| Message ID | 20170607232607.26870-2-Jason@zx2c4.com (mailing list archive) |
|---|---|
| State | Not Applicable |
| Delegated to: | Herbert Xu |
| Headers | show |
On 2017-06-08 01:25:55 [+0200], Jason A. Donenfeld wrote: > It's possible that get_random_{u32,u64} is used before the crng has > initialized, in which case, its output might not be cryptographically > secure. For this problem, directly, this patch set is introducing the > *_wait variety of functions, but even with that, there's a subtle issue: > what happens to our batched entropy that was generated before > initialization. Prior to this commit, it'd stick around, supplying bad > numbers. After this commit, we force the entropy to be re-extracted > after each phase of the crng has initialized. > > In order to avoid a race condition with the position counter, we > introduce a simple rwlock for this invalidation. Since it's only during > this awkward transition period, after things are all set up, we stop > using it, so that it doesn't have an impact on performance. > > This should probably be backported to 4.11. > > (Also: adding my copyright to the top. With the patch series from > January, this patch, and then the ones that come after, I think there's > a relevant amount of code in here to add my name to the top.) > > Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com> > Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> I don't understand why lockdep notices this possible deadlock only in RT: | the existing dependency chain (in reverse order) is: | | -> #1 (primary_crng.lock){+.+...}: | lock_acquire+0xb5/0x2b0 | rt_spin_lock+0x46/0x50 | _extract_crng+0x39/0xa0 | extract_crng+0x3a/0x40 | get_random_u64+0x17a/0x200 | cache_random_seq_create+0x51/0x100 | init_cache_random_seq+0x35/0x90 | __kmem_cache_create+0xd3/0x560 | create_boot_cache+0x8c/0xb2 | create_kmalloc_cache+0x54/0x9f | create_kmalloc_caches+0xe3/0xfd | kmem_cache_init+0x14f/0x1f0 | start_kernel+0x1e7/0x3b3 | x86_64_start_reservations+0x2a/0x2c | x86_64_start_kernel+0x13d/0x14c | verify_cpu+0x0/0xfc | | -> #0 (batched_entropy_reset_lock){+.+...}: | __lock_acquire+0x11b4/0x1320 | lock_acquire+0xb5/0x2b0 | rt_write_lock+0x26/0x40 | rt_write_lock_irqsave+0x9/0x10 | invalidate_batched_entropy+0x28/0xb0 | crng_fast_load+0xb5/0xe0 | add_interrupt_randomness+0x16c/0x1a0 | irq_thread+0x15c/0x1e0 | kthread+0x112/0x150 | ret_from_fork+0x31/0x40 We have the path add_interrupt_randomness() -> crng_fast_load() which take primary_crng.lock -> batched_entropy_reset_lock and we have get_random_uXX() -> extract_crng() which take the locks in opposite order: batched_entropy_reset_lock -> crng->lock however batched_entropy_reset_lock() is only taken if "crng_init < 2" so it is possible RT hits this constraint more reliably. > --- > drivers/char/random.c | 37 +++++++++++++++++++++++++++++++++++++ > 1 file changed, 37 insertions(+) > > diff --git a/drivers/char/random.c b/drivers/char/random.c > index a561f0c2f428..d35da1603e12 100644 > --- a/drivers/char/random.c > +++ b/drivers/char/random.c > @@ -1,6 +1,9 @@ > /* > * random.c -- A strong random number generator > * > + * Copyright (C) 2017 Jason A. Donenfeld <Jason@zx2c4.com>. All > + * Rights Reserved. > + * > * Copyright Matt Mackall <mpm@selenic.com>, 2003, 2004, 2005 > * > * Copyright Theodore Ts'o, 1994, 1995, 1996, 1997, 1998, 1999. All > @@ -762,6 +765,8 @@ static DECLARE_WAIT_QUEUE_HEAD(crng_init_wait); > static struct crng_state **crng_node_pool __read_mostly; > #endif > > +static void invalidate_batched_entropy(void); > + > static void crng_initialize(struct crng_state *crng) > { > int i; > @@ -799,6 +804,7 @@ static int crng_fast_load(const char *cp, size_t len) > cp++; crng_init_cnt++; len--; > } > if (crng_init_cnt >= CRNG_INIT_CNT_THRESH) { > + invalidate_batched_entropy(); > crng_init = 1; > wake_up_interruptible(&crng_init_wait); > pr_notice("random: fast init done\n"); > @@ -836,6 +842,7 @@ static void crng_reseed(struct crng_state *crng, struct entropy_store *r) > memzero_explicit(&buf, sizeof(buf)); > crng->init_time = jiffies; > if (crng == &primary_crng && crng_init < 2) { > + invalidate_batched_entropy(); > crng_init = 2; > process_random_ready_list(); > wake_up_interruptible(&crng_init_wait); > @@ -2023,6 +2030,7 @@ struct batched_entropy { > }; > unsigned int position; > }; > +static rwlock_t batched_entropy_reset_lock = __RW_LOCK_UNLOCKED(batched_entropy_reset_lock); > > /* > * Get a random word for internal kernel use only. The quality of the random > @@ -2033,6 +2041,8 @@ static DEFINE_PER_CPU(struct batched_entropy, batched_entropy_u64); > u64 get_random_u64(void) > { > u64 ret; > + const bool use_lock = READ_ONCE(crng_init) < 2; > + unsigned long flags = 0; > struct batched_entropy *batch; > > #if BITS_PER_LONG == 64 > @@ -2045,11 +2055,15 @@ u64 get_random_u64(void) > #endif > > batch = &get_cpu_var(batched_entropy_u64); > + if (use_lock) > + read_lock_irqsave(&batched_entropy_reset_lock, flags); > if (batch->position % ARRAY_SIZE(batch->entropy_u64) == 0) { > extract_crng((u8 *)batch->entropy_u64); > batch->position = 0; > } > ret = batch->entropy_u64[batch->position++]; > + if (use_lock) > + read_unlock_irqrestore(&batched_entropy_reset_lock, flags); > put_cpu_var(batched_entropy_u64); > return ret; > } > @@ -2059,22 +2073,45 @@ static DEFINE_PER_CPU(struct batched_entropy, batched_entropy_u32); > u32 get_random_u32(void) > { > u32 ret; > + const bool use_lock = READ_ONCE(crng_init) < 2; > + unsigned long flags = 0; > struct batched_entropy *batch; > > if (arch_get_random_int(&ret)) > return ret; > > batch = &get_cpu_var(batched_entropy_u32); > + if (use_lock) > + read_lock_irqsave(&batched_entropy_reset_lock, flags); > if (batch->position % ARRAY_SIZE(batch->entropy_u32) == 0) { > extract_crng((u8 *)batch->entropy_u32); > batch->position = 0; > } > ret = batch->entropy_u32[batch->position++]; > + if (use_lock) > + read_unlock_irqrestore(&batched_entropy_reset_lock, flags); > put_cpu_var(batched_entropy_u32); > return ret; > } > EXPORT_SYMBOL(get_random_u32); > > +/* It's important to invalidate all potential batched entropy that might > + * be stored before the crng is initialized, which we can do lazily by > + * simply resetting the counter to zero so that it's re-extracted on the > + * next usage. */ > +static void invalidate_batched_entropy(void) > +{ > + int cpu; > + unsigned long flags; > + > + write_lock_irqsave(&batched_entropy_reset_lock, flags); > + for_each_possible_cpu (cpu) { > + per_cpu_ptr(&batched_entropy_u32, cpu)->position = 0; > + per_cpu_ptr(&batched_entropy_u64, cpu)->position = 0; > + } > + write_unlock_irqrestore(&batched_entropy_reset_lock, flags); > +} > + > /** > * randomize_page - Generate a random, page aligned address > * @start: The smallest acceptable address the caller will take. > -- > 2.13.0 Sebastian
There's a potential race that I fixed in my v5 of that patch set, but Ted only took v4, and for whatever reason has been to busy to submit the additional patch I already posted showing the diff between v4&v5. Hopefully he actually gets around to it and sends this for the next rc. Here it is: https://patchwork.kernel.org/patch/9774563/
On 2017-06-15 00:33:12 [+0200], Jason A. Donenfeld wrote: > There's a potential race that I fixed in my v5 of that patch set, but > Ted only took v4, and for whatever reason has been to busy to submit > the additional patch I already posted showing the diff between v4&v5. > Hopefully he actually gets around to it and sends this for the next > rc. Here it is: > > https://patchwork.kernel.org/patch/9774563/ So you replace "crng_init < 2" with use_lock instead. That is not what I am talking about. Again: add_interrupt_randomness() -> crng_fast_load() spin_trylock_irqsave(&primary_crng.lock, ) -> invalidate_batched_entropy() write_lock_irqsave(&batched_entropy_reset_lock, ); in that order while the code path get_random_uXX() read_lock_irqsave(&batched_entropy_reset_lock, ); -> extract_crng() -> _extract_crng() spin_lock_irqsave(&crng->lock, ); which allocates the same lock in opposite order. That means T1 T2 crng_fast_load() get_random_u64() extract_crng() *dead lock* invalidate_batched_entropy() _extract_crng() So T1 waits for batched_entropy_reset_lock holding primary_crng.lock and T2 waits for primary_crng.lock holding batched_entropy_reset_lock. Sebastian
On Fri, Jun 16, 2017 at 10:31 AM, Sebastian Andrzej Siewior
<sebastian@breakpoint.cc> wrote:
> am talking about. Again:
I actually figured that out myself after sending the initial email, so
then I wrote a follow-up patch which I attached to this thread. You
should have received it. Can you take a look?
https://lkml.org/lkml/2017/6/14/942
On 2017-06-16 14:12:42 [+0200], Jason A. Donenfeld wrote: > I actually figured that out myself after sending the initial email, so > then I wrote a follow-up patch which I attached to this thread. You > should have received it. Can you take a look? replied to the patch. Sebastian
diff --git a/drivers/char/random.c b/drivers/char/random.c index a561f0c2f428..d35da1603e12 100644 --- a/drivers/char/random.c +++ b/drivers/char/random.c @@ -1,6 +1,9 @@ /* * random.c -- A strong random number generator * + * Copyright (C) 2017 Jason A. Donenfeld <Jason@zx2c4.com>. All + * Rights Reserved. + * * Copyright Matt Mackall <mpm@selenic.com>, 2003, 2004, 2005 * * Copyright Theodore Ts'o, 1994, 1995, 1996, 1997, 1998, 1999. All @@ -762,6 +765,8 @@ static DECLARE_WAIT_QUEUE_HEAD(crng_init_wait); static struct crng_state **crng_node_pool __read_mostly; #endif +static void invalidate_batched_entropy(void); + static void crng_initialize(struct crng_state *crng) { int i; @@ -799,6 +804,7 @@ static int crng_fast_load(const char *cp, size_t len) cp++; crng_init_cnt++; len--; } if (crng_init_cnt >= CRNG_INIT_CNT_THRESH) { + invalidate_batched_entropy(); crng_init = 1; wake_up_interruptible(&crng_init_wait); pr_notice("random: fast init done\n"); @@ -836,6 +842,7 @@ static void crng_reseed(struct crng_state *crng, struct entropy_store *r) memzero_explicit(&buf, sizeof(buf)); crng->init_time = jiffies; if (crng == &primary_crng && crng_init < 2) { + invalidate_batched_entropy(); crng_init = 2; process_random_ready_list(); wake_up_interruptible(&crng_init_wait); @@ -2023,6 +2030,7 @@ struct batched_entropy { }; unsigned int position; }; +static rwlock_t batched_entropy_reset_lock = __RW_LOCK_UNLOCKED(batched_entropy_reset_lock); /* * Get a random word for internal kernel use only. The quality of the random @@ -2033,6 +2041,8 @@ static DEFINE_PER_CPU(struct batched_entropy, batched_entropy_u64); u64 get_random_u64(void) { u64 ret; + const bool use_lock = READ_ONCE(crng_init) < 2; + unsigned long flags = 0; struct batched_entropy *batch; #if BITS_PER_LONG == 64 @@ -2045,11 +2055,15 @@ u64 get_random_u64(void) #endif batch = &get_cpu_var(batched_entropy_u64); + if (use_lock) + read_lock_irqsave(&batched_entropy_reset_lock, flags); if (batch->position % ARRAY_SIZE(batch->entropy_u64) == 0) { extract_crng((u8 *)batch->entropy_u64); batch->position = 0; } ret = batch->entropy_u64[batch->position++]; + if (use_lock) + read_unlock_irqrestore(&batched_entropy_reset_lock, flags); put_cpu_var(batched_entropy_u64); return ret; } @@ -2059,22 +2073,45 @@ static DEFINE_PER_CPU(struct batched_entropy, batched_entropy_u32); u32 get_random_u32(void) { u32 ret; + const bool use_lock = READ_ONCE(crng_init) < 2; + unsigned long flags = 0; struct batched_entropy *batch; if (arch_get_random_int(&ret)) return ret; batch = &get_cpu_var(batched_entropy_u32); + if (use_lock) + read_lock_irqsave(&batched_entropy_reset_lock, flags); if (batch->position % ARRAY_SIZE(batch->entropy_u32) == 0) { extract_crng((u8 *)batch->entropy_u32); batch->position = 0; } ret = batch->entropy_u32[batch->position++]; + if (use_lock) + read_unlock_irqrestore(&batched_entropy_reset_lock, flags); put_cpu_var(batched_entropy_u32); return ret; } EXPORT_SYMBOL(get_random_u32); +/* It's important to invalidate all potential batched entropy that might + * be stored before the crng is initialized, which we can do lazily by + * simply resetting the counter to zero so that it's re-extracted on the + * next usage. */ +static void invalidate_batched_entropy(void) +{ + int cpu; + unsigned long flags; + + write_lock_irqsave(&batched_entropy_reset_lock, flags); + for_each_possible_cpu (cpu) { + per_cpu_ptr(&batched_entropy_u32, cpu)->position = 0; + per_cpu_ptr(&batched_entropy_u64, cpu)->position = 0; + } + write_unlock_irqrestore(&batched_entropy_reset_lock, flags); +} + /** * randomize_page - Generate a random, page aligned address * @start: The smallest acceptable address the caller will take.
It's possible that get_random_{u32,u64} is used before the crng has initialized, in which case, its output might not be cryptographically secure. For this problem, directly, this patch set is introducing the *_wait variety of functions, but even with that, there's a subtle issue: what happens to our batched entropy that was generated before initialization. Prior to this commit, it'd stick around, supplying bad numbers. After this commit, we force the entropy to be re-extracted after each phase of the crng has initialized. In order to avoid a race condition with the position counter, we introduce a simple rwlock for this invalidation. Since it's only during this awkward transition period, after things are all set up, we stop using it, so that it doesn't have an impact on performance. This should probably be backported to 4.11. (Also: adding my copyright to the top. With the patch series from January, this patch, and then the ones that come after, I think there's a relevant amount of code in here to add my name to the top.) Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> --- drivers/char/random.c | 37 +++++++++++++++++++++++++++++++++++++ 1 file changed, 37 insertions(+)