calibrate: now in a separate file (still not working)

f7f60598 · Alessandro Rubini · 8779d604 · f7f60598 · f7f60598 · f7f60598
Commit f7f60598 authored Apr 25, 2012 by Alessandro Rubini
Hide whitespace changes
Inline Side-by-side

Showing with 215 additions and 51 deletions

Makefile Makefile +1 -1

acam.c acam.c +6 -48

calibrate.c calibrate.c +179 -0

fine-delay.h fine-delay.h +29 -2

No files found.
--- a/Makefile
+++ b/Makefile
@@ -13,7 +13,7 @@ obj-m := spec-fine-delay.o
 spec-fine-delay-objs	=  fd-zio.o fd-spec.o fd-core.o
 spec-fine-delay-objs	+= onewire.o spi.o i2c.o gpio.o
-spec-fine-delay-objs	+= acam.o pll.o time.o
+spec-fine-delay-objs	+= acam.o calibrate.o pll.o time.o
 all: modules

--- a/acam.c
+++ b/acam.c
@@ -92,14 +92,14 @@ static void acam_set_address(struct spec_fd *fd, int addr)
 }
 /* Warning: acam_readl and acam_writel only work if GCR.BYPASS is set */
-static uint32_t acam_readl(struct spec_fd *fd, int reg)
+uint32_t acam_readl(struct spec_fd *fd, int reg)
 {
 	acam_set_address(fd, reg);
 	fd_writel(fd, FD_TDCSR_READ, FD_REG_TDCSR);
 	return fd_readl(fd, FD_REG_TDR) & ACAM_MASK;
 }
-static void acam_writel(struct spec_fd *fd, int val, int reg)
+void acam_writel(struct spec_fd *fd, int val, int reg)
 {
 	acam_set_address(fd, reg);
 	fd_writel(fd, val, FD_REG_TDR);
@@ -183,48 +183,6 @@ out:
 	return -EIO;
 }
-static int acam_test_delay_transfer_function(struct spec_fd *fd)
-{
-	/* FIXME */
-	return 0;
-}
-/* Evaluates 2nd order polynomial. Coefs have 32 fractional bits. */
-static int fd_eval_polynomial(struct spec_fd *fd)
-{
-	int64_t x = fd->temp;
-	int64_t *coef = fd->calib.frr_poly;
-	return (coef[0] * x * x + coef[1] * x + coef[2]) >> 32;
-}
-static int fd_find_8ns_tap(struct spec_fd *fd, int ch)
-{
-	/* FIXME */
-	return 0;
-}
-static int acam_calibrate_outputs(struct spec_fd *fd)
-{
-	int ret, ch;
-	int measured, fitted;
-	if ((ret = acam_test_delay_transfer_function(fd)) < 0)
-		return ret;
-	fitted = fd_eval_polynomial(fd);
-	for (ch = FD_CH_1; ch <= FD_CH_LAST; ch++) {
-		fd_read_temp(fd, 0);
-		measured = fd_find_8ns_tap(fd, ch);
-		fd->ch[ch].frr_cur = measured;
-		fd->ch[ch].frr_offset = measured - fitted;
-		pr_info("%s: ch %i: 8ns @ %i (f %i, offset %i, t %i.%02i)\n",
-			__func__, FD_CH_EXT(ch),
-			fd->ch[ch].frr_cur, fitted, fd->ch[ch].frr_offset,
-			fd->temp / 16, (fd->temp & 0xf) * 100 / 16);
-	}
-	return 0;
-}
 /* We need to write come static configuration in the registers */
 struct acam_init_data {
@@ -288,16 +246,16 @@ static struct acam_mode_setup fd_acam_table[] = {
 /* To configure the thing, follow the table, but treat 5 and 7 as special */
 static int __acam_config(struct spec_fd *fd, struct acam_mode_setup *s)
 {
-	int i, bin, hsdiv, refdiv, reg7val;
+	int i, hsdiv, refdiv, reg7val;
 	struct acam_init_data *p;
 	uint32_t regval;
 	unsigned long j;
-	bin = acam_calc_pll(ACAM_FP_TREF, ACAM_FP_BIN, &hsdiv, &refdiv);
+	fd->bin = acam_calc_pll(ACAM_FP_TREF, ACAM_FP_BIN, &hsdiv, &refdiv);
 	reg7val = AR7_HSDiv(hsdiv) | AR7_RefClkDiv(refdiv);
 	pr_debug("%s: config for %s-mode (bin 0x%x, hsdiv %i, refdiv %i)\n",
-		 __func__, s->name, bin, hsdiv, refdiv);
+		 __func__, s->name, fd->bin, hsdiv, refdiv);
 	/* Disable TDC inputs prior to configuring */
 	fd_writel(fd, FD_TDCSR_STOP_DIS | FD_TDCSR_START_DIS, FD_REG_TDCSR);
@@ -352,7 +310,7 @@ int fd_acam_init(struct spec_fd *fd)
 	if ( (ret = fd_acam_config(fd, ACAM_IMODE)) )
 		return ret;
-	if ( (ret = acam_calibrate_outputs(fd)) )
+	if ( (ret = fd_calibrate_outputs(fd)) )
 		return ret;
 	if ( (ret = fd_acam_config(fd, ACAM_RMODE)) )

--- a/calibrate.c
+++ b/calibrate.c
+/*
+ * Calibrate the output path.
+ *
+ * Copyright (C) 2012 CERN (www.cern.ch)
+ * Author: Tomasz Wlostowski <tomasz.wlostowski@cern.ch>
+ * Author: Alessandro Rubini <rubini@gnudd.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public License
+ * version 2 as published by the Free Software Foundation or, at your
+ * option, any later version.
+ */
+#include <linux/kernel.h>
+#include <linux/slab.h>
+#include <linux/io.h>
+#include <linux/delay.h>
+#include "fine-delay.h"
+#include "hw/fd_main_regs.h"
+#include "hw/acam_gpx.h"
+#include "hw/fd_channel_regs.h"
+/* TEMP! */
+static void acam_set_bypass(struct spec_fd *fd, int on)
+{
+	/* FIXME: this zeroes all other GCR bits */
+	fd_writel(fd, on ? FD_GCR_BYPASS : 0, FD_REG_GCR);
+}
+static int acam_test_delay_transfer_function(struct spec_fd *fd)
+{
+	/* FIXME */
+	return 0;
+}
+/* Evaluates 2nd order polynomial. Coefs have 32 fractional bits. */
+static int fd_eval_polynomial(struct spec_fd *fd)
+{
+	int64_t x = fd->temp;
+	int64_t *coef = fd->calib.frr_poly;
+	return (coef[0] * x * x + coef[1] * x + coef[2]) >> 32;
+}
+/*
+ * Measures the the FPGA-generated TDC start and the output of one of
+ * the fine delay chips (channel) at a pre-defined number of taps
+ * (fine). Retuns the delay in picoseconds. The measurement is
+ * repeated and averaged (n_avgs) times. Also, the standard deviation
+ * of the result can be written to (sdev) if it's not NULL.
+ */
+struct delay_stats {
+	uint32_t avg;
+	uint32_t min;
+	uint32_t max;
+};
+/* Note: channel is the "internal" one: 0..3 */
+static uint32_t output_delay_ps(struct spec_fd *fd, int ch, int fine, int n,
+				struct delay_stats *stats)
+{
+	int i;
+	uint32_t res, *results;
+	uint64_t acc = 0;
+	results = kmalloc(n * sizeof(*results), GFP_KERNEL);
+	if (!results)
+		return -ENOMEM;
+	/* Disable the output for the channel being calibrated */
+	fd_gpio_clr(fd, FD_GPIO_OUTPUT_EN(FD_CH_EXT(ch)));
+	/* Enable the stop input in ACAM for the channel being calibrated */
+	acam_writel(fd, AR0_TRiseEn(0) | AR0_TRiseEn(FD_CH_EXT(ch))
+		    | AR0_HQSel | AR0_ROsc, 0);
+	/* Program the output delay line setpoint */
+	fd_ch_writel(fd, ch, fine, FD_REG_FRR);
+	fd_ch_writel(fd, ch, FD_DCR_ENABLE | FD_DCR_MODE | FD_DCR_UPDATE,
+		    FD_REG_DCR);
+	fd_ch_writel(fd, ch, FD_DCR_FORCE_DLY | FD_DCR_ENABLE, FD_REG_DCR);
+	/*
+	 * Set the calibration pulse mask to genrate calibration
+	 * pulses only on one channel at a time.  This minimizes the
+	 * crosstalk in the output buffer which can severely decrease
+	 * the accuracy of calibration measurements
+	 */
+	fd_writel(fd, FD_CALR_PSEL_W(1 << ch), FD_REG_CALR);
+	udelay(100);
+	/* Do n_avgs single measurements and average */
+	for (i = 0; i < n; i++) {
+		uint32_t fr;
+		/* Re-arm the ACAM (it's working in a single-shot mode) */
+		fd_writel(fd, FD_TDCSR_ALUTRIG, FD_REG_TDCSR);
+		udelay(100);
+		/* Produce a calib pulse on the TDC start and the output ch */
+		fd_writel(fd, FD_CALR_CAL_PULSE |
+			  FD_CALR_PSEL_W(1 << ch), FD_REG_CALR);
+		udelay(10000);
+		/* read the tag, convert to picoseconds (fixed point: 16.16) */
+		fr = acam_readl(fd, 8 /* fifo */);
+		res = fr & 0x1ffff * fd->bin * 3; /* bin is 16.16 already */
+		printk("%i: %08x, 0x%08x\n", fine, fr, res);
+		results[i] = res;
+		acc += res;
+	}
+	fd_ch_writel(fd, ch, 0, FD_REG_DCR);
+	/* Calculate avg, min max */
+	acc = (acc + n / 2) / n;
+	if (stats) {
+		stats->avg = acc;
+		stats->min = ~0;
+		stats->max = 0;
+		for (i = 0; i < n; i++) {
+			if (results[i] > stats->max) stats->max = results[i];
+			if (results[i] < stats->min) stats->min = results[i];
+		}
+	}
+	kfree(results);
+	return acc;
+}
+static int fd_find_8ns_tap(struct spec_fd *fd, int ch)
+{
+	int l = 0, mid, r = FD_NUM_TAPS - 1;
+	uint32_t bias, dly;
+	/*
+	 * Measure the delay at zero setting, so it can be further
+	 * subtracted to get only the delay part introduced by the
+	 * delay line (ingoring the TDC, FPGA and routing delays).
+	 */
+	bias = output_delay_ps(fd, ch, 0, FD_CAL_STEPS, NULL);
+	while( r - l > 1) {
+		mid = ( l + r) / 2;
+		dly = output_delay_ps(fd, ch, mid, FD_CAL_STEPS, NULL)
+			- bias;
+		printk("%i %i %i %08xx (bias %08x)\n", l, r, mid, dly, bias);
+		if(dly < 8000 << 16)
+			l = mid;
+		else
+			r = mid;
+	}
+	return l;
+}
+int fd_calibrate_outputs(struct spec_fd *fd)
+{
+	int ret, ch;
+	int measured, fitted;
+	acam_set_bypass(fd, 1); /* not useful */
+	fd_writel(fd, FD_TDCSR_START_EN | FD_TDCSR_STOP_EN, FD_REG_TDCSR);
+	if ((ret = acam_test_delay_transfer_function(fd)) < 0)
+		return ret;
+	fitted = fd_eval_polynomial(fd);
+	for (ch = FD_CH_1; ch <= FD_CH_LAST; ch++) {
+		fd_read_temp(fd, 0);
+		measured = fd_find_8ns_tap(fd, ch);
+		fd->ch[ch].frr_cur = measured;
+		fd->ch[ch].frr_offset = measured - fitted;
+		pr_info("%s: ch %i: 8ns @ %i (f %i, offset %i, t %i.%02i)\n",
+			__func__, FD_CH_EXT(ch),
+			fd->ch[ch].frr_cur, fitted, fd->ch[ch].frr_offset,
+			fd->temp / 16, (fd->temp & 0xf) * 100 / 16);
+	}
+	return 0;
+}
--- a/fine-delay.h
+++ b/fine-delay.h
@@ -20,6 +20,9 @@ struct fd_calib {
 #define FD_CH_INT(i)	((i) - 1)
 #define FD_CH_EXT(i)	((i) + 1)
+#define FD_NUM_TAPS	1024
+#define FD_CAL_STEPS	16 //1024
 struct fd_ch {
 	/* Offset between FRR measured at known T at startup and poly-fitted */
 	uint32_t frr_offset;
@@ -36,6 +39,7 @@ struct spec_fd {
 	unsigned char __iomem *base;	/* regs files are byte-oriented */
 	unsigned char __iomem *regs;
 	unsigned char __iomem *ow_regs;
+	uint32_t bin;
 	int acam_addr;			/* cache of currently active addr */
 	uint8_t ds18_id[8];
 	unsigned long next_t;
@@ -61,6 +65,26 @@ static inline void fd_writel(struct spec_fd *fd, uint32_t v, unsigned long reg)
 	writel(v, fd->regs + reg);
 }
+static inline void __check_chan(int x)
+{
+	BUG_ON(x < 0 || x > 3);
+}
+static inline uint32_t fd_ch_readl(struct spec_fd *fd, int ch,
+				   unsigned long reg)
+{
+	__check_chan(ch);
+	return fd_readl(fd, 0x100 + ch * 0x100 + reg);
+}
+static inline void fd_ch_writel(struct spec_fd *fd, int ch,
+				uint32_t v, unsigned long reg)
+{
+	__check_chan(ch);
+	fd_writel(fd, v, 0x100 + ch * 0x100 + reg);
+}
 #define FD_REGS_OFFSET	0x84000
 #define FD_MAGIC_FPGA	0xf19ede1a	/* FD_REG_IDR content */
@@ -134,6 +158,11 @@ extern int fd_read_temp(struct spec_fd *fd, int verbose);
 /* Functions exported by acam.c */
 extern int fd_acam_init(struct spec_fd *fd);
 extern void fd_acam_exit(struct spec_fd *fd);
+extern uint32_t acam_readl(struct spec_fd *fd, int reg);
+extern void acam_writel(struct spec_fd *fd, int val, int reg);
+/* Functions exported by calibrate.c, called within acam.c */
+extern int fd_calibrate_outputs(struct spec_fd *fd);
 /* Functions exported by gpio.c */
 extern int fd_gpio_init(struct spec_fd *fd);
@@ -152,8 +181,6 @@ extern int fd_time_set(struct spec_fd *fd, struct fd_time *t,
 extern int fd_time_get(struct spec_fd *fd, struct fd_time *t,
 		       struct timespec *ts);
 /* Functions exported by fd-zio.c */
 extern int fd_zio_register(void);
 extern void fd_zio_unregister(void);