kernel/arch/s390/include/asm/archrandom.h, branch linux-6.2.y

Merge tag 'random-6.0-rc1-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/crng/random

2022-08-03T00:31:35Z

Pull random number generator updates from Jason Donenfeld: "Though there's been a decent amount of RNG-related development during this last cycle, not all of it is coming through this tree, as this cycle saw a shift toward tackling early boot time seeding issues, which took place in other trees as well. Here's a summary of the various patches: - The CONFIG_ARCH_RANDOM .config option and the "nordrand" boot option have been removed, as they overlapped with the more widely supported and more sensible options, CONFIG_RANDOM_TRUST_CPU and "random.trust_cpu". This change allowed simplifying a bit of arch code. - x86's RDRAND boot time test has been made a bit more robust, with RDRAND disabled if it's clearly producing bogus results. This would be a tip.git commit, technically, but I took it through random.git to avoid a large merge conflict. - The RNG has long since mixed in a timestamp very early in boot, on the premise that a computer that does the same things, but does so starting at different points in wall time, could be made to still produce a different RNG state. Unfortunately, the clock isn't set early in boot on all systems, so now we mix in that timestamp when the time is actually set. - User Mode Linux now uses the host OS's getrandom() syscall to generate a bootloader RNG seed and later on treats getrandom() as the platform's RDRAND-like faculty. - The arch_get_random_{seed_,}_long() family of functions is now arch_get_random_{seed_,}_longs(), which enables certain platforms, such as s390, to exploit considerable performance advantages from requesting multiple CPU random numbers at once, while at the same time compiling down to the same code as before on platforms like x86. - A small cleanup changing a cmpxchg() into a try_cmpxchg(), from Uros. - A comment spelling fix" More info about other random number changes that come in through various architecture trees in the full commentary in the pull request: https://lore.kernel.org/all/20220731232428.2219258-1-Jason@zx2c4.com/ * tag 'random-6.0-rc1-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/crng/random: random: correct spelling of "overwrites" random: handle archrandom with multiple longs um: seed rng using host OS rng random: use try_cmpxchg in _credit_init_bits timekeeping: contribute wall clock to rng on time change x86/rdrand: Remove "nordrand" flag in favor of "random.trust_cpu" random: remove CONFIG_ARCH_RANDOM

random: handle archrandom with multiple longs

2022-07-25T11:26:14Z

The archrandom interface was originally designed for x86, which supplies RDRAND/RDSEED for receiving random words into registers, resulting in one function to generate an int and another to generate a long. However, other architectures don't follow this. On arm64, the SMCCC TRNG interface can return between one and three longs. On s390, the CPACF TRNG interface can return arbitrary amounts, with four longs having the same cost as one. On UML, the os_getrandom() interface can return arbitrary amounts. So change the api signature to take a "max_longs" parameter designating the maximum number of longs requested, and then return the number of longs generated. Since callers need to check this return value and loop anyway, each arch implementation does not bother implementing its own loop to try again to fill the maximum number of longs. Additionally, all existing callers pass in a constant max_longs parameter. Taken together, these two things mean that the codegen doesn't really change much for one-word-at-a-time platforms, while performance is greatly improved on platforms such as s390. Acked-by: Heiko Carstens Acked-by: Catalin Marinas Acked-by: Mark Rutland Acked-by: Michael Ellerman Acked-by: Borislav Petkov Signed-off-by: Jason A. Donenfeld

s390/archrandom: prevent CPACF trng invocations in interrupt context

2022-07-21T18:03:03Z

This patch slightly reworks the s390 arch_get_random_seed_{int,long} implementation: Make sure the CPACF trng instruction is never called in any interrupt context. This is done by adding an additional condition in_task(). Justification: There are some constrains to satisfy for the invocation of the arch_get_random_seed_{int,long}() functions: - They should provide good random data during kernel initialization. - They should not be called in interrupt context as the TRNG instruction is relatively heavy weight and may for example make some network loads cause to timeout and buck. However, it was not clear what kind of interrupt context is exactly encountered during kernel init or network traffic eventually calling arch_get_random_seed_long(). After some days of investigations it is clear that the s390 start_kernel function is not running in any interrupt context and so the trng is called: Jul 11 18:33:39 t35lp54 kernel: [<00000001064e90ca>] arch_get_random_seed_long.part.0+0x32/0x70 Jul 11 18:33:39 t35lp54 kernel: [<000000010715f246>] random_init+0xf6/0x238 Jul 11 18:33:39 t35lp54 kernel: [<000000010712545c>] start_kernel+0x4a4/0x628 Jul 11 18:33:39 t35lp54 kernel: [<000000010590402a>] startup_continue+0x2a/0x40 The condition in_task() is true and the CPACF trng provides random data during kernel startup. The network traffic however, is more difficult. A typical call stack looks like this: Jul 06 17:37:07 t35lp54 kernel: [<000000008b5600fc>] extract_entropy.constprop.0+0x23c/0x240 Jul 06 17:37:07 t35lp54 kernel: [<000000008b560136>] crng_reseed+0x36/0xd8 Jul 06 17:37:07 t35lp54 kernel: [<000000008b5604b8>] crng_make_state+0x78/0x340 Jul 06 17:37:07 t35lp54 kernel: [<000000008b5607e0>] _get_random_bytes+0x60/0xf8 Jul 06 17:37:07 t35lp54 kernel: [<000000008b56108a>] get_random_u32+0xda/0x248 Jul 06 17:37:07 t35lp54 kernel: [<000000008aefe7a8>] kfence_guarded_alloc+0x48/0x4b8 Jul 06 17:37:07 t35lp54 kernel: [<000000008aeff35e>] __kfence_alloc+0x18e/0x1b8 Jul 06 17:37:07 t35lp54 kernel: [<000000008aef7f10>] __kmalloc_node_track_caller+0x368/0x4d8 Jul 06 17:37:07 t35lp54 kernel: [<000000008b611eac>] kmalloc_reserve+0x44/0xa0 Jul 06 17:37:07 t35lp54 kernel: [<000000008b611f98>] __alloc_skb+0x90/0x178 Jul 06 17:37:07 t35lp54 kernel: [<000000008b6120dc>] __napi_alloc_skb+0x5c/0x118 Jul 06 17:37:07 t35lp54 kernel: [<000000008b8f06b4>] qeth_extract_skb+0x13c/0x680 Jul 06 17:37:07 t35lp54 kernel: [<000000008b8f6526>] qeth_poll+0x256/0x3f8 Jul 06 17:37:07 t35lp54 kernel: [<000000008b63d76e>] __napi_poll.constprop.0+0x46/0x2f8 Jul 06 17:37:07 t35lp54 kernel: [<000000008b63dbec>] net_rx_action+0x1cc/0x408 Jul 06 17:37:07 t35lp54 kernel: [<000000008b937302>] __do_softirq+0x132/0x6b0 Jul 06 17:37:07 t35lp54 kernel: [<000000008abf46ce>] __irq_exit_rcu+0x13e/0x170 Jul 06 17:37:07 t35lp54 kernel: [<000000008abf531a>] irq_exit_rcu+0x22/0x50 Jul 06 17:37:07 t35lp54 kernel: [<000000008b922506>] do_io_irq+0xe6/0x198 Jul 06 17:37:07 t35lp54 kernel: [<000000008b935826>] io_int_handler+0xd6/0x110 Jul 06 17:37:07 t35lp54 kernel: [<000000008b9358a6>] psw_idle_exit+0x0/0xa Jul 06 17:37:07 t35lp54 kernel: ([<000000008ab9c59a>] arch_cpu_idle+0x52/0xe0) Jul 06 17:37:07 t35lp54 kernel: [<000000008b933cfe>] default_idle_call+0x6e/0xd0 Jul 06 17:37:07 t35lp54 kernel: [<000000008ac59f4e>] do_idle+0xf6/0x1b0 Jul 06 17:37:07 t35lp54 kernel: [<000000008ac5a28e>] cpu_startup_entry+0x36/0x40 Jul 06 17:37:07 t35lp54 kernel: [<000000008abb0d90>] smp_start_secondary+0x148/0x158 Jul 06 17:37:07 t35lp54 kernel: [<000000008b935b9e>] restart_int_handler+0x6e/0x90 which confirms that the call is in softirq context. So in_task() covers exactly the cases where we want to have CPACF trng called: not in nmi, not in hard irq, not in soft irq but in normal task context and during kernel init. Signed-off-by: Harald Freudenberger Acked-by: Jason A. Donenfeld Reviewed-by: Juergen Christ Link: https://lore.kernel.org/r/20220713131721.257907-1-freude@linux.ibm.com Fixes: e4f74400308c ("s390/archrandom: simplify back to earlier design and initialize earlier") [agordeev@linux.ibm.com changed desc, added Fixes and Link, removed -stable] Signed-off-by: Alexander Gordeev

random: remove CONFIG_ARCH_RANDOM

2022-07-18T13:03:37Z

When RDRAND was introduced, there was much discussion on whether it should be trusted and how the kernel should handle that. Initially, two mechanisms cropped up, CONFIG_ARCH_RANDOM, a compile time switch, and "nordrand", a boot-time switch. Later the thinking evolved. With a properly designed RNG, using RDRAND values alone won't harm anything, even if the outputs are malicious. Rather, the issue is whether those values are being *trusted* to be good or not. And so a new set of options were introduced as the real ones that people use -- CONFIG_RANDOM_TRUST_CPU and "random.trust_cpu". With these options, RDRAND is used, but it's not always credited. So in the worst case, it does nothing, and in the best case, maybe it helps. Along the way, CONFIG_ARCH_RANDOM's meaning got sort of pulled into the center and became something certain platforms force-select. The old options don't really help with much, and it's a bit odd to have special handling for these instructions when the kernel can deal fine with the existence or untrusted existence or broken existence or non-existence of that CPU capability. Simplify the situation by removing CONFIG_ARCH_RANDOM and using the ordinary asm-generic fallback pattern instead, keeping the two options that are actually used. For now it leaves "nordrand" for now, as the removal of that will take a different route. Acked-by: Michael Ellerman Acked-by: Catalin Marinas Acked-by: Borislav Petkov Acked-by: Heiko Carstens Acked-by: Greg Kroah-Hartman Signed-off-by: Jason A. Donenfeld

s390/archrandom: simplify back to earlier design and initialize earlier

2022-06-30T17:40:36Z

s390x appears to present two RNG interfaces: - a "TRNG" that gathers entropy using some hardware function; and - a "DRBG" that takes in a seed and expands it. Previously, the TRNG was wired up to arch_get_random_{long,int}(), but it was observed that this was being called really frequently, resulting in high overhead. So it was changed to be wired up to arch_get_random_ seed_{long,int}(), which was a reasonable decision. Later on, the DRBG was then wired up to arch_get_random_{long,int}(), with a complicated buffer filling thread, to control overhead and rate. Fortunately, none of the performance issues matter much now. The RNG always attempts to use arch_get_random_seed_{long,int}() first, which means a complicated implementation of arch_get_random_{long,int}() isn't really valuable or useful to have around. And it's only used when reseeding, which means it won't hit the high throughput complications that were faced before. So this commit returns to an earlier design of just calling the TRNG in arch_get_random_seed_{long,int}(), and returning false in arch_get_ random_{long,int}(). Part of what makes the simplification possible is that the RNG now seeds itself using the TRNG at bootup. But this only works if the TRNG is detected early in boot, before random_init() is called. So this commit also causes that check to happen in setup_arch(). Cc: stable@vger.kernel.org Cc: Harald Freudenberger Cc: Ingo Franzki Cc: Juergen Christ Cc: Heiko Carstens Signed-off-by: Jason A. Donenfeld Link: https://lore.kernel.org/r/20220610222023.378448-1-Jason@zx2c4.com Reviewed-by: Harald Freudenberger Acked-by: Heiko Carstens Signed-off-by: Alexander Gordeev

s390/crypto: add arch_get_random_long() support

2020-12-10T20:02:08Z

The random longs to be pulled by arch_get_random_long() are prepared in an 4K buffer which is filled from the NIST 800-90 compliant s390 drbg. By default the random long buffer is refilled 256 times before the drbg itself needs a reseed. The reseed of the drbg is done with 32 bytes fetched from the high quality (but slow) trng which is assumed to deliver 100% entropy. So the 32 * 8 = 256 bits of entropy are spread over 256 * 4KB = 1MB serving 131072 arch_get_random_long() invocations before reseeded. How often the 4K random long buffer is refilled with the drbg before the drbg is reseeded can be adjusted. There is a module parameter 's390_arch_rnd_long_drbg_reseed' accessible via /sys/module/arch_random/parameters/rndlong_drbg_reseed or as kernel command line parameter arch_random.rndlong_drbg_reseed= This parameter tells how often the drbg fills the 4K buffer before it is re-seeded by fresh entropy from the trng. A value of 16 results in reseeding the drbg at every 16 * 4 KB = 64 KB with 32 bytes of fresh entropy pulled from the trng. So a value of 16 would result in 256 bits entropy per 64 KB. A value of 256 results in 1MB of drbg output before a reseed of the drbg is done. So this would spread the 256 bits of entropy among 1MB. Setting this parameter to 0 forces the reseed to take place every time the 4K buffer is depleted, so the entropy rises to 256 bits entropy per 4K or 0.5 bit entropy per arch_get_random_long(). With setting this parameter to negative values all this effort is disabled, arch_get_random long() returns false and thus indicating that the arch_get_random_long() feature is disabled at all. arch_get_random_long() is used by random.c among others to provide an initial hash value to be mixed with the entropy pool on every random data pull. For about 64 bytes read from /dev/urandom there is one call to arch_get_random_long(). So these additional random long values count for performance of /dev/urandom with measurable but low penalty. Signed-off-by: Harald Freudenberger Reviewed-by: Ingo Franzki Reviewed-by: Juergen Christ Signed-off-by: Heiko Carstens

s390x: Mark archrandom.h functions __must_check

2020-01-25T17:18:51Z

We must not use the pointer output without validating the success of the random read. Reviewed-by: Harald Freudenberger Reviewed-by: Ard Biesheuvel Signed-off-by: Richard Henderson Signed-off-by: Mark Brown Link: https://lore.kernel.org/r/20200110145422.49141-11-broonie@kernel.org Signed-off-by: Theodore Ts'o

s390: Remove arch_has_random, arch_has_random_seed

2020-01-25T17:18:50Z

These symbols are currently part of the generic archrandom.h interface, but are currently unused and can be removed. Signed-off-by: Richard Henderson Signed-off-by: Mark Brown Link: https://lore.kernel.org/r/20200110145422.49141-4-broonie@kernel.org Signed-off-by: Theodore Ts'o

s390/archrandom: Rework arch random implementation.

2018-05-30T09:18:04Z

The arch_get_random_seed_long() invocation done by the random device driver is done in interrupt context and may be invoked very very frequently. The existing s390 arch_get_random_seed*() implementation uses the PRNO(TRNG) instruction which produces excellent high quality entropy but is relatively slow and thus expensive. This fix reworks the arch_get_random_seed* implementation. It introduces a buffer concept to decouple the delivery of random data via arch_get_random_seed*() from the generation of new random bytes. The buffer of random data is filled asynchronously by a workqueue thread. If there are enough bytes in the buffer the s390_arch_random_generate() just delivers these bytes. Otherwise false is returned until the worker thread refills the buffer. The worker fills the rng buffer by pulling fresh entropy from the high quality (but slow) true hardware random generator. This entropy is then spread over the buffer with an pseudo random generator. As the arch_get_random_seed_long() fetches 8 bytes and the calling function add_interrupt_randomness() counts this as 1 bit entropy the distribution needs to make sure there is in fact 1 bit entropy contained in 8 bytes of the buffer. The current values pull 32 byte entropy and scatter this into a 2048 byte buffer. So 8 byte in the buffer will contain 1 bit of entropy. The worker thread is rescheduled based on the charge level of the buffer but at least with 500 ms delay to avoid too much cpu consumption. So the max. amount of rng data delivered via arch_get_random_seed is limited to 4Kb per second. Signed-off-by: Harald Freudenberger Reviewed-by: Patrick Steuer Signed-off-by: Martin Schwidefsky

Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux

2017-11-13T19:47:01Z

Pull s390 updates from Heiko Carstens: "Since Martin is on vacation you get the s390 pull request for the v4.15 merge window this time from me. Besides a lot of cleanups and bug fixes these are the most important changes: - a new regset for runtime instrumentation registers - hardware accelerated AES-GCM support for the aes_s390 module - support for the new CEX6S crypto cards - support for FORTIFY_SOURCE - addition of missing z13 and new z14 instructions to the in-kernel disassembler - generate opcode tables for the in-kernel disassembler out of a simple text file instead of having to manually maintain those tables - fast memset16, memset32 and memset64 implementations - removal of named saved segment support - hardware counter support for z14 - queued spinlocks and queued rwlocks implementations for s390 - use the stack_depth tracking feature for s390 BPF JIT - a new s390_sthyi system call which emulates the sthyi (store hypervisor information) instruction - removal of the old KVM virtio transport - an s390 specific CPU alternatives implementation which is used in the new spinlock code" * 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux: (88 commits) MAINTAINERS: add virtio-ccw.h to virtio/s390 section s390/noexec: execute kexec datamover without DAT s390: fix transactional execution control register handling s390/bpf: take advantage of stack_depth tracking s390: simplify transactional execution elf hwcap handling s390/zcrypt: Rework struct ap_qact_ap_info. s390/virtio: remove unused header file kvm_virtio.h s390: avoid undefined behaviour s390/disassembler: generate opcode tables from text file s390/disassembler: remove insn_to_mnemonic() s390/dasd: avoid calling do_gettimeofday() s390: vfio-ccw: Do not attempt to free no-op, test and tic cda. s390: remove named saved segment support s390/archrandom: Reconsider s390 arch random implementation s390/pci: do not require AIS facility s390/qdio: sanitize put_indicator s390/qdio: use atomic_cmpxchg s390/nmi: avoid using long-displacement facility s390: pass endianness info to sparse s390/decompressor: remove informational messages ...