* [Xenomai-help] imx1 port.
@ 2007-10-29 12:54 garryt
2007-10-29 13:38 ` Gilles Chanteperdrix
0 siblings, 1 reply; 4+ messages in thread
From: garryt @ 2007-10-29 12:54 UTC (permalink / raw)
To: xenomai
[-- Attachment #1: Type: text/plain, Size: 4251 bytes --]
Hello all,
I had time to try again adeos+xenomai on a imx1 board. (i know this is quite
old).
This looks stable, i have been testing for several hours.
I have followed the wiki documentation and performed some tests with xeno-test +
ping flooding on this board, again i have no other *similar* board to compare
results with, but here they are.
I wonder if the difference between user-mode and in-kernel mode should be as
much as this...
By the way enabling ipipe tracer increase response time by a factor 2! is this
ok with such HW ?
I have attached the patch if someone is interested in, any remarks regarding
this are welcome.
All tests have been performed with xeno-test + ping flooding the board.
----
# cat /proc/ipipe/version
1.7-06
# cat /proc/xenomai/version
2.3.4
running: ./run -- -p 500 -T 120 -t0 # latency
*
*
* Type ^C to stop this application.
*
*
== Sampling period: 500 us
== Test mode: periodic user-mode task
== All results in microseconds
warming up...
RTT| 00:00:01 (periodic user-mode task, 500 us period, priority 99)
RTH|-----lat min|-----lat avg|-----lat max|-overrun|----lat best|---lat worst
RTD| 83.062| 116.437| 125.437| 0| 83.062| 125.437
RTD| 54.500| 118.375| 141.437| 0| 54.500| 141.437
RTD| 53.437| 120.187| 145.500| 0| 53.437| 145.500
RTD| 50.312| 119.375| 149.000| 0| 50.312| 149.000
RTD| 55.562| 116.375| 124.750| 0| 50.312| 149.000
...
---|------------|------------|------------|--------|-------------------------
RTS| 45.250| 118.500| 150.937| 0| 00:02:00/00:02:00
Thu Jan 1 00:24:02 GMT 1970
running: ./run -- -p 500 -T 120 -t1 # latency
*
*
* Type ^C to stop this application.
*
*
== Sampling period: 500 us
== Test mode: in-kernel periodic task
== All results in microseconds
warming up...
RTT| 00:00:01 (in-kernel periodic task, 500 us period, priority 99)
RTH|-----lat min|-----lat avg|-----lat max|-overrun|----lat best|---lat worst
RTD| 8.812| 19.329| 57.625| 0| 8.812| 57.625
RTD| 8.812| 15.628| 57.125| 0| 8.812| 57.625
RTD| 8.812| 16.827| 70.562| 0| 8.812| 70.562
RTD| 8.812| 29.156| 56.562| 0| 8.812| 70.562
...
RTS| 8.437| 17.812| 78.437| 0| 00:02:00/00:02:00
Thu Jan 1 00:26:09 GMT 1970
running: ./run -- -p 500 -T 120 -t2 # latency
*
*
* Type ^C to stop this application.
*
*
== Sampling period: 500 us
== Test mode: in-kernel timer handler
== All results in microseconds
warming up...
RTT| 00:00:01 (in-kernel timer handler, 500 us period, priority 99)
RTH|-----lat min|-----lat avg|-----lat max|-overrun|----lat best|---lat worst
RTD| 1.250| 7.492| 21.000| 0| 1.250| 21.000
RTD| 1.250| 5.211| 27.375| 0| 1.250| 27.375
RTD| 1.312| 5.299| 28.812| 0| 1.250| 28.812
...
RTS| 1.187| 4.335| 44.437| 0| 00:02:00/00:02:00
Thu Jan 1 00:28:15 GMT 1970
running: ./run -- -T 120 # switchtest
*
*
* Type ^C to stop this application.
*
*
== Testing FPU check routines...
== FPU check routines: unimplemented, skipping FPU switches tests.
== Threads: sleeper-0 rtk-1 rtk-2 rtup-3 rtup-4 rtus-5 rtus-6 rtuo-7 rtuo-8
RTT| 00:00:01
RTH|ctx switches|-------total
RTD| 900| 900
RTD| 900| 1800
RTD| 909| 2709
...
running: ./run -- -p 500 -h # switchbench
*
*
* Type ^C to stop this application.
*
*
== Sampling period: 500 us
== Do not interrupt this program
RTH| lat min| lat avg| lat max| lost
RTD| 47.937| 50.250| 88.687| 0
---|---range-|---samples
HSD| 47 - 48 | 1
HSD| 48 - 49 | 40757
HSD| 49 - 50 | 55940
HSD| 50 - 51 | 9
HSD| 67 - 68 | 1
HSD| 68 - 69 | 4
HSD| 86 - 87 | 75
HSD| 87 - 88 | 2741
HSD| 88 - 89 | 468
HSS| 99996| 49.845| 6.888
[-- Attachment #2: 100-linux-2.6.20-adeos-ipipe.diff --]
[-- Type: application/octet-stream, Size: 11292 bytes --]
--- linux-2.6.20.12/include/asm-arm/arch-imx/irqs.h 2007-05-24 23:21:02.000000000 +0200
+++ linux-2.6.20.12-ipipe/include/asm-arm/arch-imx/irqs.h 2007-07-04 11:47:34.000000000 +0200
@@ -113,4 +113,17 @@
#define NR_IRQS (IRQ_GPIOD(32) + 1)
#define IRQ_GPIO(x)
+
+#ifdef CONFIG_IPIPE
+#define __ipipe_irqbit(irq) (1ULL << (irq))
+
+#define __ipipe_muxed_irqmask (__ipipe_irqbit(GPIO_INT_PORTA) | \
+ __ipipe_irqbit(GPIO_INT_PORTB) | \
+ __ipipe_irqbit(GPIO_INT_PORTC) | \
+ __ipipe_irqbit(GPIO_INT_PORTD))
+
+#define __ipipe_mach_irq_mux_p(irq) (__ipipe_irqbit(irq) \
+ & __ipipe_muxed_irqmask)
+#endif /* CONFIG_IPIPE */
+
#endif
--- linux-2.6.20.12/include/asm-arm/arch-imx/imx-regs.h 2007-09-05 11:58:55.000000000 +0200
+++ linux-2.6.20.12-ipipe/include/asm-arm/arch-imx/imx-regs.h 2007-05-31 14:36:31.000000000 +0200
@@ -345,7 +345,16 @@
#define IMX_INTDISNUM __REG(IMX_AITC_BASE+0x0c)
#define IMX_INTENABLEH __REG(IMX_AITC_BASE+0x10)
#define IMX_INTENABLEL __REG(IMX_AITC_BASE+0x14)
+#define IMX_PRIO7 __REG(IMX_AITC_BASE+0x20)
+#define IMX_PRIO6 __REG(IMX_AITC_BASE+0x24)
+#define IMX_PRIO5 __REG(IMX_AITC_BASE+0x28)
+#define IMX_PRIO4 __REG(IMX_AITC_BASE+0x2C)
+#define IMX_PRIO3 __REG(IMX_AITC_BASE+0x30)
+#define IMX_PRIO2 __REG(IMX_AITC_BASE+0x34)
+#define IMX_PRIO1 __REG(IMX_AITC_BASE+0x38)
+#define IMX_PRIO0 __REG(IMX_AITC_BASE+0x3C)
+#define IMX_PRIO(x) __REG2(IMX_AITC_BASE+0x20,((7-((x)>>3)) <<2) )
/*
* General purpose timers
*/
@@ -415,6 +424,7 @@
#define PCR_BPIX_8 (3<<25)
#define PCR_BPIX_12 (4<<25)
#define PCR_BPIX_16 (4<<25)
+#define PCR_BPIX_MASK (7<<25)
#define PCR_PIXPOL (1<<24)
#define PCR_FLMPOL (1<<23)
#define PCR_LPPOL (1<<22)
--- linux-2.6.20.12/arch/arm/mach-imx/irq.c 2007-09-05 16:28:42.000000000 +0200
+++ linux-2.6.20.12/arch/arm/mach-imx/irq.c 2007-09-05 16:36:02.000000000 +0200
@@ -26,12 +26,87 @@
#include <linux/init.h>
#include <linux/list.h>
#include <linux/timer.h>
+#ifdef CONFIG_IPIPE
+#include <linux/ipipe.h>
+#endif /* CONFIG_IPIPE */
#include <asm/hardware.h>
#include <asm/irq.h>
#include <asm/io.h>
#include <asm/mach/irq.h>
+#include <asm/arch/imx-regs.h>
+
+/* Used for IMX INTERRUPT priority: Still Experimental */
+#ifdef CONFIG_IPIPE
+
+static unsigned int imx_default_irq_priority[IMX_IRQS] __initdata = {
+ 0, /* unused */
+ 0, /* unused */
+ 0, /* unused */
+ 0, /* unused */
+ 0, /* unused */
+ 0, /* unused */
+ 5, /* Common Sensor Interface */
+ 5, /* Multimedia Accelerator MAC */
+ 5, /* Multimedia Accelerator */
+ 0, /* unused */
+ 5, /* MS */
+ 4, /* GPIO Port A */
+ 4, /* GPIO Port B */
+ 4, /* GPIO Port C */
+ 5, /* LCDC */
+ 0, /* unused */
+ 0, /* unused */
+ 7, /* RTC */
+ 7, /* RTC */
+ 5, /* UART 2 PFerr*/
+ 5, /* UART 2 RTS */
+ 5, /* UART 2 DTR */
+ 5, /* UART 2 UARTC*/
+ 5, /* UART 2 Tx*/
+ 5, /* UART 2 Rx*/
+ 5, /* UART 1 PFerr*/
+ 5, /* UART 1 RTS */
+ 5, /* UART 1 DTR */
+ 5, /* UART 1 UARTC*/
+ 5, /* UART 1 Tx*/
+ 5, /* UART 1 Rx*/
+ 0, /* unused */
+ 0, /* unused */
+ 0, /* unused */
+ 5,/* PWM */
+ 5, /* MMC */
+ 0, /* unused */
+ 0, /* unused */
+ 0, /* unused */
+ 5, /* I2C */
+ 5, /*SPI 2 */
+ 5, /*SPI 1 */
+ 5, /* SSI Tx*/
+ 5, /* SSI Tx Err*/
+ 5, /* SSI Rx */
+ 5, /* SSI Rx Err*/
+ 0, /* unused */
+ 5, /* USB0 */
+ 5, /* USB1 */
+ 5, /* USB2 */
+ 5, /* USB3 */
+ 5, /* USB4 */
+ 5, /* USB5 */
+ 5, /* USB6 */
+ 0, /* unused */
+ 0, /* unused */
+ 0, /* unused */
+ 0, /* unused */
+ 5, /* Timer2 */
+ 15, /* Timer1 This is use by IPIPE */
+ 12, /* DMA Err */
+ 12, /* DMA */
+ 4, /* GPIO D */
+ 12 /* Watch dog */ /* Advanced Interrupt Controller (IRQ6) */
+ };
+#endif
/*
*
@@ -222,6 +296,25 @@
.set_type = imx_gpio_irq_type,
};
+#ifdef CONFIG_IPIPE
+void __init imx_init_priority(void)
+{
+ unsigned int irq;
+ IMX_PRIO0=0;
+ IMX_PRIO1=0;
+ IMX_PRIO2=0;
+ IMX_PRIO3=0;
+ IMX_PRIO4=0;
+ IMX_PRIO5=0;
+ IMX_PRIO6=0;
+ IMX_PRIO7=0;
+ printk(KERN_INFO "Initializing imx interrupt priorities\n");
+ for (irq = 0; irq < IMX_IRQS; irq++) {
+ IMX_PRIO(irq)|=((imx_default_irq_priority[irq]&15)<<((irq%8)*4));
+ }
+}
+#endif
+
void __init
imx_init_irq(void)
{
@@ -246,7 +339,9 @@
set_irq_handler(irq, handle_edge_irq);
set_irq_flags(irq, IRQF_VALID);
}
-
+#ifdef CONFIG_IPIPE
+ imx_init_priority();
+#endif
set_irq_chained_handler(GPIO_INT_PORTA, imx_gpioa_demux_handler);
set_irq_chained_handler(GPIO_INT_PORTB, imx_gpiob_demux_handler);
set_irq_chained_handler(GPIO_INT_PORTC, imx_gpioc_demux_handler);
@@ -255,3 +350,29 @@
/* Disable all interrupts initially. */
/* In IMX this is done in the bootloader. */
}
+
+#ifdef CONFIG_IPIPE
+void __ipipe_mach_demux_irq(unsigned irq, struct pt_regs *regs)
+{
+ struct irq_desc *desc_unused = irq_desc + irq;
+ unsigned irq_unused = irq;
+ unsigned int i, mask;
+
+ /* GPPIOA, GPIOB, GPIOC, GPIOD are INT 11,12,13 and 62 */
+ i=((irq&7)-3);
+
+ /* Get all multiplexed ITs from given GPIO */
+ mask=ISR(i);
+ irq=(i<<5)+IMX_IRQS;
+ do {
+ if (mask & 1 ) {
+ __ipipe_handle_irq(irq, regs);
+ // Ack multiplexed IT from given GPIO
+ ISR(i)=1<<(irq&0x1f);
+ }
+ mask >>=1;
+ irq++;
+ } while (mask);
+ desc_unused->chip->unmask(irq_unused);
+}
+#endif /* CONFIG_IPIPE */
--- linux-2.6.20.12/arch/arm/mach-imx/time.c 2007-10-10 17:11:18.000000000 +0200
+++ linux-2.6.20.12-ipipe/arch/arm/mach-imx/time.c 2007-10-11 12:42:07.000000000 +0200
--- time.c.orig 2007-10-10 17:11:18.000000000 +0200
+++ time.c 2007-10-29 10:18:02.000000000 +0100
@@ -15,6 +15,7 @@
#include <linux/irq.h>
#include <linux/time.h>
#include <linux/clocksource.h>
+#include <linux/module.h>
#include <asm/hardware.h>
#include <asm/io.h>
@@ -22,6 +23,55 @@
#include <asm/irq.h>
#include <asm/mach/time.h>
+#ifdef CONFIG_IPIPE
+#ifdef CONFIG_NO_IDLE_HZ
+#error "dynamic tick timer not yet supported with IPIPE"
+#endif /* CONFIG_NO_IDLE_HZ */
+int __ipipe_mach_timerint = TIM1_INT;
+EXPORT_SYMBOL(__ipipe_mach_timerint);
+
+int __ipipe_mach_timerstolen = 0;
+EXPORT_SYMBOL(__ipipe_mach_timerstolen);
+
+unsigned int __ipipe_mach_ticks_per_jiffy = LATCH;
+EXPORT_SYMBOL(__ipipe_mach_ticks_per_jiffy);
+
+static int imx_timer_initialized;
+static unsigned long last_jiffy_time;
+union tsc_reg {
+#ifdef __BIG_ENDIAN
+ struct {
+ unsigned long high;
+ unsigned long low;
+ };
+#else /* __LITTLE_ENDIAN */
+ struct {
+ unsigned long low;
+ unsigned long high;
+ };
+#endif /* __LITTLE_ENDIAN */
+ unsigned long long full;
+};
+#ifdef CONFIG_SMP
+static union tsc_reg tsc[NR_CPUS];
+
+void __ipipe_mach_get_tscinfo(struct __ipipe_tscinfo *info)
+{
+ info->type = IPIPE_TSC_TYPE_NONE;
+}
+#else /* !CONFIG_SMP */
+static union tsc_reg *tsc;
+
+void __ipipe_mach_get_tscinfo(struct __ipipe_tscinfo *info)
+{
+ info->type = IPIPE_TSC_TYPE_FREERUNNING;
+ info->u.fr.counter = (unsigned *)(0x10 + IMX_TIM1_BASE);
+ info->u.fr.mask = 0xffffffff;
+ info->u.fr.tsc = &tsc->full;
+}
+#endif /* !CONFIG_SMP */
+#endif /* CONFIG_IPIPE */
+
/* Use timer 1 as system timer */
#define TIMER_BASE IMX_TIM1_BASE
@@ -30,26 +80,43 @@
/*
* IRQ handler for the timer
*/
-static irqreturn_t
-imx_timer_interrupt(int irq, void *dev_id)
+static irqreturn_t imx_timer_interrupt(int irq, void *dev_id)
{
uint32_t tstat;
+#ifdef CONFIG_IPIPE
+ /*
+ * - if Linux is running natively (no ipipe), ack and reprogram the timer
+ * - if Linux is running under ipipe, but it still has the control over
+ * the timer (no Xenomai for example), then reprogram the timer (ipipe
+ * has already acked it)
+ * - if some other domain has taken over the timer, then do nothing
+ * (ipipe has acked it, and the other domain has reprogramed it)
+ */
+ if (__ipipe_mach_timerstolen) {
+ timer_tick();
+ last_jiffy_time += LATCH;
+ } else
+#else /* !CONFIG_IPIPE */
/* clear the interrupt */
tstat = IMX_TSTAT(TIMER_BASE);
IMX_TSTAT(TIMER_BASE) = 0;
if (tstat & TSTAT_COMP) {
+#endif
do {
-
write_seqlock(&xtime_lock);
timer_tick();
+#ifdef CONFIG_IPIPE
+ last_jiffy_time += LATCH;
+#endif /* CONFIG_IPIPE */
write_sequnlock(&xtime_lock);
IMX_TCMP(TIMER_BASE) += evt_diff;
-
- } while (unlikely((int32_t)(IMX_TCMP(TIMER_BASE)
+ } while (unlikely((int32_t) (IMX_TCMP(TIMER_BASE)
- IMX_TCN(TIMER_BASE)) < 0));
+#ifndef CONFIG_IPIPE
}
+#endif /* CONFIG_IPIPE */
return IRQ_HANDLED;
}
@@ -108,8 +175,91 @@
* Make irqs happen for the system timer
*/
setup_irq(TIM1_INT, &imx_timer_irq);
+
+#ifdef CONFIG_IPIPE
+#ifndef CONFIG_SMP
+ tsc = (union tsc_reg *)__ipipe_tsc_area;
+ barrier();
+#endif /* CONFIG_SMP */
+ imx_timer_initialized = 1;
+
+#endif /* CONFIG_IPIPE */
}
struct sys_timer imx_timer = {
.init = imx_timer_init,
};
+
+#ifdef CONFIG_IPIPE
+void __ipipe_mach_acktimer(void)
+{
+ uint32_t tstat;
+ tstat = IMX_TSTAT(TIMER_BASE);
+ IMX_TSTAT(TIMER_BASE) = 0;
+ if (likely(imx_timer_initialized)) {
+ union tsc_reg *local_tsc;
+ unsigned long stamp, flags;
+
+ local_irq_save_hw(flags);
+ local_tsc = &tsc[ipipe_processor_id()];
+ stamp = IMX_TCN(TIMER_BASE);
+ if (unlikely(stamp < local_tsc->low))
+ /* 32 bit counter wrapped, increment high word. */
+ local_tsc->high++;
+ local_tsc->low = stamp;
+ local_irq_restore_hw(flags);
+ }
+}
+
+notrace unsigned long long __ipipe_mach_get_tsc(void)
+{
+ if (likely(imx_timer_initialized)) {
+ union tsc_reg *local_tsc, result;
+ unsigned long stamp;
+
+ local_tsc = &tsc[ipipe_processor_id()];
+
+ __asm__("ldmia %1, %M0\n":"=r"(result.full),
+ "+&r"
+ (local_tsc)
+ : "m"(*local_tsc));
+ barrier();
+ stamp = IMX_TCN(TIMER_BASE);
+ if (unlikely(stamp < result.low))
+ result.high++;
+ result.low = stamp;
+ return result.full;
+ }
+ return 0;
+}
+
+EXPORT_SYMBOL(__ipipe_mach_get_tsc);
+
+/*
+ * Reprogram the timer
+ */
+void __ipipe_mach_set_dec(unsigned long delay)
+{
+ unsigned long flags;
+ if (delay > 8) {
+ local_irq_save_hw(flags);
+ IMX_TCMP(TIMER_BASE) = IMX_TCN(TIMER_BASE) + delay;
+ local_irq_restore_hw(flags);
+ } else
+ ipipe_trigger_irq(TIM1_INT);
+}
+
+EXPORT_SYMBOL(__ipipe_mach_set_dec);
+
+void __ipipe_mach_release_timer(void)
+{
+ __ipipe_mach_set_dec(__ipipe_mach_ticks_per_jiffy);
+}
+
+EXPORT_SYMBOL(__ipipe_mach_release_timer);
+
+unsigned long __ipipe_mach_get_dec(void)
+{
+ return IMX_TCMP(TIMER_BASE) - IMX_TCN(TIMER_BASE);
+}
+#endif /* CONFIG_IPIPE */
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [Xenomai-help] imx1 port.
2007-10-29 12:54 [Xenomai-help] imx1 port garryt
@ 2007-10-29 13:38 ` Gilles Chanteperdrix
2007-10-29 16:30 ` garryt
0 siblings, 1 reply; 4+ messages in thread
From: Gilles Chanteperdrix @ 2007-10-29 13:38 UTC (permalink / raw)
To: garryt; +Cc: xenomai
On 10/29/07, garryt <garryt@domain.hid> wrote:
> Hello all,
>
> I had time to try again adeos+xenomai on a imx1 board. (i know this is quite
> old).
Thanks. The imx port should be integrated in the next I-pipe patch.
>
> This looks stable, i have been testing for several hours.
>
> I have followed the wiki documentation and performed some tests with xeno-test +
> ping flooding on this board, again i have no other *similar* board to compare
> results with, but here they are.
>
> I wonder if the difference between user-mode and in-kernel mode should be as
> much as this...
>From the tests I have seen so far, except those published by Adeneo,
there is always a great difference between user-mode an kernel-mode.
This is the price to pay for the comfort of user-space on ARM.
>
> By the way enabling ipipe tracer increase response time by a factor 2! is this
> ok with such HW ?
I guess so.
>
> I have attached the patch if someone is interested in, any remarks regarding
> this are welcome.
There is something suspicious, it is the loop:
do {
write_seqlock(&xtime_lock);
timer_tick();
+#ifdef CONFIG_IPIPE
+ last_jiffy_time += LATCH;
+#endif /* CONFIG_IPIPE */
write_sequnlock(&xtime_lock);
IMX_TCMP(TIMER_BASE) += evt_diff;
} while (unlikely((int32_t) (IMX_TCMP(TIMER_BASE)
- IMX_TCN(TIMER_BASE)) < 0));
First, you should not touch IMX_TCMP(TIMER_BASE) in the loop, this is
done by __ipipe_mach_set_dec.
Second, with the current version of the I-pipe, the do/while loop
should be disabled when CONFIG_IPIPE is enabled. However, this is
going to change in the next version of I-pipe because of the so-called
"flat log" optimization, but I think IMX_TCMP(TIMER_BASE) should be
replaced with last_jiffy_time, because IMX_TCMP(TIMER_BASE) is the
date of the previous hardware timer tick, but the previous hardware
timer tick is not necessarily Linux timer previous tick. Anyway, I
will try and fix this and send a patch for testing.
>
>
> All tests have been performed with xeno-test + ping flooding the board.
For a load that give worse figures, run the calibrator
http://www.cwi.nl/~manegold/Calibrator
--
Gilles Chanteperdrix
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [Xenomai-help] imx1 port.
2007-10-29 13:38 ` Gilles Chanteperdrix
@ 2007-10-29 16:30 ` garryt
2007-10-29 16:46 ` Gilles Chanteperdrix
0 siblings, 1 reply; 4+ messages in thread
From: garryt @ 2007-10-29 16:30 UTC (permalink / raw)
To: Gilles Chanteperdrix; +Cc: garryt, xenomai
Hello,
Selon Gilles Chanteperdrix <gilles.chanteperdrix@xenomai.org>:
> On 10/29/07, garryt <garryt@domain.hid> wrote:
> > Hello all,
> >
> > I had time to try again adeos+xenomai on a imx1 board. (i know this is
> quite
> > old).
>
> Thanks. The imx port should be integrated in the next I-pipe patch.
>
> >
> > This looks stable, i have been testing for several hours.
> >
> > I have followed the wiki documentation and performed some tests with
> xeno-test +
> > ping flooding on this board, again i have no other *similar* board to
> compare
> > results with, but here they are.
> >
> > I wonder if the difference between user-mode and in-kernel mode should be
> as
> > much as this...
>
> From the tests I have seen so far, except those published by Adeneo,
> there is always a great difference between user-mode an kernel-mode.
> This is the price to pay for the comfort of user-space on ARM.
>
> >
> > By the way enabling ipipe tracer increase response time by a factor 2! is
> this
> > ok with such HW ?
>
> I guess so.
>
> >
> > I have attached the patch if someone is interested in, any remarks
> regarding
> > this are welcome.
>
> There is something suspicious, it is the loop:
>
> do {
> write_seqlock(&xtime_lock);
> timer_tick();
> +#ifdef CONFIG_IPIPE
> + last_jiffy_time += LATCH;
> +#endif /* CONFIG_IPIPE */
> write_sequnlock(&xtime_lock);
> IMX_TCMP(TIMER_BASE) += evt_diff;
> } while (unlikely((int32_t) (IMX_TCMP(TIMER_BASE)
> - IMX_TCN(TIMER_BASE)) < 0));
>
> First, you should not touch IMX_TCMP(TIMER_BASE) in the loop, this is
> done by __ipipe_mach_set_dec.
>
> Second, with the current version of the I-pipe, the do/while loop
> should be disabled when CONFIG_IPIPE is enabled. However, this is
> going to change in the next version of I-pipe because of the so-called
> "flat log" optimization, but I think IMX_TCMP(TIMER_BASE) should be
> replaced with last_jiffy_time, because IMX_TCMP(TIMER_BASE) is the
> date of the previous hardware timer tick, but the previous hardware
> timer tick is not necessarily Linux timer previous tick. Anyway, I
> will try and fix this and send a patch for testing.
>
Please correct me if i am wrong, this will help, my understanding from the
comment in the function (in 1.7-06.patch), was that if another domain (xenomai)
has stolen the timer we should do nothing neither ack it (already done), nor
reprogram the timer in linux interrupt handler...just run timer_tick. (should i
used time_lock)
I though that if we enter the loop this means that the timer programming is
still held by linux not xenomai.
> >
> >
> > All tests have been performed with xeno-test + ping flooding the board.
>
> For a load that give worse figures, run the calibrator
> http://www.cwi.nl/~manegold/Calibrator
>
>
> --
> Gilles Chanteperdrix
>
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [Xenomai-help] imx1 port.
2007-10-29 16:30 ` garryt
@ 2007-10-29 16:46 ` Gilles Chanteperdrix
0 siblings, 0 replies; 4+ messages in thread
From: Gilles Chanteperdrix @ 2007-10-29 16:46 UTC (permalink / raw)
To: garryt; +Cc: xenomai
On 10/29/07, garryt <garryt@domain.hid> wrote:
> Please correct me if i am wrong, this will help, my understanding from the
> comment in the function (in 1.7-06.patch), was that if another domain (xenomai)
> has stolen the timer we should do nothing neither ack it (already done), nor
> reprogram the timer in linux interrupt handler...just run timer_tick. (should i
> used time_lock)
>
> I though that if we enter the loop this means that the timer programming is
> still held by linux not xenomai.
Right, the only thing which should be skipped is what is in
__ipipe_mach_acktimer. So, sorry for the noise, your patch is OK.
--
Gilles Chanteperdrix
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2007-10-29 16:46 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-10-29 12:54 [Xenomai-help] imx1 port garryt
2007-10-29 13:38 ` Gilles Chanteperdrix
2007-10-29 16:30 ` garryt
2007-10-29 16:46 ` Gilles Chanteperdrix
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.