10-i386-r9c6.patch 2005-03-14 06:24:15.611540350 -0700
@@ -701,7 +701,7 @@
+
+ /* Make the current domain (== sleeper) wait for the mutex to be
+ released. Adeos' pipelined scheme guarantees that the new
-+ sleeper _is_ more prioritary than any aslept domain since we
++ sleeper _is_ more priority than any aslept domain since we
+ have stalled each sleeper's stage. Must be called with local hw
+ interrupts off. */
+
@@ -734,7 +734,7 @@
+ flags = __test_and_set_bit(IPIPE_STALL_FLAG,&adp->cpudata[cpuid].status);
+
+ /* Two cases to handle here on SMP systems, only one for UP:
-+ 1) in case of a conflicting access from a prioritary domain
++ 1) in case of a conflicting access from a priority domain
+ running on the same cpu, make this domain sleep on the mutex,
+ and resume the current owner so it can release the lock asap.
+ 2) in case of a conflicting access from any domain on a
@@ -797,7 +797,7 @@
+ if (mutex->sleepq != NULL)
+ {
+ adomain_t *sleeper = mutex->sleepq;
-+ /* Wake up the first most prioritary sleeper. */
++ /* Wake up the first highest priority sleeper. */
+ mutex->sleepq = sleeper->m_link;
+ __adeos_switch_to(adp_cpu_current[cpuid],sleeper,cpuid);
+ adeos_load_cpuid();
@@ -2225,7 +2225,7 @@
+ if (adp == adp_root)
+ {
+ /* Make sure to re-enable hw interrupts to reduce
-+ preemption latency by more prioritary domains when
++ preemption latency by more priority domains when
+ calling Linux handlers. */
+
+ adeos_hw_sti();
@@ -2401,8 +2401,8 @@
+ in mind that domains having a higher priority than the current
+ one are sleeping on the adeos_suspend_domain() service. In
+ addition, domains having a lower priority have been preempted
-+ by an interrupt dispatched to a more prioritary domain. Once
-+ the first and most prioritary stage has been selected here, the
++ by an interrupt dispatched to a more priority domain. Once
++ the first and highest priority stage has been selected here, the
+ subsequent stages will be activated in turn when each visited
+ domain calls adeos_suspend_domain() to wake up its neighbour
+ down the pipeline. */
@@ -2419,7 +2419,7 @@
+ just like if it has been actually received from a hw source. This
+ both works for real and virtual interrupts. This also means that
+ the current domain might be immediately preempted by a more
-+ prioritary domain who happens to handle this interrupt. */
++ priority domain who happens to handle this interrupt. */
+
+int fastcall adeos_trigger_irq (unsigned irq)
+
@@ -5245,7 +5245,7 @@
+ in the pipeline will get a chance to process the event. The latter
+ will eventually be allowed to process its own event too if a valid
+ handler exists for it. Handler executions are always scheduled by
-+ the domain which raised the event for the prioritary domains
++ the domain which raised the event for the priority domains
+ wanting to be notified of such event. Note: evdata might be
+ NULL. */
+
@@ -5289,7 +5289,7 @@
+ next_domain->cpudata[cpuid].event_info.propagate = 0;
+ __set_bit(IPIPE_XPEND_FLAG,&next_domain->cpudata[cpuid].status);
+
-+ /* Let the prioritary domain process the event. */
++ /* Let the priority domain process the event. */
+ __adeos_switch_to(this_domain,next_domain,cpuid);
+
+ adeos_load_cpuid(); /* Processor might have changed. */
