Skip to contentSkip to conversation

ARM: omap2: throw the die id into the entropy pool

PATCH Superseded Linus Walleij 3 participants 5 messages Sep 5, 2013 – Oct 3, 2013

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 ARM: omap2: throw the die id into the entropy pool

Atleast eight bytes of this number are totally unique for the device it seems, so this is a perfect candidate for feeding the entropy pool. One byte more or less of constants does not matter so feed in the entire OID struct.

---
ChangeLog v1->v2:
- Use omap_device_initcall() to avoid calling on non-OMAPs.

---
 arch/arm/mach-omap2/id.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)
arch/arm/mach-omap2/id.c +12 -0
index 2dc62a2..9260d09 100644
@@ -18,6 +18,7 @@
 #include <linux/kernel.h>
 #include <linux/init.h>
 #include <linux/io.h>
+#include <linux/random.h>
 #include <linux/slab.h>
 
 #ifdef CONFIG_SOC_BUS
@@ -130,6 +131,17 @@ void omap_get_die_id(struct omap_die_id *odi)
 	odi->id_3 = read_tap_reg(OMAP_TAP_DIE_ID_3);
 }
 
+static int __init omap_feed_randpool(void)
+{
+	struct omap_die_id odi;
+
+	/* Throw the die ID into the entropy pool at boot */
+	omap_get_die_id(&odi);
+	add_device_randomness(&odi, sizeof(odi));
+	return 0;
+}
+omap_device_initcall(omap_feed_randpool);
+
 void __init omap2xxx_check_revision(void)
 {
 	int i, j;
Signature · 1 line
1.8.1.4
#1 lore raw

PATCH v2 1/1 ARM: omap2: throw the die id into the entropy pool

Hi Linus,

Quoted 9 lines · Linus Walleij

Atleast eight bytes of this number are totally unique for the device it seems, so this is a perfect candidate for feeding the entropy pool. One byte more or less of constants does not matter so feed in the entire OID struct.

Heh, that function name "add_device_randomness()" is a bit misleading. It's not actually intended to add "randomness": from drivers/char/random.c:

/*
* Add device- or boot-specific data to the input and nonblocking
* pools to help initialize them to unique values.
*
* None of this adds any entropy, it is meant to avoid the
* problem of the nonblocking pool having similar initial state
* across largely identical devices.
*/

But of course the function name is not your fault :-) The entropy count isn't increased by this, so:

Thanks Linus.

- Paul

#2 lore raw

PATCH v2 1/1 ARM: omap2: throw the die id into the entropy pool

By the way, if anyone wants to expand on Linus' patch for OMAP, there are several other register fields that could be mixed in, which, based on their names, might vary on a per-chip basis. For example, on OMAP4, the CONTROL_STD_FUSE_OPP* and CONTROL_DPLL_NWELL_TRIM* registers.

- Paul

#3 lore raw

PATCH v2 1/1 ARM: omap2: throw the die id into the entropy pool

In reply to Paul Walmsley
Quoted 3 lines · Paul Walmsley

Heh, that function name "add_device_randomness()" is a bit misleading. It's not actually intended to add "randomness": from drivers/char/random.c:

Yeah you're right...

Tony feel free to edit the commit message when applying.

Quoted 3 lines

* None of this adds any entropy, it is meant to avoid the
* problem of the nonblocking pool having similar initial state
* across largely identical devices.

It's noble enough, just a few years back I ran into the problem where all boards in a test farm came up with the same ethernet MAC address due to the initialization of the nonblocking pool being constant.

Yours,
Linus Walleij

maintainer
#4 lore raw

PATCH v2 1/1 ARM: omap2: throw the die id into the entropy pool

Quoted 10 lines · Linus Walleij
Quoted 3 lines · Paul Walmsley

Heh, that function name "add_device_randomness()" is a bit misleading. It's not actually intended to add "randomness": from drivers/char/random.c:

Yeah you're right...

Tony feel free to edit the commit message when applying.

Quoted 8 lines
Quoted 3 lines

* None of this adds any entropy, it is meant to avoid the
* problem of the nonblocking pool having similar initial state
* across largely identical devices.

It's noble enough, just a few years back I ran into the problem where all boards in a test farm came up with the same ethernet MAC address due to the initialization of the nonblocking pool being constant.

Thanks, I'll apply this into omap-for-v3.13/fixes-not-urgent with updated comments.

Tony