LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: "Arve Hjønnevåg" <arve@android.com>
To: linux-arm-kernel@lists.infradead.org
Cc: linux-omap@vger.kernel.org, "Arve Hjønnevåg" <arve@android.com>,
	"Russell King" <linux@arm.linux.org.uk>,
	"Alexander Shishkin" <virtuoso@slind.org>,
	"Jason Wessel" <jason.wessel@windriver.com>,
	"Greg Kroah-Hartman" <gregkh@suse.de>,
	linux-kernel@vger.kernel.org
Subject: [PATCH 3/8] ARM: etm: Don't try to clear the buffer full status after reading the buffer
Date: Mon, 14 Feb 2011 22:11:11 -0800	[thread overview]
Message-ID: <1297750276-12475-3-git-send-email-arve@android.com> (raw)
In-Reply-To: <1296701663-12168-1-git-send-email-arve@android.com>

If the write address was at the end of the buffer, toggling the trace
capture bit would set the RAM-full status instead of clearing it, and
if any of the stop bits in the formatter is set toggling the trace
capture bit may not do anything.

Instead use the read position to find out if the data has already
been returned.

This also fixes the read function so it works when the trace buffer is
larger than the buffer passed in from user space. The old version
would reset the trace buffer pointers after every read, so the second
call to read would always return 0.

Signed-off-by: Arve Hjønnevåg <arve@android.com>
---
 arch/arm/kernel/etm.c |   58 ++++++++++++++++++++++---------------------------
 1 files changed, 26 insertions(+), 32 deletions(-)

diff --git a/arch/arm/kernel/etm.c b/arch/arm/kernel/etm.c
index 6e15c48..bc7d8f2 100644
--- a/arch/arm/kernel/etm.c
+++ b/arch/arm/kernel/etm.c
@@ -91,6 +91,7 @@ static int trace_start(struct tracectx *t)
 
 	etb_unlock(t);
 
+	etb_writel(t, 0, ETBR_WRITEADDR);
 	etb_writel(t, 0, ETBR_FORMATTERCTRL);
 	etb_writel(t, 1, ETBR_CTRL);
 
@@ -184,24 +185,15 @@ static int trace_stop(struct tracectx *t)
 static int etb_getdatalen(struct tracectx *t)
 {
 	u32 v;
-	int rp, wp;
+	int wp;
 
 	v = etb_readl(t, ETBR_STATUS);
 
 	if (v & 1)
 		return t->etb_bufsz;
 
-	rp = etb_readl(t, ETBR_READADDR);
 	wp = etb_readl(t, ETBR_WRITEADDR);
-
-	if (rp > wp) {
-		etb_writel(t, 0, ETBR_READADDR);
-		etb_writel(t, 0, ETBR_WRITEADDR);
-
-		return 0;
-	}
-
-	return wp - rp;
+	return wp;
 }
 
 /* sysrq+v will always stop the running trace and leave it at that */
@@ -234,14 +226,6 @@ static void etm_dump(void)
 		printk("%08x", cpu_to_be32(etb_readl(t, ETBR_READMEM)));
 	printk(KERN_INFO "\n--- ETB buffer end ---\n");
 
-	/* deassert the overflow bit */
-	etb_writel(t, 1, ETBR_CTRL);
-	etb_writel(t, 0, ETBR_CTRL);
-
-	etb_writel(t, 0, ETBR_TRIGGERCOUNT);
-	etb_writel(t, 0, ETBR_READADDR);
-	etb_writel(t, 0, ETBR_WRITEADDR);
-
 	etb_lock(t);
 }
 
@@ -275,6 +259,10 @@ static ssize_t etb_read(struct file *file, char __user *data,
 	struct tracectx *t = file->private_data;
 	u32 first = 0;
 	u32 *buf;
+	int wpos;
+	int skip;
+	long wlength;
+	loff_t pos = *ppos;
 
 	mutex_lock(&t->mutex);
 
@@ -289,28 +277,34 @@ static ssize_t etb_read(struct file *file, char __user *data,
 	if (total == t->etb_bufsz)
 		first = etb_readl(t, ETBR_WRITEADDR);
 
+	if (pos > total * 4) {
+		skip = 0;
+		wpos = total;
+	} else {
+		skip = (int)pos % 4;
+		wpos = (int)pos / 4;
+	}
+	total -= wpos;
+	first = (first + wpos) % t->etb_bufsz;
+
 	etb_writel(t, first, ETBR_READADDR);
 
-	length = min(total * 4, (int)len);
-	buf = vmalloc(length);
+	wlength = min(total, DIV_ROUND_UP(skip + (int)len, 4));
+	length = min(total * 4 - skip, (int)len);
+	buf = vmalloc(wlength * 4);
 
-	dev_dbg(t->dev, "ETB buffer length: %d\n", total);
+	dev_dbg(t->dev, "ETB read %ld bytes to %lld from %ld words at %d\n",
+		length, pos, wlength, first);
+	dev_dbg(t->dev, "ETB buffer length: %d\n", total + wpos);
 	dev_dbg(t->dev, "ETB status reg: %x\n", etb_readl(t, ETBR_STATUS));
-	for (i = 0; i < length / 4; i++)
+	for (i = 0; i < wlength; i++)
 		buf[i] = etb_readl(t, ETBR_READMEM);
 
-	/* the only way to deassert overflow bit in ETB status is this */
-	etb_writel(t, 1, ETBR_CTRL);
-	etb_writel(t, 0, ETBR_CTRL);
-
-	etb_writel(t, 0, ETBR_WRITEADDR);
-	etb_writel(t, 0, ETBR_READADDR);
-	etb_writel(t, 0, ETBR_TRIGGERCOUNT);
-
 	etb_lock(t);
 
-	length -= copy_to_user(data, buf, length);
+	length -= copy_to_user(data, (u8 *)buf + skip, length);
 	vfree(buf);
+	*ppos = pos + length;
 
 out:
 	mutex_unlock(&t->mutex);
-- 
1.7.3.1


  parent reply	other threads:[~2011-02-15  6:12 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-02-03  2:54 [PATCH 1/5] ARM: etm: Don't require clock control Arve Hjønnevåg
2011-02-03  2:54 ` [PATCH 2/5] ARM: etm: Don't limit tracing to only non-secure code Arve Hjønnevåg
2011-02-03  2:54 ` [PATCH 3/5] ARM: etm: Don't trigger another overflow when trying to clear the RAM-full status Arve Hjønnevåg
2011-02-03  2:54 ` [PATCH 4/5] ARM: etm: Allow range selection Arve Hjønnevåg
2011-02-03  2:54 ` [PATCH 5/5] ARM: etm: Configure data tracing Arve Hjønnevåg
2011-02-03 12:45 ` [PATCH 1/5] ARM: etm: Don't require clock control Mark Brown
2011-02-04  0:30   ` Arve Hjønnevåg
2011-02-04 14:31     ` Mark Brown
2011-02-15  6:11 ` [PATCH 1/8] " Arve Hjønnevåg
2011-02-15  6:11 ` [PATCH 2/8] ARM: etm: Don't limit tracing to only non-secure code Arve Hjønnevåg
2011-02-15  6:11 ` Arve Hjønnevåg [this message]
2011-02-15  6:11 ` [PATCH 4/8] ARM: etm: Allow range selection Arve Hjønnevåg
2011-02-15 13:50   ` Alexander Shishkin
2011-02-15 23:04     ` Arve Hjønnevåg
2011-02-15  6:11 ` [PATCH 5/8] ARM: etm: Configure data tracing Arve Hjønnevåg
2011-02-15  6:11 ` [PATCH 6/8] ARM: etm: Add some missing locks and error checks Arve Hjønnevåg
2011-02-15  6:11 ` [PATCH 7/8] ARM: etm: Return the entire trace buffer if it is empty after reset Arve Hjønnevåg
2011-02-15  6:11 ` [PATCH 8/8] ARM: etm: Support multiple ETMs/PTMs Arve Hjønnevåg
2011-02-24  1:36 ` [PATCH 9/9 (originally 8)] ARM: etm: Power down etm(s) when tracing is not enabled Arve Hjønnevåg

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1297750276-12475-3-git-send-email-arve@android.com \
    --to=arve@android.com \
    --cc=gregkh@suse.de \
    --cc=jason.wessel@windriver.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-omap@vger.kernel.org \
    --cc=linux@arm.linux.org.uk \
    --cc=virtuoso@slind.org \
    --subject='Re: [PATCH 3/8] ARM: etm: Don'\''t try to clear the buffer full status after reading the buffer' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).