Replies: 2 comments 2 replies
-
|
@1Mateus Thanks for a great description of the problem! Could I ask you to change to
Without looking at the Gramine logs, I can't say why you see SIGSEGV. (Also IIUC Rust's |
Beta Was this translation helpful? Give feedback.
2 replies
-
|
@dimakuv Hey, thanks! I changed the log_level to 'all' as you suggested: https://github.com/hack-a-chain-software/bazk/blob/feat/new-api/packages/gramine/bazk-build/bazk.manifest.template#L7 And here is the output: I tried making changes the image version of gramine that I use for building and running the container, now both use the version: https://github.com/hack-a-chain-software/bazk/blob/feat/new-api/packages/gramine/gramine-image/Dockerfile#L1 Now, when I send a command to execute the 'new' binary and receive this log: |
Beta Was this translation helpful? Give feedback.
-
|
@dimakuv Hey, we run more tests with the binaries. We updated lib.rs so that the binary runs on a single core only: https://github.com/hack-a-chain-software/phase2-bn254/blob/feat/build-release/bellman/src/lib.rs#L39 and update the worker usage: https://github.com/hack-a-chain-software/phase2-bn254/blob/feat/build-release/phase2/src/parameters.rs#L8 and https://github.com/hack-a-chain-software/phase2-bn254/blob/feat/build-release/phase2/src/parameters.rs#L238 And now, when I send the "new" command to be executed, everything works correctly: (...)
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- write(1, 0x673c928, 0x27) ...
[Enclave] Computed result valid: true
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- return from write(...) = 0x27
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- write(1, 0x673c950, 0x20) ...
[Enclave] Adding to contract...
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- return from write(...) = 0x20
(...)
[Phala] Calling method addContribution with args: 1709904124 2 Postman prod ceremony Lorem ipsum dolor sit amet 1709221725 1709904124 [
{ name: 'power', value: '10' },
{ name: 'bash', value: '256' },
{ name: 'nPrvInputs', value: '0' },
{ name: 'nPubInputs', value: '1' },
{ name: 'nInputs', value: '1' },
{ name: 'nOutputs', value: '253' },
{ name: 'nVars', value: '508' },
{ name: 'nConstants', value: '0' },
{ name: 'nSignals', value: '508' }
] [
{
hash: 'QmcE8DfXtNKrg4R5nQNkb9qoKNw1u8QyBhKfhmR3YpmBY8',
name: 'circom1.params',
timestamp: 1709904163
}
]
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- return from write(...) = 0x24f
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- write(1, 0x673cbc0, 0x3e) ...
cert.address 5CvhWGmHRX29nMr22dSmmyRC786ryGEbqW8DiiRJmZ1ep2ub
(...)But, when I try to run the 'contribute' binary, the previous error reappears: (libos_parser.c:1658:buf_write_all) [P2:T12:contribute] trace: ---- openat(AT_FDCWD, "/proc/self/cgroup", O_RDONLY|0x80000, 0000) = -2
(...)
(pal_exception.c:289:_PalExceptionHandler) error: Host injected malicious signal 2
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- return from read(...) = 0x1c
(pal_process.c:248:_PalProcessExit) debug: PalProcessExit: Returning exit code 1
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- clock_gettime(6, 0x49388910) = 0x0
(...)
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- epoll_wait(13, 0x49385900, 1024, 4612) ...
(libos_ipc_child.c:21:ipc_child_disconnect_callback) [P1:libos] error: Child process (vmid: 0x2) got disconnected
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- return from epoll_wait(...) = 0x2
(libos_context.c:279:prepare_sigframe) [P1:T1:node] debug: Created sigframe for sig: 17 at 0x49385250 (handler: 0x1556a40, restorer: 0x476036f0)
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- read(4, 0x4938520f, 0x1) ...
(...)
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- return from write(...) = 0x1
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- rt_sigreturn()
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- clock_gettime(6, 0x49385800) = 0x0
(...)
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- clock_gettime(1, 0x493849f0) = 0x0
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- mprotect(0x37f43000, 0x3c000, PROT_READ|PROT_EXEC) ...
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- return from mprotect(...) = 0x0
(libos_epoll.c:433:do_epoll_del) [P1:T1:node] debug: epoll: deleted 26 (0x7f8b7aa8) from epoll handle 0x7f8b5c08
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- epoll_ctl(13, DEL, 26, {.events=0, .data=0}) = 0x0
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- close(26) = 0x0
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- read(28, 0x59f2390, 0x10000) ...
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- return from read(...) = 0x0
(libos_epoll.c:433:do_epoll_del) [P1:T1:node] debug: epoll: deleted 28 (0x7f62a018) from epoll handle 0x7f8b5c08
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- epoll_ctl(13, DEL, 28, {.events=0, .data=0}) = 0x0
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- close(28) = 0x0
(...)
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- epoll_ctl(13, DEL, 23, {.events=0, .data=0}) = -2
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- close(23) = 0x0
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- rt_sigprocmask(SETMASK, [SIGHUP,SIGINT,SIGQUIT,SIGILL,SIGTRAP,SIGABRT,SIGBUS,SIGFPE,SIGKILL,SIGUSR1,SIGSEGV,SIGUSR2,SIGPIPE,SIGALRM,SIGTERM,SIGSTKFLT,SIGCHLD,SIGCONT,SIGSTOP,SIGTSTP,SIGTTIN,SIGTTOU,SIGURG,SIGXCPU,SIGXFSZ,SIGVTALRM,SIGPROF,SIGWINCH,SIG
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: IO,SIGPWR,SIGSYS,SIG34,SIG35,SIG36,SIG37,SIG38,SIG39,SIG40,SIG41,SIG42,SIG43,SIG44,SIG45,SIG46,SIG47,SIG48,SIG49,SIG50,SIG51,SIG52,SIG53,SIG54,SIG55,SIG56,SIG57,SIG58,SIG59,SIG60,SIG61,SIG62,SIG63,SIG64,], [], 0x8) = 0x0
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- read(4, 0x49384a8f, 0x1) ...
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- return from read(...) = 0x1
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- rt_sigaction([SIGCHLD], 0x493849e0, 0, 0x8) = 0x0
(...)
[Enclave] Error: Error: Command failed: ./app/bin/contribute circom1.params circom2.params
at ChildProcess.exithandler (node:child_process:402:12)
at ChildProcess.emit (node:events:513:28)
at maybeClose (node:internal/child_process:1100:16)
at Process.ChildProcess._handle.onexit (node:internal/child_process:304:5) {
code: null,
killed: false,
signal: 'SIGPWR',
cmd: './app/bin/contribute circom1.params circom2.params',
stdout: 'Contributing to circom1.params...\n' +
'MPCParameters::contribute()\n' +
'MPCParameters::batch_exp1()\n' +
'MPCParameters::batch_exp2()\n' +
'MPCParameters::delta_inv\n',
stderr: ''
}
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- return from write(...) = 0x294
(libos_parser.c:1658:buf_write_all) [P1:T1:node] trace: ---- clock_gettime(1, 0x49384540) = 0x0And by adding more logs to the binary's dependencies, I found out it breaks when executing this function: https://github.com/hack-a-chain-software/phase2-bn254/blob/feat/build-release/phase2/src/parameters.rs#L465 Do you have any idea what the problem might be? |
Beta Was this translation helpful? Give feedback.
-
|
Thanks for the logs and experiments! I think we have our suspect: (same for other logs that you showed) Gramine does not implement It looks like whatever Rust crate you're using ( |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hello everyone, I'm working with Gramine on a project where we're running a Node.js instance that executes some binaries that were built using Rust.
heres the manifest: https://github.com/hack-a-chain-software/bazk/blob/main/packages/gramine/bazk-build/bazk.manifest.template
We use Node.js to execute some binaries that were built with Rust:
https://github.com/hack-a-chain-software/bazk/blob/main/packages/gramine/src/index.ts#L177
I can correctly compile the binaries and launch the enclave:
When i send a command to execute the 'new_constrained' binary, it executes correctly: https://github.com/hack-a-chain-software/phase2-bn254/blob/feat/gramine/powersoftau/src/bin/new_constrained.rs#L12
But, when I send a command to execute the 'new' binary: https://github.com/hack-a-chain-software/phase2-bn254/blob/feat/gramine/phase2/src/bin/new.rs#L9
I receive this error:
Signal: 'SIGSEGV'
I add logs inside the binaries and discovered that the process fails when this function inside the 'new' command binary tries to execute a new worker: https://github.com/hack-a-chain-software/phase2-bn254/blob/feat/gramine/phase2/src/parameters.rs#L238
thanks for help.:
Beta Was this translation helpful? Give feedback.
All reactions