[PATCH] crypto: algapi - Remove test larvals to fix error paths

[PATCH] crypto: algapi - Remove test larvals to fix error paths

[Philipp Zabel]

If crypto_unregister_alg() is called with an algorithm that still has a
pending test larval, the algorithm will have a reference count of 2 and
crypto_unregister_alg() will trigger a BUG(). This can happen during
cleanup if the error path is taken for a built-in algorithm, before
crypto_algapi_init() was called.

Kill test larvals for untested algorithms during removal to fix the
reference count.

Fixes: adad556efcdd ("crypto: api - Fix built-in testing dependency failures")
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
---
 crypto/algapi.c | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)

diff --git a/crypto/algapi.c b/crypto/algapi.c
index a366cb3e8aa1..fc3b75e59d0e 100644
--- a/crypto/algapi.c
+++ b/crypto/algapi.c
@@ -458,6 +458,28 @@ void crypto_unregister_alg(struct crypto_alg *alg)
 	if (WARN(ret, "Algorithm %s is not registered", alg->cra_driver_name))
 		return;
 
+	/* If there is a test larval holding a reference, remove it */
+	if (!(alg->cra_flags & CRYPTO_ALG_TESTED)) {
+		struct crypto_alg *q, *n;
+
+		list_for_each_entry_safe(q, n, &crypto_alg_list, cra_list) {
+			struct crypto_larval *l;
+
+			if (!crypto_is_larval(q))
+				continue;
+
+			l = (void *)q;
+
+			if (!crypto_is_test_larval(l))
+				continue;
+
+			if (l->adult != alg)
+				continue;
+
+			crypto_larval_kill(q);
+		}
+	}
+
 	BUG_ON(refcount_read(&alg->cra_refcnt) != 1);
 	if (alg->cra_destroy)
 		alg->cra_destroy(alg);

base-commit: e783362eb54cd99b2cac8b3a9aeac942e6f6ac07
-- 
2.30.2

Re: [PATCH] crypto: algapi - Remove test larvals to fix error paths

[Philipp Zabel]

On Wed, 2022-01-26 at 15:53 +0100, Philipp Zabel wrote:
[...]
> This can happen during cleanup if the error path is taken for a built-
> in algorithm, before crypto_algapi_init() was called.

I see this happen on ARM with CONFIG_CRYPTO_AES_ARM_BS=y since v5.16-rc1
because the simd_skcipher_create_compat("ecb(aes)", "ecb-aes-neonbs",
"__ecb-aes-neonbs") call in arch/arm/crypto/aes-neonbs-glue.c returns
-ENOENT. I believe that is the same issue as reported in [1].

[1] https://lore.kernel.org/lkml/CA+G9fYvyATAWicFbnKnOTqc=MKUXNrbCBYP6zej3FJJyA31WPQ@mail.gmail.com/

[    3.921012] ------------[ cut here ]------------
[    3.925715] kernel BUG at crypto/algapi.c:461!
[    3.930282] Internal error: Oops - BUG: 0 [#1] PREEMPT SMP ARM
[    3.936187] Modules linked in:
[    3.939318] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.17.0-rc1-20220124-1 #1
[    3.946614] Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
[    3.953263] PC is at crypto_unregister_alg+0x11c/0x124
[    3.958486] LR is at __this_cpu_preempt_check+0x1c/0x20
[    3.963845] pc : [<805611d4>]    lr : [<80bac708>]    psr: 20000113
[    3.970180] sp : 81ce7e40  ip : 81ce7db0  fp : 81ce7e6c
[    3.975524] r10: 8104a878  r9 : 00000000  r8 : 000000ff
[    3.980818] r7 : 812af474  r6 : 00000001  r5 : 812af488  r4 : 82e64a80
[    3.987468] r3 : 00000002  r2 : 00000001  r1 : 00000000  r0 : 00000000
[    3.994066] Flags: nzCv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment none
[    4.001326] Control: 10c5387d  Table: 1000404a  DAC: 00000051
[    4.007139] Register r0 information: NULL pointer
[    4.011973] Register r1 information: NULL pointer
[    4.016750] Register r2 information: non-paged memory
[    4.021930] Register r3 information: non-paged memory
[    4.027053] Register r4 information: slab kmalloc-512 start 82e64a00 pointer offset 128 size 512
[    4.035991] Register r5 information: non-slab/vmalloc memory
[    4.041723] Register r6 information: non-paged memory
[    4.046899] Register r7 information: non-slab/vmalloc memory
[    4.052630] Register r8 information: non-paged memory
[    4.057807] Register r9 information: NULL pointer
[    4.062584] Register r10 information: non-slab/vmalloc memory
[    4.068456] Register r11 information: non-slab/vmalloc memory
[    4.074276] Register r12 information: non-slab/vmalloc memory
[    4.080147] Process swapper/0 (pid: 1, stack limit = 0xbd1e2d0f)
[    4.086226] Stack: (0x81ce7e40 to 0x81ce8000)
[    4.090712] 7e40: 00002080 81110580 00000000 81ce7e4c 81ce7e4c 81108ec8 00000000 82e64a00
[    4.098962] 7e60: 81ce7e7c 81ce7e70 80563044 805610c4 81ce7e94 81ce7e80 80574e94 80563034
[    4.107264] 7e80: 812af478 812af488 81ce7eac 81ce7e98 80122554 80574e80 81110580 fffffffe
[    4.115566] 7ea0: 81ce7ecc 81ce7eb0 81008f88 80122530 81da8fc0 8128f100 81108f10 81008f00
[    4.123866] 7ec0: 81ce7f4c 81ce7ed0 80102684 81008f0c 81ce7eec 81ce7ee0 80bac708 80bac500
[    4.132114] 7ee0: 81ce7f1c 81ce7ef0 80babd54 80bac6f8 00000001 81da8fc0 8128f100 81108f10
[    4.140415] 7f00: 8104a858 812af000 80fac7c4 8104a874 81ce7f2c 81ce7f20 801b91dc 81108ec8
[    4.148717] 7f20: 81ce7f4c 8107d0bc 81d240c0 00000008 8104a858 000000ff 812af000 8104a878
[    4.156964] 7f40: 81ce7f94 81ce7f50 81001428 8010260c 00000007 00000007 00000000 8100044c
[    4.165266] 7f60: 80fac7c4 8100044c 81ce7f8c 81108ec0 80bac9bc 00000000 00000000 00000000
[    4.173566] 7f80: 00000000 00000000 81ce7fac 81ce7f98 80bac9e4 8100125c 00000000 80bac9bc
[    4.181866] 7fa0: 00000000 81ce7fb0 80100108 80bac9c8 00000000 00000000 00000000 00000000
[    4.190113] 7fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[    4.198413] 7fe0: 00000000 00000000 00000000 00000000 00000013 00000000 00000000 00000000
[    4.206709] Backtrace: 
[    4.209230]  crypto_unregister_alg from crypto_unregister_skcipher+0x1c/0x20
[    4.216417]  r4:82e64a00
[    4.219019]  crypto_unregister_skcipher from simd_skcipher_free+0x20/0x2c
[    4.225892]  simd_skcipher_free from aes_exit+0x30/0x4c
[    4.231257]  r5:812af488 r4:812af478
[    4.234900]  aes_exit from aes_init+0x88/0xa4
[    4.239393]  r5:fffffffe r4:81110580
[    4.243034]  aes_init from do_one_initcall+0x84/0x3c4
[    4.248168]  r7:81008f00 r6:81108f10 r5:8128f100 r4:81da8fc0
[    4.253949]  do_one_initcall from kernel_init_freeable+0x1d8/0x230
[    4.260218]  r10:8104a878 r9:812af000 r8:000000ff r7:8104a858 r6:00000008 r5:81d240c0
[    4.268168]  r4:8107d0bc
[    4.270769]  kernel_init_freeable from kernel_init+0x28/0x144
[    4.276656]  r10:00000000 r9:00000000 r8:00000000 r7:00000000 r6:00000000 r5:80bac9bc
[    4.284608]  r4:81108ec0
[    4.287208]  kernel_init from ret_from_fork+0x14/0x2c
[    4.292337] Exception stack(0x81ce7fb0 to 0x81ce7ff8)
[    4.297517] 7fa0:                                     00000000 00000000 00000000 00000000
[    4.305766] 7fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[    4.314067] 7fe0: 00000000 00000000 00000000 00000000 00000013 00000000
[    4.320803]  r5:80bac9bc r4:00000000
[    4.324450] Code: e34800eb e58d4000 eb18fc40 eaffffe9 (e7f001f2) 
[    4.330669] ---[ end trace 0000000000000000 ]---
[    4.346697] Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
[    4.354500] CPU2: stopping
[    4.357283] CPU: 2 PID: 0 Comm: swapper/2 Tainted: G      D           5.17.0-rc1-20220124-1 #1
[    4.366022] Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
[    4.372618] Backtrace: 
[    4.375137]  dump_backtrace from show_stack+0x20/0x24
[    4.380328]  r7:81cc3f40 r6:00000080 r5:80e968a4 r4:60000193
[    4.386055]  show_stack from dump_stack_lvl+0x60/0x78
[    4.391238]  dump_stack_lvl from dump_stack+0x18/0x1c
[    4.396369]  r7:81cc3f40 r6:8128f1a0 r5:00000002 r4:00000004
[    4.402146]  dump_stack from do_handle_IPI+0x330/0x35c
[    4.407364]  do_handle_IPI from ipi_handler+0x28/0x30
[    4.412550]  r10:bfd360b8 r9:81dc1f80 r8:81d09e38 r7:81cc3f40 r6:811096e4 r5:00000014
[    4.420447]  r4:81cc1600 r3:00000010
[    4.424142]  ipi_handler from handle_percpu_devid_irq+0xac/0x180
[    4.430229]  handle_percpu_devid_irq from generic_handle_domain_irq+0x50/0x94
[    4.437505]  r9:81dc1f80 r8:f400010c r7:81085b54 r6:81dc1f80 r5:00000004 r4:81cc4400
[    4.445367]  generic_handle_domain_irq from gic_handle_irq+0x80/0x94
[    4.451805]  r7:81085b54 r6:f4000100 r5:8123d2b4 r4:811096e4
[    4.457584]  gic_handle_irq from generic_handle_arch_irq+0x60/0x80
[    4.463847]  r9:81dc1f80 r8:00000000 r7:81d09ed0 r6:80eba814 r5:80e97b0c r4:81085b30
[    4.471707]  generic_handle_arch_irq from __irq_svc+0x54/0x88
[    4.477529] Exception stack(0x81d09ed0 to 0x81d09f18)
[    4.482706] 9ec0:                                     808ae49c 00000000 3ecb0000 00000000
[    4.490956] 9ee0: 00000001 038c318c 00000001 00000001 00000001 811116a8 bfd360b8 81d09f54
[    4.499203] 9f00: 81d09ef8 81d09f20 802434b8 808ae4a0 60000013 ffffffff
[    4.505942]  r9:81dc1f80 r8:00000001 r7:81d09f04 r6:ffffffff r5:60000013 r4:808ae4a0
[    4.513749]  cpuidle_enter_state from cpuidle_enter+0x40/0x50
[    4.519584]  r10:80eba814 r9:00000001 r8:0347bb80 r7:81108f50 r6:00000001 r5:811116a8
[    4.527535]  r4:bfd360b8
[    4.530133]  cpuidle_enter from do_idle+0x21c/0x2b4
[    4.535096]  r9:811116a8 r8:bfd360b8 r7:81108f50 r6:00000002 r5:81dc1f80 r4:81108f10
[    4.542959]  do_idle from cpu_startup_entry+0x28/0x2c
[    4.548092]  r10:00000000 r9:412fc09a r8:1000406a r7:812af324 r6:81dc1f80 r5:00000002
[    4.556043]  r4:00000097
[    4.558644]  cpu_startup_entry from secondary_start_kernel+0x174/0x198
[    4.565301]  secondary_start_kernel from 0x10101bd4
[    4.570258]  r7:812af324 r6:10c0387d r5:00000051 r4:11d6806a
[    4.576037] CPU3: stopping
[    4.578819] CPU: 3 PID: 0 Comm: swapper/3 Tainted: G      D           5.17.0-rc1-20220124-1 #1
[    4.587561] Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
[    4.594158] Backtrace: 
[    4.596732]  dump_backtrace from show_stack+0x20/0x24
[    4.601870]  r7:81cc3f40 r6:00000080 r5:80e968a4 r4:60000193
[    4.607650]  show_stack from dump_stack_lvl+0x60/0x78
[    4.612781]  dump_stack_lvl from dump_stack+0x18/0x1c
[    4.617969]  r7:81cc3f40 r6:8128f1a0 r5:00000003 r4:00000004
[    4.623695]  dump_stack from do_handle_IPI+0x330/0x35c
[    4.628912]  do_handle_IPI from ipi_handler+0x28/0x30
[    4.634099]  r10:bfd470b8 r9:81dc0fc0 r8:81d0be38 r7:81cc3f40 r6:811096e4 r5:00000014
[    4.642050]  r4:81cc1600 r3:00000010
[    4.645692]  ipi_handler from handle_percpu_devid_irq+0xac/0x180
[    4.651778]  handle_percpu_devid_irq from generic_handle_domain_irq+0x50/0x94
[    4.659053]  r9:81dc0fc0 r8:f400010c r7:81085b54 r6:81dc0fc0 r5:00000004 r4:81cc4400
[    4.666916]  generic_handle_domain_irq from gic_handle_irq+0x80/0x94
[    4.673352]  r7:81085b54 r6:f4000100 r5:8123d2b4 r4:811096e4
[    4.679132]  gic_handle_irq from generic_handle_arch_irq+0x60/0x80
[    4.685396]  r9:81dc0fc0 r8:00000000 r7:81d0bed0 r6:80eba814 r5:80e97b0c r4:81085b30
[    4.693259]  generic_handle_arch_irq from __irq_svc+0x54/0x88
[    4.699135] Exception stack(0x81d0bed0 to 0x81d0bf18)
[    4.704259] bec0:                                     808ae49c 00000000 3ecc1000 00000000
[    4.712562] bee0: 00000001 038c32da 00000001 00000001 00000001 811116a8 bfd470b8 81d0bf54
[    4.720810] bf00: 81d0bef8 81d0bf20 802434b8 808ae4a0 60000013 ffffffff
[    4.727549]  r9:81dc0fc0 r8:00000001 r7:81d0bf04 r6:ffffffff r5:60000013 r4:808ae4a0
[    4.735412]  cpuidle_enter_state from cpuidle_enter+0x40/0x50
[    4.741243]  r10:80eba814 r9:00000001 r8:0347bb80 r7:81108f50 r6:00000001 r5:811116a8
[    4.749195]  r4:bfd470b8
[    4.751796]  cpuidle_enter from do_idle+0x21c/0x2b4
[    4.756810]  r9:811116a8 r8:bfd470b8 r7:81108f50 r6:00000003 r5:81dc0fc0 r4:81108f10
[    4.764619]  do_idle from cpu_startup_entry+0x28/0x2c
[    4.769806]  r10:00000000 r9:412fc09a r8:1000406a r7:812af324 r6:81dc0fc0 r5:00000003
[    4.777703]  r4:00000097
[    4.780357]  cpu_startup_entry from secondary_start_kernel+0x174/0x198
[    4.786963]  secondary_start_kernel from 0x10101bd4
[    4.791970]  r7:812af324 r6:10c0387d r5:00000051 r4:11d6806a
[    4.797697] CPU1: stopping
[    4.800480] CPU: 1 PID: 0 Comm: swapper/1 Tainted: G      D           5.17.0-rc1-20220124-1 #1
[    4.809219] Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
[    4.815870] Backtrace: 
[    4.818389]  dump_backtrace from show_stack+0x20/0x24
[    4.823527]  r7:81cc3f40 r6:00000080 r5:80e968a4 r4:60000193
[    4.829311]  show_stack from dump_stack_lvl+0x60/0x78
[    4.834442]  dump_stack_lvl from dump_stack+0x18/0x1c
[    4.839630]  r7:81cc3f40 r6:8128f1a0 r5:00000001 r4:00000004
[    4.845357]  dump_stack from do_handle_IPI+0x330/0x35c
[    4.850628]  do_handle_IPI from ipi_handler+0x28/0x30
[    4.855761]  r10:bfd250b8 r9:81dbee40 r8:81d07e38 r7:81cc3f40 r6:811096e4 r5:00000014
[    4.863714]  r4:81cc1600 r3:00000010
[    4.867357]  ipi_handler from handle_percpu_devid_irq+0xac/0x180
[    4.873497]  handle_percpu_devid_irq from generic_handle_domain_irq+0x50/0x94
[    4.880716]  r9:81dbee40 r8:f400010c r7:81085b54 r6:81dbee40 r5:00000004 r4:81cc4400
[    4.888578]  generic_handle_domain_irq from gic_handle_irq+0x80/0x94
[    4.895067]  r7:81085b54 r6:f4000100 r5:8123d2b4 r4:811096e4
[    4.900793]  gic_handle_irq from generic_handle_arch_irq+0x60/0x80
[    4.907110]  r9:81dbee40 r8:00000000 r7:81d07ed0 r6:80eba814 r5:80e97b0c r4:81085b30
[    4.914919]  generic_handle_arch_irq from __irq_svc+0x54/0x88
[    4.920796] Exception stack(0x81d07ed0 to 0x81d07f18)
[    4.925921] 7ec0:                                     808ae49c 00000000 3ec9f000 00000000
[    4.934226] 7ee0: 00000001 038c32da 00000001 00000001 00000001 811116a8 bfd250b8 81d07f54
[    4.942528] 7f00: 81d07ef8 81d07f20 802434b8 808ae4a0 60000013 ffffffff
[    4.949214]  r9:81dbee40 r8:00000001 r7:81d07f04 r6:ffffffff r5:60000013 r4:808ae4a0
[    4.957077]  cpuidle_enter_state from cpuidle_enter+0x40/0x50
[    4.962961]  r10:80eba814 r9:00000001 r8:0347bb80 r7:81108f50 r6:00000001 r5:811116a8
[    4.970859]  r4:bfd250b8
[    4.973459]  cpuidle_enter from do_idle+0x21c/0x2b4
[    4.978474]  r9:811116a8 r8:bfd250b8 r7:81108f50 r6:00000001 r5:81dbee40 r4:81108f10
[    4.986338]  do_idle from cpu_startup_entry+0x28/0x2c
[    4.991471]  r10:00000000 r9:412fc09a r8:1000406a r7:812af324 r6:81dbee40 r5:00000001
[    4.999424]  r4:00000097
[    5.002025]  cpu_startup_entry from secondary_start_kernel+0x174/0x198
[    5.008683]  secondary_start_kernel from 0x10101bd4
[    5.013639]  r7:812af324 r6:10c0387d r5:00000051 r4:11d6806a
[    5.034197] ---[ end Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b ]---

regards
Philipp

Re: [PATCH] crypto: algapi - Remove test larvals to fix error paths

[Herbert Xu]

On Wed, Jan 26, 2022 at 04:01:04PM +0100, Philipp Zabel wrote:
> On Wed, 2022-01-26 at 15:53 +0100, Philipp Zabel wrote:
> [...]
> > This can happen during cleanup if the error path is taken for a built-
> > in algorithm, before crypto_algapi_init() was called.
> 
> I see this happen on ARM with CONFIG_CRYPTO_AES_ARM_BS=y since v5.16-rc1
> because the simd_skcipher_create_compat("ecb(aes)", "ecb-aes-neonbs",
> "__ecb-aes-neonbs") call in arch/arm/crypto/aes-neonbs-glue.c returns
> -ENOENT. I believe that is the same issue as reported in [1].

Good catch.  We definitely need to fix the crash from the unregister
while the algorithm is still being referenced.

However, the fact that the simd create is failing is probably a bug
too.  Could you take a look at that aspect and whether the failure
is related to the testmgr changes in question?

Thanks!
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

Re: [PATCH] crypto: algapi - Remove test larvals to fix error paths

[Herbert Xu]

On Wed, Jan 26, 2022 at 03:53:22PM +0100, Philipp Zabel wrote:
> If crypto_unregister_alg() is called with an algorithm that still has a
> pending test larval, the algorithm will have a reference count of 2 and
> crypto_unregister_alg() will trigger a BUG(). This can happen during
> cleanup if the error path is taken for a built-in algorithm, before
> crypto_algapi_init() was called.
> 
> Kill test larvals for untested algorithms during removal to fix the
> reference count.
> 
> Fixes: adad556efcdd ("crypto: api - Fix built-in testing dependency failures")
> Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
> ---
>  crypto/algapi.c | 22 ++++++++++++++++++++++
>  1 file changed, 22 insertions(+)

Yes this is definitely a bug.

However, I think simply fixing this for test larvals is not enough.
After all, another thread could come through the middle and take
a reference to our newly minted algorithm before a subsequent
algorithm registration failure and thus would trigger exactly
the same crash.

So we need something a bit more general.  Previously we relied
on module reference counts to stop the unregistering of live
algorithms.  But that is clearly not of any use in this case.

It also fails for hardware devices than can be unplugged.

One solution would be to simply allow the unregistration to
continue and only free the data structures after all references
have gone away.

Cheers,
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

Re: [PATCH] crypto: algapi - Remove test larvals to fix error paths

[Herbert Xu]

On Wed, Jan 26, 2022 at 04:01:04PM +0100, Philipp Zabel wrote:
> 
> I see this happen on ARM with CONFIG_CRYPTO_AES_ARM_BS=y since v5.16-rc1
> because the simd_skcipher_create_compat("ecb(aes)", "ecb-aes-neonbs",
> "__ecb-aes-neonbs") call in arch/arm/crypto/aes-neonbs-glue.c returns
> -ENOENT. I believe that is the same issue as reported in [1].

I cannot reproduce this crash with qemu.  If you can still
reproduce this, please send me your complete kconfig file.

Thanks,
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

Re: [PATCH] crypto: algapi - Remove test larvals to fix error paths

[Ard Biesheuvel]

On Wed, 16 Mar 2022 at 02:11, Herbert Xu <herbert@gondor.apana.org.au> wrote:
>
> On Wed, Jan 26, 2022 at 04:01:04PM +0100, Philipp Zabel wrote:
> >
> > I see this happen on ARM with CONFIG_CRYPTO_AES_ARM_BS=y since v5.16-rc1
> > because the simd_skcipher_create_compat("ecb(aes)", "ecb-aes-neonbs",
> > "__ecb-aes-neonbs") call in arch/arm/crypto/aes-neonbs-glue.c returns
> > -ENOENT. I believe that is the same issue as reported in [1].
>
> I cannot reproduce this crash with qemu.  If you can still
> reproduce this, please send me your complete kconfig file.
>

According to the bisect log in the other thread,
adad556efcdd42a1d9e060cb is the culprit, which does not seem
surprising, at is would result in the SIMD skcipher being encapsulated
to not be available yet when the SIMD helper tries to take a reference
to it.

Given that the resulting algo will be subject to testing as well,
could we simply override the test requirement for the inner algo? That
should make it available immediately, IIUC, and the outer algo test is
guaranteed to cover the inner one.

Re: [PATCH] crypto: algapi - Remove test larvals to fix error paths

[Herbert Xu]

On Wed, Mar 16, 2022 at 08:23:24AM +0100, Ard Biesheuvel wrote:
>
> According to the bisect log in the other thread,
> adad556efcdd42a1d9e060cb is the culprit, which does not seem
> surprising, at is would result in the SIMD skcipher being encapsulated
> to not be available yet when the SIMD helper tries to take a reference
> to it.

It's supposed to work because any use of an algorithm prior to
the tests starting will automatically trigger the test right away.
I confirmed this by booting the kernel in qemu with AES_ARM_BS=y
and it successfully registered those algorithms and passed the
self-tests.

Someone else has already sent me a complete kconfig file which
hopefully should reproduce the crash for me.

Cheers,
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

Re: [PATCH] crypto: algapi - Remove test larvals to fix error paths

[Philipp Zabel]

On Do, 2022-03-17 at 10:10 +1200, Herbert Xu wrote:
> On Wed, Mar 16, 2022 at 08:23:24AM +0100, Ard Biesheuvel wrote:
> > 
> > According to the bisect log in the other thread,
> > adad556efcdd42a1d9e060cb is the culprit, which does not seem
> > surprising, at is would result in the SIMD skcipher being
> > encapsulated
> > to not be available yet when the SIMD helper tries to take a
> > reference
> > to it.
> 
> It's supposed to work because any use of an algorithm prior to
> the tests starting will automatically trigger the test right away.
> I confirmed this by booting the kernel in qemu with AES_ARM_BS=y
> and it successfully registered those algorithms and passed the
> self-tests.
> 
> Someone else has already sent me a complete kconfig file which
> hopefully should reproduce the crash for me.

It seems you already did. I can confirm that [1] fixes the crash for
me, as I had CRYPTO_CBC disabled. After enabling it, aes_init() doesn't
hit the error path anymore, hiding the BUG() again.

[1] https://lore.kernel.org/all/YjJq0RLIHvN7YWaT@gondor.apana.org.au/

regards
Philipp