Stream: git-wasmtime

Topic: wasmtime / issue #3303 Behavior of fd_write changed, seem...

view this post on Zulip Wasmtime GitHub notifications bot (Sep 05 2021 at 13:38):

DenialAdams opened issue #3303:

I recently updated my local version of wasmtime and noticed that the behavior of fd_write seems to have changed. I am supplying two strings to print: a supplied string, and a newline character. This used to print the newline character, which is what I expect, but seems to no longer do so. I bisected the behavior change back to PR #2444, but I didn't see this behavior change explicitly called out.

Basically, I'm wondering:

1) Is this new behavior intended, or a bug?
2) If this new behavior is intended, what am I doing wrong in the example program?

Thanks for reading!

Test Case

(module
  (import "wasi_unstable" "fd_write" (func $fd_write (param i32 i32 i32 i32) (result i32)))
  (memory 1)
  (export "memory" (memory 0))
  (data 0 (i32.const 16) "\n")
  (data 0 (i32.const 18) "world on a new line!")
  (data 0 (i32.const 47) "Hello,")
  (global $mem_address (mut i32) (i32.const 0))
  (func $print (param i32) (param i32)
    (i32.store (i32.const 0) (local.get 0))
    (i32.store (i32.const 4) (local.get 1))
    (i32.store (i32.const 8) (i32.const 16))
    (i32.store (i32.const 12) (i32.const 1))
    (call $fd_write (i32.const 1) (i32.const 0) (i32.const 2) (i32.const 0))
    drop
  )
  (func $main
    (export "_start")
    i32.const 47
    i32.const 6
    call $print
    i32.const 18
    i32.const 20
    call $print
  )
)

Steps to Reproduce

1) wasmtime repro.wast (contents above)
2) Observe that the newline is not printed

Expected Results

Hello,
world on a new line!

Actual Results

Hello,world on a new line!

Versions and Environment

I first observed this behavior on wasmtime 0.29.0. I bisected the issue back to PR #2444, which seems to be where the behavior changed.

Operating system: Windows 10

Architecture: x86

view this post on Zulip Wasmtime GitHub notifications bot (Sep 05 2021 at 13:38):

DenialAdams labeled issue #3303:

I recently updated my local version of wasmtime and noticed that the behavior of fd_write seems to have changed. I am supplying two strings to print: a supplied string, and a newline character. This used to print the newline character, which is what I expect, but seems to no longer do so. I bisected the behavior change back to PR #2444, but I didn't see this behavior change explicitly called out.

Basically, I'm wondering:

1) Is this new behavior intended, or a bug?
2) If this new behavior is intended, what am I doing wrong in the example program?

Thanks for reading!

Test Case

(module
  (import "wasi_unstable" "fd_write" (func $fd_write (param i32 i32 i32 i32) (result i32)))
  (memory 1)
  (export "memory" (memory 0))
  (data 0 (i32.const 16) "\n")
  (data 0 (i32.const 18) "world on a new line!")
  (data 0 (i32.const 47) "Hello,")
  (global $mem_address (mut i32) (i32.const 0))
  (func $print (param i32) (param i32)
    (i32.store (i32.const 0) (local.get 0))
    (i32.store (i32.const 4) (local.get 1))
    (i32.store (i32.const 8) (i32.const 16))
    (i32.store (i32.const 12) (i32.const 1))
    (call $fd_write (i32.const 1) (i32.const 0) (i32.const 2) (i32.const 0))
    drop
  )
  (func $main
    (export "_start")
    i32.const 47
    i32.const 6
    call $print
    i32.const 18
    i32.const 20
    call $print
  )
)

Steps to Reproduce

1) wasmtime repro.wast (contents above)
2) Observe that the newline is not printed

Expected Results

Hello,
world on a new line!

Actual Results

Hello,world on a new line!

Versions and Environment

I first observed this behavior on wasmtime 0.29.0. I bisected the issue back to PR #2444, which seems to be where the behavior changed.

Operating system: Windows 10

Architecture: x86

view this post on Zulip Wasmtime GitHub notifications bot (Sep 05 2021 at 13:39):

DenialAdams edited issue #3303:

I recently updated my local version of wasmtime and noticed that the behavior of fd_write seems to have changed. I am supplying two strings to print: a supplied string, and a newline character. This used to print the newline character, which is what I expect, but seems to no longer do so. I bisected the behavior change back to PR #2444, but I didn't see this behavior change explicitly called out.

Basically, I'm wondering:

1) Is this new behavior intended, or a bug?
2) If this new behavior is intended, what am I doing wrong in the example program?

Thanks for reading!

Test Case

(module
  (import "wasi_unstable" "fd_write" (func $fd_write (param i32 i32 i32 i32) (result i32)))
  (memory 1)
  (export "memory" (memory 0))
  (data 0 (i32.const 16) "\n")
  (data 0 (i32.const 18) "world on a new line!")
  (data 0 (i32.const 47) "Hello,")
  (func $print (param i32) (param i32)
    (i32.store (i32.const 0) (local.get 0))
    (i32.store (i32.const 4) (local.get 1))
    (i32.store (i32.const 8) (i32.const 16))
    (i32.store (i32.const 12) (i32.const 1))
    (call $fd_write (i32.const 1) (i32.const 0) (i32.const 2) (i32.const 0))
    drop
  )
  (func $main
    (export "_start")
    i32.const 47
    i32.const 6
    call $print
    i32.const 18
    i32.const 20
    call $print
  )
)

Steps to Reproduce

1) wasmtime repro.wast (contents above)
2) Observe that the newline is not printed

Expected Results

Hello,
world on a new line!

Actual Results

Hello,world on a new line!

Versions and Environment

I first observed this behavior on wasmtime 0.29.0. I bisected the issue back to PR #2444, which seems to be where the behavior changed.

Operating system: Windows 10

Architecture: x86

view this post on Zulip Wasmtime GitHub notifications bot (Sep 05 2021 at 13:41):

bjorn3 commented on issue #3303:

fd_write returns the amount of bytes written. You have to call it in a loop until all bytes are written. This matches the posix write syscall. From the man page of the write syscall:

The number of bytes written may be less than count if, for example, there is insufficient space on the underlying physical medium, or the RLIMIT_FSIZE resource limit is encountered (see setrlimit(2)), or the call was interrupted by a signal handler after having written less than count bytes. (See also pipe(7).)

Higher level api's generally do this loop for you.

view this post on Zulip Wasmtime GitHub notifications bot (Sep 05 2021 at 13:42):

DenialAdams commented on issue #3303:

fd_write returns the amount of bytes written. You have to call it in a loop until all bytes are written. This matches the posix write syscall. From the man page of the write syscall:

The number of bytes written may be less than count if, for example, there is insufficient space on the underlying physical medium, or the RLIMIT_FSIZE resource limit is encountered (see setrlimit(2)), or the call was interrupted by a signal handler after having written less than count bytes. (See also pipe(7).)

Higher level api's generally do this loop for you.

Thank you, this makes a lot of sense! Sorry for the false report

view this post on Zulip Wasmtime GitHub notifications bot (Sep 05 2021 at 13:42):

DenialAdams closed issue #3303:

I recently updated my local version of wasmtime and noticed that the behavior of fd_write seems to have changed. I am supplying two strings to print: a supplied string, and a newline character. This used to print the newline character, which is what I expect, but seems to no longer do so. I bisected the behavior change back to PR #2444, but I didn't see this behavior change explicitly called out.

Basically, I'm wondering:

1) Is this new behavior intended, or a bug?
2) If this new behavior is intended, what am I doing wrong in the example program?

Thanks for reading!

Test Case

(module
  (import "wasi_unstable" "fd_write" (func $fd_write (param i32 i32 i32 i32) (result i32)))
  (memory 1)
  (export "memory" (memory 0))
  (data 0 (i32.const 16) "\n")
  (data 0 (i32.const 18) "world on a new line!")
  (data 0 (i32.const 47) "Hello,")
  (func $print (param i32) (param i32)
    (i32.store (i32.const 0) (local.get 0))
    (i32.store (i32.const 4) (local.get 1))
    (i32.store (i32.const 8) (i32.const 16))
    (i32.store (i32.const 12) (i32.const 1))
    (call $fd_write (i32.const 1) (i32.const 0) (i32.const 2) (i32.const 0))
    drop
  )
  (func $main
    (export "_start")
    i32.const 47
    i32.const 6
    call $print
    i32.const 18
    i32.const 20
    call $print
  )
)

Steps to Reproduce

1) wasmtime repro.wast (contents above)
2) Observe that the newline is not printed

Expected Results

Hello,
world on a new line!

Actual Results

Hello,world on a new line!

Versions and Environment

I first observed this behavior on wasmtime 0.29.0. I bisected the issue back to PR #2444, which seems to be where the behavior changed.

Operating system: Windows 10

Architecture: x86

Last updated: Nov 22 2024 at 17:03 UTC