From: Ivo Sieben <meltedpianoman@gmail.com>
To: linux-serial@vger.kernel.org, RT <linux-rt-users@vger.kernel.org>,
Greg KH <gregkh@linuxfoundation.org>
Cc: Ivo Sieben <meltedpianoman@gmail.com>
Subject: [PATCH-v2] tty: Use raw spin lock to protect the TTY read section
Date: Thu, 24 Jan 2013 11:34:59 +0100 [thread overview]
Message-ID: <1359023699-27050-1-git-send-email-meltedpianoman@gmail.com> (raw)
In-Reply-To: <1348150166-4868-1-git-send-email-meltedpianoman@gmail.com>
The "normal" spin lock that guards the N_TTY line discipline read section
is replaced by a raw spin lock.
On a PREEMP_RT system this prevents unwanted scheduling overhead when data is
read at the same time as data is being received: while RX IRQ threaded handling
is busy a TTY read call is performed from a RT priority > threaded IRQ priority.
The read call tries to take the read section spin lock (held by the threaded
IRQ) which blocks and causes a context switch to/from the threaded IRQ handler
until the spin lock is unlocked.
On a 240 MHz AT91SAM9261 processor setup this fixes about 100us of scheduling
overhead on the TTY read call.
Signed-off-by: Ivo Sieben <meltedpianoman@gmail.com
---
v2:
Rebased this patch to the latest linux tree status.
I tested this on a 3.0.43-rt65 kernel, with the following Kernel Hacking options
enabled:
- CONFIG_DEBUG_RT_MUTEXES
- CONFIG_DEBUG_SPINLOCK
- CONFIG_DEBUG_MUTEXES
- CONFIG_DEBUG_LOCK_ALLOC
- CONFIG_PROVE_LOCKING
- CONFIG_DEBUG_SPINLOCK_SLEEP
drivers/tty/n_tty.c | 56 +++++++++++++++++++++++++--------------------------
1 file changed, 28 insertions(+), 28 deletions(-)
diff --git a/drivers/tty/n_tty.c b/drivers/tty/n_tty.c
index 19083ef..0950728 100644
--- a/drivers/tty/n_tty.c
+++ b/drivers/tty/n_tty.c
@@ -100,7 +100,7 @@ struct n_tty_data {
struct mutex atomic_read_lock;
struct mutex output_lock;
struct mutex echo_lock;
- spinlock_t read_lock;
+ raw_spinlock_t read_lock;
};
static inline int tty_put_user(struct tty_struct *tty, unsigned char x,
@@ -182,9 +182,9 @@ static void put_tty_queue(unsigned char c, struct n_tty_data *ldata)
* The problem of stomping on the buffers ends here.
* Why didn't anyone see this one coming? --AJK
*/
- spin_lock_irqsave(&ldata->read_lock, flags);
+ raw_spin_lock_irqsave(&ldata->read_lock, flags);
put_tty_queue_nolock(c, ldata);
- spin_unlock_irqrestore(&ldata->read_lock, flags);
+ raw_spin_unlock_irqrestore(&ldata->read_lock, flags);
}
/**
@@ -218,9 +218,9 @@ static void reset_buffer_flags(struct tty_struct *tty)
struct n_tty_data *ldata = tty->disc_data;
unsigned long flags;
- spin_lock_irqsave(&ldata->read_lock, flags);
+ raw_spin_lock_irqsave(&ldata->read_lock, flags);
ldata->read_head = ldata->read_tail = ldata->read_cnt = 0;
- spin_unlock_irqrestore(&ldata->read_lock, flags);
+ raw_spin_unlock_irqrestore(&ldata->read_lock, flags);
mutex_lock(&ldata->echo_lock);
ldata->echo_pos = ldata->echo_cnt = ldata->echo_overrun = 0;
@@ -276,7 +276,7 @@ static ssize_t n_tty_chars_in_buffer(struct tty_struct *tty)
unsigned long flags;
ssize_t n = 0;
- spin_lock_irqsave(&ldata->read_lock, flags);
+ raw_spin_lock_irqsave(&ldata->read_lock, flags);
if (!ldata->icanon) {
n = ldata->read_cnt;
} else if (ldata->canon_data) {
@@ -284,7 +284,7 @@ static ssize_t n_tty_chars_in_buffer(struct tty_struct *tty)
ldata->canon_head - ldata->read_tail :
ldata->canon_head + (N_TTY_BUF_SIZE - ldata->read_tail);
}
- spin_unlock_irqrestore(&ldata->read_lock, flags);
+ raw_spin_unlock_irqrestore(&ldata->read_lock, flags);
return n;
}
@@ -915,19 +915,19 @@ static void eraser(unsigned char c, struct tty_struct *tty)
kill_type = WERASE;
else {
if (!L_ECHO(tty)) {
- spin_lock_irqsave(&ldata->read_lock, flags);
+ raw_spin_lock_irqsave(&ldata->read_lock, flags);
ldata->read_cnt -= ((ldata->read_head - ldata->canon_head) &
(N_TTY_BUF_SIZE - 1));
ldata->read_head = ldata->canon_head;
- spin_unlock_irqrestore(&ldata->read_lock, flags);
+ raw_spin_unlock_irqrestore(&ldata->read_lock, flags);
return;
}
if (!L_ECHOK(tty) || !L_ECHOKE(tty) || !L_ECHOE(tty)) {
- spin_lock_irqsave(&ldata->read_lock, flags);
+ raw_spin_lock_irqsave(&ldata->read_lock, flags);
ldata->read_cnt -= ((ldata->read_head - ldata->canon_head) &
(N_TTY_BUF_SIZE - 1));
ldata->read_head = ldata->canon_head;
- spin_unlock_irqrestore(&ldata->read_lock, flags);
+ raw_spin_unlock_irqrestore(&ldata->read_lock, flags);
finish_erasing(ldata);
echo_char(KILL_CHAR(tty), tty);
/* Add a newline if ECHOK is on and ECHOKE is off. */
@@ -961,10 +961,10 @@ static void eraser(unsigned char c, struct tty_struct *tty)
break;
}
cnt = (ldata->read_head - head) & (N_TTY_BUF_SIZE-1);
- spin_lock_irqsave(&ldata->read_lock, flags);
+ raw_spin_lock_irqsave(&ldata->read_lock, flags);
ldata->read_head = head;
ldata->read_cnt -= cnt;
- spin_unlock_irqrestore(&ldata->read_lock, flags);
+ raw_spin_unlock_irqrestore(&ldata->read_lock, flags);
if (L_ECHO(tty)) {
if (L_ECHOPRT(tty)) {
if (!ldata->erasing) {
@@ -1344,12 +1344,12 @@ send_signal:
put_tty_queue(c, ldata);
handle_newline:
- spin_lock_irqsave(&ldata->read_lock, flags);
+ raw_spin_lock_irqsave(&ldata->read_lock, flags);
set_bit(ldata->read_head, ldata->read_flags);
put_tty_queue_nolock(c, ldata);
ldata->canon_head = ldata->read_head;
ldata->canon_data++;
- spin_unlock_irqrestore(&ldata->read_lock, flags);
+ raw_spin_unlock_irqrestore(&ldata->read_lock, flags);
kill_fasync(&tty->fasync, SIGIO, POLL_IN);
if (waitqueue_active(&tty->read_wait))
wake_up_interruptible(&tty->read_wait);
@@ -1423,7 +1423,7 @@ static void n_tty_receive_buf(struct tty_struct *tty, const unsigned char *cp,
unsigned long cpuflags;
if (ldata->real_raw) {
- spin_lock_irqsave(&ldata->read_lock, cpuflags);
+ raw_spin_lock_irqsave(&ldata->read_lock, cpuflags);
i = min(N_TTY_BUF_SIZE - ldata->read_cnt,
N_TTY_BUF_SIZE - ldata->read_head);
i = min(count, i);
@@ -1439,7 +1439,7 @@ static void n_tty_receive_buf(struct tty_struct *tty, const unsigned char *cp,
memcpy(ldata->read_buf + ldata->read_head, cp, i);
ldata->read_head = (ldata->read_head + i) & (N_TTY_BUF_SIZE-1);
ldata->read_cnt += i;
- spin_unlock_irqrestore(&ldata->read_lock, cpuflags);
+ raw_spin_unlock_irqrestore(&ldata->read_lock, cpuflags);
} else {
for (i = count, p = cp, f = fp; i; i--, p++) {
if (f)
@@ -1635,7 +1635,7 @@ static int n_tty_open(struct tty_struct *tty)
mutex_init(&ldata->atomic_read_lock);
mutex_init(&ldata->output_lock);
mutex_init(&ldata->echo_lock);
- spin_lock_init(&ldata->read_lock);
+ raw_spin_lock_init(&ldata->read_lock);
/* These are ugly. Currently a malloc failure here can panic */
ldata->read_buf = kzalloc(N_TTY_BUF_SIZE, GFP_KERNEL);
@@ -1703,10 +1703,10 @@ static int copy_from_read_buf(struct tty_struct *tty,
bool is_eof;
retval = 0;
- spin_lock_irqsave(&ldata->read_lock, flags);
+ raw_spin_lock_irqsave(&ldata->read_lock, flags);
n = min(ldata->read_cnt, N_TTY_BUF_SIZE - ldata->read_tail);
n = min(*nr, n);
- spin_unlock_irqrestore(&ldata->read_lock, flags);
+ raw_spin_unlock_irqrestore(&ldata->read_lock, flags);
if (n) {
retval = copy_to_user(*b, &ldata->read_buf[ldata->read_tail], n);
n -= retval;
@@ -1714,13 +1714,13 @@ static int copy_from_read_buf(struct tty_struct *tty,
ldata->read_buf[ldata->read_tail] == EOF_CHAR(tty);
tty_audit_add_data(tty, &ldata->read_buf[ldata->read_tail], n,
ldata->icanon);
- spin_lock_irqsave(&ldata->read_lock, flags);
+ raw_spin_lock_irqsave(&ldata->read_lock, flags);
ldata->read_tail = (ldata->read_tail + n) & (N_TTY_BUF_SIZE-1);
ldata->read_cnt -= n;
/* Turn single EOF into zero-length read */
if (L_EXTPROC(tty) && ldata->icanon && is_eof && !ldata->read_cnt)
n = 0;
- spin_unlock_irqrestore(&ldata->read_lock, flags);
+ raw_spin_unlock_irqrestore(&ldata->read_lock, flags);
*b += n;
*nr -= n;
}
@@ -1900,7 +1900,7 @@ do_it_again:
if (ldata->icanon && !L_EXTPROC(tty)) {
/* N.B. avoid overrun if nr == 0 */
- spin_lock_irqsave(&ldata->read_lock, flags);
+ raw_spin_lock_irqsave(&ldata->read_lock, flags);
while (nr && ldata->read_cnt) {
int eol;
@@ -1918,25 +1918,25 @@ do_it_again:
if (--ldata->canon_data < 0)
ldata->canon_data = 0;
}
- spin_unlock_irqrestore(&ldata->read_lock, flags);
+ raw_spin_unlock_irqrestore(&ldata->read_lock, flags);
if (!eol || (c != __DISABLED_CHAR)) {
if (tty_put_user(tty, c, b++)) {
retval = -EFAULT;
b--;
- spin_lock_irqsave(&ldata->read_lock, flags);
+ raw_spin_lock_irqsave(&ldata->read_lock, flags);
break;
}
nr--;
}
if (eol) {
tty_audit_push(tty);
- spin_lock_irqsave(&ldata->read_lock, flags);
+ raw_spin_lock_irqsave(&ldata->read_lock, flags);
break;
}
- spin_lock_irqsave(&ldata->read_lock, flags);
+ raw_spin_lock_irqsave(&ldata->read_lock, flags);
}
- spin_unlock_irqrestore(&ldata->read_lock, flags);
+ raw_spin_unlock_irqrestore(&ldata->read_lock, flags);
if (retval)
break;
} else {
--
1.7.9.5
next prev parent reply other threads:[~2013-01-24 10:35 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-09-20 14:09 [PATCH] tty: Use raw spin lock to protect the TTY read section Ivo Sieben
2013-01-24 10:34 ` Ivo Sieben [this message]
2013-01-25 16:13 ` [PATCH-v2] " Greg KH
2013-01-28 12:15 ` Ivo Sieben
2013-01-28 12:32 ` [PATCH-v3] " Ivo Sieben
2013-02-04 15:33 ` Thomas Gleixner
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=1359023699-27050-1-git-send-email-meltedpianoman@gmail.com \
--to=meltedpianoman@gmail.com \
--cc=gregkh@linuxfoundation.org \
--cc=linux-rt-users@vger.kernel.org \
--cc=linux-serial@vger.kernel.org \
/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
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).