This reintroduces the entropy sampling of the AB3100 IRQ as the
IRQF_SAMPLE_RANDOM is going out according to the feature removal
schedule. I'm trying to do this the right way then, so CC:ing some
random people for a quick review. We add entropy for interrupt
events in the AB3100 which are truly random in nature, like
external cables being connected, voltages on batteries dropping
below certain ranges, ADC triggers or overheating.
Signed-off-by: Linus Walleij <[email protected]>
Cc: Robin Getz <[email protected]>
Cc: Matt Mackall <[email protected]>
---
drivers/mfd/ab3100-core.c | 17 +++++++++--------
1 files changed, 9 insertions(+), 8 deletions(-)
diff --git a/drivers/mfd/ab3100-core.c b/drivers/mfd/ab3100-core.c
index aa3824a..a2ce3b6 100644
--- a/drivers/mfd/ab3100-core.c
+++ b/drivers/mfd/ab3100-core.c
@@ -1,5 +1,5 @@
/*
- * Copyright (C) 2007-2009 ST-Ericsson
+ * Copyright (C) 2007-2010 ST-Ericsson
* License terms: GNU General Public License (GPL) version 2
* Low-level core for exclusive access to the AB3100 IC on the I2C bus
* and some basic chip-configuration.
@@ -14,6 +14,7 @@
#include <linux/platform_device.h>
#include <linux/device.h>
#include <linux/interrupt.h>
+#include <linux/random.h>
#include <linux/debugfs.h>
#include <linux/seq_file.h>
#include <linux/uaccess.h>
@@ -376,6 +377,8 @@ static irqreturn_t ab3100_irq_handler(int irq, void *data)
u32 fatevent;
int err;
+ add_interrupt_randomness(irq);
+
err = ab3100_get_register_page_interruptible(ab3100, AB3100_EVENTA1,
event_regs, 3);
if (err)
@@ -720,10 +723,7 @@ static struct platform_device ab3100_##devname##_device = { \
.id = -1, \
}
-/*
- * This lists all the subdevices and corresponding register
- * ranges.
- */
+/* This lists all the subdevices */
AB3100_DEVICE(dac, "ab3100-dac");
AB3100_DEVICE(leds, "ab3100-leds");
AB3100_DEVICE(power, "ab3100-power");
@@ -889,10 +889,11 @@ static int __init ab3100_probe(struct i2c_client *client,
if (err)
goto exit_no_setup;
- /* This real unpredictable IRQ is of course sampled for entropy */
err = request_threaded_irq(client->irq, NULL, ab3100_irq_handler,
- IRQF_ONESHOT,
- "ab3100-core", ab3100);
+ IRQF_ONESHOT, "ab3100-core", ab3100);
+ /* This real unpredictable IRQ is of course sampled for entropy */
+ rand_initialize_irq(client->irq);
+
if (err)
goto exit_no_irq;
--
1.6.5.2
On Sat, 2010-01-09 at 01:38 +0100, Linus Walleij wrote:
> This reintroduces the entropy sampling of the AB3100 IRQ as the
> IRQF_SAMPLE_RANDOM is going out according to the feature removal
> schedule. I'm trying to do this the right way then, so CC:ing some
> random people for a quick review. We add entropy for interrupt
> events in the AB3100 which are truly random in nature, like
> external cables being connected, voltages on batteries dropping
> below certain ranges, ADC triggers or overheating.
>
> Signed-off-by: Linus Walleij <[email protected]>
> Cc: Robin Getz <[email protected]>
> Cc: Matt Mackall <[email protected]>
> ---
> drivers/mfd/ab3100-core.c | 17 +++++++++--------
> 1 files changed, 9 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/mfd/ab3100-core.c b/drivers/mfd/ab3100-core.c
> index aa3824a..a2ce3b6 100644
> --- a/drivers/mfd/ab3100-core.c
> +++ b/drivers/mfd/ab3100-core.c
> @@ -1,5 +1,5 @@
> /*
> - * Copyright (C) 2007-2009 ST-Ericsson
> + * Copyright (C) 2007-2010 ST-Ericsson
> * License terms: GNU General Public License (GPL) version 2
> * Low-level core for exclusive access to the AB3100 IC on the I2C bus
> * and some basic chip-configuration.
> @@ -14,6 +14,7 @@
> #include <linux/platform_device.h>
> #include <linux/device.h>
> #include <linux/interrupt.h>
> +#include <linux/random.h>
> #include <linux/debugfs.h>
> #include <linux/seq_file.h>
> #include <linux/uaccess.h>
> @@ -376,6 +377,8 @@ static irqreturn_t ab3100_irq_handler(int irq, void *data)
> u32 fatevent;
> int err;
>
> + add_interrupt_randomness(irq);
> +
> err = ab3100_get_register_page_interruptible(ab3100, AB3100_EVENTA1,
> event_regs, 3);
> if (err)
> @@ -720,10 +723,7 @@ static struct platform_device ab3100_##devname##_device = { \
> .id = -1, \
> }
>
> -/*
> - * This lists all the subdevices and corresponding register
> - * ranges.
> - */
> +/* This lists all the subdevices */
> AB3100_DEVICE(dac, "ab3100-dac");
> AB3100_DEVICE(leds, "ab3100-leds");
> AB3100_DEVICE(power, "ab3100-power");
> @@ -889,10 +889,11 @@ static int __init ab3100_probe(struct i2c_client *client,
> if (err)
> goto exit_no_setup;
>
> - /* This real unpredictable IRQ is of course sampled for entropy */
> err = request_threaded_irq(client->irq, NULL, ab3100_irq_handler,
> - IRQF_ONESHOT,
> - "ab3100-core", ab3100);
> + IRQF_ONESHOT, "ab3100-core", ab3100);
> + /* This real unpredictable IRQ is of course sampled for entropy */
> + rand_initialize_irq(client->irq);
> +
> if (err)
> goto exit_no_irq;
Acked-by: Matt Mackall <[email protected]>
--
http://selenic.com : development and support for Mercurial and Linux
2010/1/11 Matt Mackall <[email protected]>:
> Acked-by: Matt Mackall <[email protected]>
Sam will you take this patch then?
Yours,
Linus Walleij
Hi Linus,
On Sun, Jan 17, 2010 at 08:18:58PM +0100, Linus Walleij wrote:
> 2010/1/11 Matt Mackall <[email protected]>:
>
> > Acked-by: Matt Mackall <[email protected]>
>
> Sam will you take this patch then?
Applied, thanks a lot.
Cheers,
Samuel.
> Yours,
> Linus Walleij
--
Intel Open Source Technology Centre
http://oss.intel.com/
Hi Linus,
On Sun, Jan 17, 2010 at 08:18:58PM +0100, Linus Walleij wrote:
> 2010/1/11 Matt Mackall <[email protected]>:
>
> > Acked-by: Matt Mackall <[email protected]>
>
> Sam will you take this patch then?
I had to change the ab3100 Kconfig entry as those random.c symbols are not
exported. The ab3100 driver is now a bool, depending on I2C=y.
Matt, would it make sense to export those symbols ?
Cheers,
Samuel.
> Yours,
> Linus Walleij
--
Intel Open Source Technology Centre
http://oss.intel.com/
2010/1/19 Samuel Ortiz <[email protected]>:
> I had to change the ab3100 Kconfig entry as those random.c symbols are not
> exported. The ab3100 driver is now a bool, depending on I2C=y.
Fine with me. We mainly need it to be a module when developing initial
support anyway. If you activate stuff like regulators it will be required to
compile in.
Acked-by: Linus Walleij <[email protected]>
Linus Walleij
On Tue, 2010-01-19 at 12:12 +0100, Samuel Ortiz wrote:
> Hi Linus,
>
> On Sun, Jan 17, 2010 at 08:18:58PM +0100, Linus Walleij wrote:
> > 2010/1/11 Matt Mackall <[email protected]>:
> >
> > > Acked-by: Matt Mackall <[email protected]>
> >
> > Sam will you take this patch then?
> I had to change the ab3100 Kconfig entry as those random.c symbols are not
> exported. The ab3100 driver is now a bool, depending on I2C=y.
>
> Matt, would it make sense to export those symbols ?
Let's not for now, I plan to change this API.
--
http://selenic.com : development and support for Mercurial and Linux