diff mbox

[v10] iopoll: Introduce memory-mapped IO polling macros

Message ID 1418687243-16395-1-git-send-email-mitchelh@codeaurora.org (mailing list archive)
State New, archived
Headers show

Commit Message

Mitchel Humpherys Dec. 15, 2014, 11:47 p.m. UTC
From: Matt Wagantall <mattw@codeaurora.org>

It is sometimes necessary to poll a memory-mapped register until its value
satisfies some condition. Introduce a family of convenience macros that do
this. Tight-looping, sleeping, and timing out can all be accomplished using
these macros.

Cc: Thierry Reding <thierry.reding@gmail.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Robert Elliott <elliott@hp.com>
Signed-off-by: Matt Wagantall <mattw@codeaurora.org>
Signed-off-by: Mitchel Humpherys <mitchelh@codeaurora.org>
---
v9..10:
  - Actually added the comments mentioned in v8..v9 (doh!)

v8..v9:
  - Added note in comments about max sleep time (Rob Elliott)

v7..v8:
  - sorted helper macros by size (b, w, l, q)
  - removed some of the more esoteric (or flat-out bogus) helper macros

This patch was originally part of a series [1] for adding support for IOMMU
address translations through an ARM SMMU hardware register.  The other
patch in the series (the one that actually uses these macros and implements
said hardware address translations) was Ack'd by the driver maintainer
there (Will Deacon) so I've pulled this patch out to avoid resending an
already Ack'd patch over and over again.

In short, please see [1] for previous discussion and the first user of
these macros.

Will also acked this patch in [2].  I didn't retain his Ack here because I
added to the macro comments.

[1] http://thread.gmane.org/gmane.linux.kernel.iommu/7140
[2] http://thread.gmane.org/gmane.linux.kernel.iommu/7449

---
 include/linux/iopoll.h | 144 +++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 144 insertions(+)
 create mode 100644 include/linux/iopoll.h

Comments

Will Deacon Dec. 16, 2014, 9:45 a.m. UTC | #1
On Mon, Dec 15, 2014 at 11:47:23PM +0000, Mitchel Humpherys wrote:
> From: Matt Wagantall <mattw@codeaurora.org>
> 
> It is sometimes necessary to poll a memory-mapped register until its value
> satisfies some condition. Introduce a family of convenience macros that do
> this. Tight-looping, sleeping, and timing out can all be accomplished using
> these macros.
> 
> Cc: Thierry Reding <thierry.reding@gmail.com>
> Cc: Will Deacon <will.deacon@arm.com>
> Cc: Arnd Bergmann <arnd@arndb.de>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: Robert Elliott <elliott@hp.com>
> Signed-off-by: Matt Wagantall <mattw@codeaurora.org>
> Signed-off-by: Mitchel Humpherys <mitchelh@codeaurora.org>
> ---
> v9..10:
>   - Actually added the comments mentioned in v8..v9 (doh!)
> 
> v8..v9:
>   - Added note in comments about max sleep time (Rob Elliott)
> 
> v7..v8:
>   - sorted helper macros by size (b, w, l, q)
>   - removed some of the more esoteric (or flat-out bogus) helper macros
> 
> This patch was originally part of a series [1] for adding support for IOMMU
> address translations through an ARM SMMU hardware register.  The other
> patch in the series (the one that actually uses these macros and implements
> said hardware address translations) was Ack'd by the driver maintainer
> there (Will Deacon) so I've pulled this patch out to avoid resending an
> already Ack'd patch over and over again.
> 
> In short, please see [1] for previous discussion and the first user of
> these macros.
> 
> Will also acked this patch in [2].  I didn't retain his Ack here because I
> added to the macro comments.

You can keep the ack, it still looks good to me and I'm not really fussed
about the comments.

Will
Mitchel Humpherys Jan. 14, 2015, 7:42 p.m. UTC | #2
On Tue, Dec 16 2014 at 01:45:27 AM, Will Deacon <will.deacon@arm.com> wrote:
> On Mon, Dec 15, 2014 at 11:47:23PM +0000, Mitchel Humpherys wrote:
>> From: Matt Wagantall <mattw@codeaurora.org>
>> 
>> It is sometimes necessary to poll a memory-mapped register until its value
>> satisfies some condition. Introduce a family of convenience macros that do
>> this. Tight-looping, sleeping, and timing out can all be accomplished using
>> these macros.
>> 
>> Cc: Thierry Reding <thierry.reding@gmail.com>
>> Cc: Will Deacon <will.deacon@arm.com>
>> Cc: Arnd Bergmann <arnd@arndb.de>
>> Cc: Andrew Morton <akpm@linux-foundation.org>
>> Cc: Robert Elliott <elliott@hp.com>
>> Signed-off-by: Matt Wagantall <mattw@codeaurora.org>
>> Signed-off-by: Mitchel Humpherys <mitchelh@codeaurora.org>
>> ---
>> v9..10:
>>   - Actually added the comments mentioned in v8..v9 (doh!)
>> 
>> v8..v9:
>>   - Added note in comments about max sleep time (Rob Elliott)
>> 
>> v7..v8:
>>   - sorted helper macros by size (b, w, l, q)
>>   - removed some of the more esoteric (or flat-out bogus) helper macros
>> 
>> This patch was originally part of a series [1] for adding support for IOMMU
>> address translations through an ARM SMMU hardware register.  The other
>> patch in the series (the one that actually uses these macros and implements
>> said hardware address translations) was Ack'd by the driver maintainer
>> there (Will Deacon) so I've pulled this patch out to avoid resending an
>> already Ack'd patch over and over again.
>> 
>> In short, please see [1] for previous discussion and the first user of
>> these macros.
>> 
>> Will also acked this patch in [2].  I didn't retain his Ack here because I
>> added to the macro comments.
>
> You can keep the ack, it still looks good to me and I'm not really fussed
> about the comments.
>
> Will

This hasn't gotten any further comments.  Would someone be willing to
take it?

Joerg, maybe you could take this through the IOMMU tree since the first
user is an IOMMU driver?  Currently we can't move [1] forward because of
this dependency...

[1] http://thread.gmane.org/gmane.linux.kernel.iommu/7837


-Mitch
Will Deacon Jan. 15, 2015, 10:25 a.m. UTC | #3
On Wed, Jan 14, 2015 at 07:42:53PM +0000, Mitchel Humpherys wrote:
> On Tue, Dec 16 2014 at 01:45:27 AM, Will Deacon <will.deacon@arm.com> wrote:
> > On Mon, Dec 15, 2014 at 11:47:23PM +0000, Mitchel Humpherys wrote:
> >> From: Matt Wagantall <mattw@codeaurora.org>
> >> 
> >> It is sometimes necessary to poll a memory-mapped register until its value
> >> satisfies some condition. Introduce a family of convenience macros that do
> >> this. Tight-looping, sleeping, and timing out can all be accomplished using
> >> these macros.
> >> 
> >> Cc: Thierry Reding <thierry.reding@gmail.com>
> >> Cc: Will Deacon <will.deacon@arm.com>
> >> Cc: Arnd Bergmann <arnd@arndb.de>
> >> Cc: Andrew Morton <akpm@linux-foundation.org>
> >> Cc: Robert Elliott <elliott@hp.com>
> >> Signed-off-by: Matt Wagantall <mattw@codeaurora.org>
> >> Signed-off-by: Mitchel Humpherys <mitchelh@codeaurora.org>
> >> ---
> >> v9..10:
> >>   - Actually added the comments mentioned in v8..v9 (doh!)
> >> 
> >> v8..v9:
> >>   - Added note in comments about max sleep time (Rob Elliott)
> >> 
> >> v7..v8:
> >>   - sorted helper macros by size (b, w, l, q)
> >>   - removed some of the more esoteric (or flat-out bogus) helper macros
> >> 
> >> This patch was originally part of a series [1] for adding support for IOMMU
> >> address translations through an ARM SMMU hardware register.  The other
> >> patch in the series (the one that actually uses these macros and implements
> >> said hardware address translations) was Ack'd by the driver maintainer
> >> there (Will Deacon) so I've pulled this patch out to avoid resending an
> >> already Ack'd patch over and over again.
> >> 
> >> In short, please see [1] for previous discussion and the first user of
> >> these macros.
> >> 
> >> Will also acked this patch in [2].  I didn't retain his Ack here because I
> >> added to the macro comments.
> >
> > You can keep the ack, it still looks good to me and I'm not really fussed
> > about the comments.
> >
> > Will
> 
> This hasn't gotten any further comments.  Would someone be willing to
> take it?

If you get an Ack from any of Arnd/Joerg/akpm then I'm happy to take it via
the arm-smmu pull (along with the patch making use of it).

Joerg, would you be ok with that?

Will

> Joerg, maybe you could take this through the IOMMU tree since the first
> user is an IOMMU driver?  Currently we can't move [1] forward because of
> this dependency...
> 
> [1] http://thread.gmane.org/gmane.linux.kernel.iommu/7837
> 
> 
> -Mitch
> 
> -- 
> Qualcomm Innovation Center, Inc.
> The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
> a Linux Foundation Collaborative Project
>
Joerg Roedel Jan. 19, 2015, 12:43 p.m. UTC | #4
On Thu, Jan 15, 2015 at 10:25:11AM +0000, Will Deacon wrote:
> If you get an Ack from any of Arnd/Joerg/akpm then I'm happy to take it via
> the arm-smmu pull (along with the patch making use of it).
> 
> Joerg, would you be ok with that?

I am ok with that, but like to have another Ack for it.


	Joerg
Will Deacon Jan. 19, 2015, 2:40 p.m. UTC | #5
On Mon, Jan 19, 2015 at 12:43:48PM +0000, Joerg Roedel wrote:
> On Thu, Jan 15, 2015 at 10:25:11AM +0000, Will Deacon wrote:
> > If you get an Ack from any of Arnd/Joerg/akpm then I'm happy to take it via
> > the arm-smmu pull (along with the patch making use of it).
> > 
> > Joerg, would you be ok with that?
> 
> I am ok with that, but like to have another Ack for it.

That's fair enough.

Arnd, could you take another look at this patch please? You had some
comments in the past.

Will
Arnd Bergmann Jan. 19, 2015, 3:19 p.m. UTC | #6
On Monday 19 January 2015 14:40:28 Will Deacon wrote:
> On Mon, Jan 19, 2015 at 12:43:48PM +0000, Joerg Roedel wrote:
> > On Thu, Jan 15, 2015 at 10:25:11AM +0000, Will Deacon wrote:
> > > If you get an Ack from any of Arnd/Joerg/akpm then I'm happy to take it via
> > > the arm-smmu pull (along with the patch making use of it).
> > > 
> > > Joerg, would you be ok with that?
> > 
> > I am ok with that, but like to have another Ack for it.
> 
> That's fair enough.
> 
> Arnd, could you take another look at this patch please? You had some
> comments in the past.

No further objections from my side. I don't think it's really needed,
but it seems harmless, so

Acked-by: Arnd Bergmann <arnd@arndb.de>
diff mbox

Patch

diff --git a/include/linux/iopoll.h b/include/linux/iopoll.h
new file mode 100644
index 000000000000..08fd52cdb5a0
--- /dev/null
+++ b/include/linux/iopoll.h
@@ -0,0 +1,144 @@ 
+/*
+ * Copyright (c) 2012-2014 The Linux Foundation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ */
+
+#ifndef _LINUX_IOPOLL_H
+#define _LINUX_IOPOLL_H
+
+#include <linux/kernel.h>
+#include <linux/types.h>
+#include <linux/hrtimer.h>
+#include <linux/delay.h>
+#include <linux/errno.h>
+#include <linux/io.h>
+
+/**
+ * readx_poll_timeout - Periodically poll an address until a condition is met or a timeout occurs
+ * @op: accessor function (takes @addr as its only argument)
+ * @addr: Address to poll
+ * @val: Variable to read the value into
+ * @cond: Break condition (usually involving @val)
+ * @sleep_us: Maximum time to sleep between reads in us (0
+ *            tight-loops).  Should be less than ~20ms since usleep_range
+ *            is used (see Documentation/timers/timers-howto.txt).
+ * @timeout_us: Timeout in us, 0 means never timeout
+ *
+ * Returns 0 on success and -ETIMEDOUT upon a timeout. In either
+ * case, the last read value at @addr is stored in @val. Must not
+ * be called from atomic context if sleep_us or timeout_us are used.
+ *
+ * When available, you'll probably want to use one of the specialized
+ * macros defined below rather than this macro directly.
+ */
+#define readx_poll_timeout(op, addr, val, cond, sleep_us, timeout_us)	\
+({ \
+	ktime_t timeout = ktime_add_us(ktime_get(), timeout_us); \
+	might_sleep_if(sleep_us); \
+	for (;;) { \
+		(val) = op(addr); \
+		if (cond) \
+			break; \
+		if (timeout_us && ktime_compare(ktime_get(), timeout) > 0) { \
+			(val) = op(addr); \
+			break; \
+		} \
+		if (sleep_us) \
+			usleep_range((sleep_us >> 2) + 1, sleep_us); \
+	} \
+	(cond) ? 0 : -ETIMEDOUT; \
+})
+
+/**
+ * readx_poll_timeout_atomic - Periodically poll an address until a condition is met or a timeout occurs
+ * @op: accessor function (takes @addr as its only argument)
+ * @addr: Address to poll
+ * @val: Variable to read the value into
+ * @cond: Break condition (usually involving @val)
+ * @delay_us: Time to udelay between reads in us (0 tight-loops).  Should
+ *            be less than ~10us since udelay is used (see
+ *            Documentation/timers/timers-howto.txt).
+ * @timeout_us: Timeout in us, 0 means never timeout
+ *
+ * Returns 0 on success and -ETIMEDOUT upon a timeout. In either
+ * case, the last read value at @addr is stored in @val.
+ *
+ * When available, you'll probably want to use one of the specialized
+ * macros defined below rather than this macro directly.
+ */
+#define readx_poll_timeout_atomic(op, addr, val, cond, delay_us, timeout_us) \
+({ \
+	ktime_t timeout = ktime_add_us(ktime_get(), timeout_us); \
+	for (;;) { \
+		(val) = op(addr); \
+		if (cond) \
+			break; \
+		if (timeout_us && ktime_compare(ktime_get(), timeout) > 0) { \
+			(val) = op(addr); \
+			break; \
+		} \
+		if (delay_us) \
+			udelay(delay_us);	\
+	} \
+	(cond) ? 0 : -ETIMEDOUT; \
+})
+
+
+#define readb_poll_timeout(addr, val, cond, delay_us, timeout_us) \
+	readx_poll_timeout(readb, addr, val, cond, delay_us, timeout_us)
+
+#define readb_poll_timeout_atomic(addr, val, cond, delay_us, timeout_us) \
+	readx_poll_timeout_atomic(readb, addr, val, cond, delay_us, timeout_us)
+
+#define readw_poll_timeout(addr, val, cond, delay_us, timeout_us) \
+	readx_poll_timeout(readw, addr, val, cond, delay_us, timeout_us)
+
+#define readw_poll_timeout_atomic(addr, val, cond, delay_us, timeout_us) \
+	readx_poll_timeout_atomic(readw, addr, val, cond, delay_us, timeout_us)
+
+#define readl_poll_timeout(addr, val, cond, delay_us, timeout_us) \
+	readx_poll_timeout(readl, addr, val, cond, delay_us, timeout_us)
+
+#define readl_poll_timeout_atomic(addr, val, cond, delay_us, timeout_us) \
+	readx_poll_timeout_atomic(readl, addr, val, cond, delay_us, timeout_us)
+
+#define readq_poll_timeout(addr, val, cond, delay_us, timeout_us) \
+	readx_poll_timeout(readq, addr, val, cond, delay_us, timeout_us)
+
+#define readq_poll_timeout_atomic(addr, val, cond, delay_us, timeout_us) \
+	readx_poll_timeout_atomic(readq, addr, val, cond, delay_us, timeout_us)
+
+#define readb_relaxed_poll_timeout(addr, val, cond, delay_us, timeout_us) \
+	readx_poll_timeout(readb_relaxed, addr, val, cond, delay_us, timeout_us)
+
+#define readb_relaxed_poll_timeout_atomic(addr, val, cond, delay_us, timeout_us) \
+	readx_poll_timeout_atomic(readb_relaxed, addr, val, cond, delay_us, timeout_us)
+
+#define readw_relaxed_poll_timeout(addr, val, cond, delay_us, timeout_us) \
+	readx_poll_timeout(readw_relaxed, addr, val, cond, delay_us, timeout_us)
+
+#define readw_relaxed_poll_timeout_atomic(addr, val, cond, delay_us, timeout_us) \
+	readx_poll_timeout_atomic(readw_relaxed, addr, val, cond, delay_us, timeout_us)
+
+#define readl_relaxed_poll_timeout(addr, val, cond, delay_us, timeout_us) \
+	readx_poll_timeout(readl_relaxed, addr, val, cond, delay_us, timeout_us)
+
+#define readl_relaxed_poll_timeout_atomic(addr, val, cond, delay_us, timeout_us) \
+	readx_poll_timeout_atomic(readl_relaxed, addr, val, cond, delay_us, timeout_us)
+
+#define readq_relaxed_poll_timeout(addr, val, cond, delay_us, timeout_us) \
+	readx_poll_timeout(readq_relaxed, addr, val, cond, delay_us, timeout_us)
+
+#define readq_relaxed_poll_timeout_atomic(addr, val, cond, delay_us, timeout_us) \
+	readx_poll_timeout_atomic(readq_relaxed, addr, val, cond, delay_us, timeout_us)
+
+#endif /* _LINUX_IOPOLL_H */