x/crypto/openpgp/clearsign: TestMultiSign go test fails on armhf system

[rogers0]

What version of Go are you using (go version)?

1.14

Does this issue reproduce with the latest release?

Not sure

What operating system and processor architecture are you using (go env)?

debian unstable

What did you do?

golang-go.crypto test fails on reproducible builds

log:

=== RUN   TestMultiSign
panic: test timed out after 10m0s

goroutine 7 [running]:
testing.(*M).startAlarm.func1()
	/usr/lib/go-1.14/src/testing/testing.go:1459 +0xbc
created by time.goFunc
	/usr/lib/go-1.14/src/time/sleep.go:168 +0x34

goroutine 1 [chan receive, 9 minutes]:
testing.(*T).Run(0x1106140, 0x1c338a, 0xd, 0x1cf1e4, 0x301)
	/usr/lib/go-1.14/src/testing/testing.go:1043 +0x2d0
testing.runTests.func1(0x109a000)
	/usr/lib/go-1.14/src/testing/testing.go:1284 +0x68
testing.tRunner(0x109a000, 0x1045ef8)
	/usr/lib/go-1.14/src/testing/testing.go:991 +0xbc
testing.runTests(0x100e100, 0x308678, 0x6, 0x6, 0x0)
	/usr/lib/go-1.14/src/testing/testing.go:1282 +0x238
testing.(*M).Run(0x1058100, 0x0)
	/usr/lib/go-1.14/src/testing/testing.go:1199 +0x134
main.main()
	_testmain.go:54 +0x120

goroutine 20 [runnable]:
math/big.nat.montgomery(0x136aa20, 0x20, 0x44, 0x136ab40, 0x20, 0x46, 0x136ab40, 0x20, 0x46, 0x12eabd0, ...)
	/usr/lib/go-1.14/src/math/big/nat.go:213 +0x3bc
math/big.nat.expNNMontgomery(0x136aa20, 0x20, 0x46, 0x1345d40, 0x20, 0x24, 0x1345c20, 0x20, 0x24, 0x12eabd0, ...)
	/usr/lib/go-1.14/src/math/big/nat.go:1446 +0x5d4
math/big.nat.expNN(0x136aa20, 0x20, 0x44, 0x1345d40, 0x20, 0x24, 0x1345c20, 0x20, 0x24, 0x12eabd0, ...)
	/usr/lib/go-1.14/src/math/big/nat.go:1261 +0x698
math/big.nat.probablyPrimeMillerRabin(0x12eabd0, 0x20, 0x24, 0x15, 0x56be4f01, 0x24)
	/usr/lib/go-1.14/src/math/big/prime.go:106 +0x380
math/big.(*Int).ProbablyPrime(0x100e240, 0x14, 0x1040d6c)
	/usr/lib/go-1.14/src/math/big/prime.go:78 +0x224
crypto/rand.Prime(0x1fba08, 0x110ab54, 0x400, 0x1100020, 0x16, 0x1018860)
	/usr/lib/go-1.14/src/crypto/rand/util.go:99 +0x244
crypto/rsa.GenerateMultiPrimeKey(0x1fba08, 0x110ab54, 0x2, 0x800, 0x1c18cc, 0x5, 0x14920a0)
	/usr/lib/go-1.14/src/crypto/rsa/rsa.go:268 +0x10c
crypto/rsa.GenerateKey(...)
	/usr/lib/go-1.14/src/crypto/rsa/rsa.go:208
golang.org/x/crypto/openpgp.NewEntity(0x1c1748, 0x4, 0x1c1daa, 0x7, 0x1c18cc, 0x5, 0x1040f7c, 0x2, 0x0, 0x0)
	/build/1st/golang-go.crypto-0.0~git20200604.70a84ac/obj-arm-linux-gnueabihf/src/golang.org/x/crypto/openpgp/keys.go:518 +0xc0
golang.org/x/crypto/openpgp/clearsign.TestMultiSign(0x1106140)
	/build/1st/golang-go.crypto-0.0~git20200604.70a84ac/obj-arm-linux-gnueabihf/src/golang.org/x/crypto/openpgp/clearsign/clearsign_test.go:153 +0x108
testing.tRunner(0x1106140, 0x1cf1e4)
	/usr/lib/go-1.14/src/testing/testing.go:991 +0xbc
created by testing.(*T).Run
	/usr/lib/go-1.14/src/testing/testing.go:1042 +0x2b0
FAIL	golang.org/x/crypto/openpgp/clearsign	600.048s

https://tests.reproducible-builds.org/debian/rbuild/unstable/armhf/golang-go.crypto_0.0~git20200604.70a84ac-1.rbuild.log.gz

[cagedmantis]

/cc @FiloSottile

[mangalaman93]

I observe this in our jenkins pipeline that runs periodic builds. Jenkins build runs on a gcp node with amd 64 bit architecture running a linux distribution.

panic: test timed out after 10m0s

goroutine 1323 [running]:
testing.(*M).startAlarm.func1()
	/usr/local/go/src/testing/testing.go:1700 +0xe5
created by time.goFunc
	/usr/local/go/src/time/sleep.go:180 +0x45

goroutine 1 [chan receive]:
testing.(*T).Run(0xc000594300, 0xd03bf7, 0x1b, 0xd2bc68, 0x495301)
	/usr/local/go/src/testing/testing.go:1239 +0x2da
testing.runTests.func1(0xc0004d5080)
	/usr/local/go/src/testing/testing.go:1511 +0x78
testing.tRunner(0xc0004d5080, 0xc00049fda8)
	/usr/local/go/src/testing/testing.go:1193 +0xef
testing.runTests(0xc00011c0c0, 0x128b3c0, 0x13, 0x13, 0xc042730e2abde59c, 0x8bf42f1a2c, 0x129b540, 0xcf9016)
	/usr/local/go/src/testing/testing.go:1509 +0x2fe
testing.(*M).Run(0xc00037c300, 0x0)
	/usr/local/go/src/testing/testing.go:1417 +0x1eb
main.main()
	_testmain.go:135 +0x1c5

goroutine 1463 [runnable]:
math/big.nat.montgomery(0xc000972a00, 0x10, 0x26, 0xc00097d320, 0x10, 0x24, 0xc00097d320, 0x10, 0x24, 0xc00070af00, ...)
	/usr/local/go/src/math/big/nat.go:213 +0x50d
math/big.nat.expNNMontgomery(0xc00097d320, 0x10, 0x24, 0xc00096f5e0, 0x10, 0x14, 0xc00096f4a0, 0x10, 0x14, 0xc00070af00, ...)
	/usr/local/go/src/math/big/nat.go:1449 +0x91d
math/big.nat.expNN(0xc00096f680, 0x10, 0x14, 0xc00096f5e0, 0x10, 0x14, 0xc00096f4a0, 0x10, 0x14, 0xc00070af00, ...)
	/usr/local/go/src/math/big/nat.go:1262 +0xa06
math/big.nat.probablyPrimeMillerRabin(0xc00070af00, 0x10, 0x14, 0x15, 0x54a32ebdd2fd1201, 0x0)
	/usr/local/go/src/math/big/prime.go:106 +0x48e
math/big.(*Int).ProbablyPrime(0xc00031c600, 0x14, 0xc0009574e8)
	/usr/local/go/src/math/big/prime.go:78 +0x232
crypto/rand.Prime(0xde2360, 0xc0000fa1e0, 0x400, 0xc0004b6130, 0xc000850038, 0x0)
	/usr/local/go/src/crypto/rand/util.go:99 +0x185
crypto/rsa.GenerateMultiPrimeKey(0xde2360, 0xc0000fa1e0, 0x2, 0x800, 0x0, 0xd2dcd8, 0x0)
	/usr/local/go/src/crypto/rsa/rsa.go:302 +0x170
crypto/rsa.GenerateKey(...)
	/usr/local/go/src/crypto/rsa/rsa.go:242
...
omitted, call to rsa.GenerateKey(crypto/rand.Reader, 2048)
....
testing.tRunner(0xc000594300, 0xd2bc68)
	/usr/local/go/src/testing/testing.go:1193 +0xef
created by testing.(*T).Run
	/usr/local/go/src/testing/testing.go:1238 +0x2b3

[mangalaman93]

I realized that our jenkins hosts were too slow for 2048 bit rsa keys for our tests. That could have caused this issue too.

[FiloSottile]

I would take a PR to reduce the size of the keys used in tests.