syscall: Getfsstat() doesn't work on freebsd/amd64 #6588
Labels
Milestone
Comments
|
Comment 2 by juphoff@vividcortex.com: Calling C.getfsstat() directly via Cgo exhibits the same behavior--a correct number of structs are returned, but they're entirely filled with 0x0's. |
|
Comment 4 by preetam@vividcortex.com: Attached. Attachments:
|
|
Comment 5 by preetam@vividcortex.com: This is for getfsstat.go from the first post. Attachments:
|
4487 getfsstat CALL getfsstat(0xc210047000,0x2,0x1<MNT_WAIT>) 4487 getfsstat RET getfsstat 0 ^ we asked for 2 fsstat results, we got 0. Moving to accepted, something strange is going on here, but I don't know if this can be prioritized for Go 1.2 Labels changed: added os-macosx. Status changed to Accepted. |
the first length = 0 is because we used the wrong pointer.
diff -r 050f1f96c25c src/pkg/syscall/zsyscall_darwin_amd64.go
--- a/src/pkg/syscall/zsyscall_darwin_amd64.go Tue Oct 15 21:35:52 2013 -0400
+++ b/src/pkg/syscall/zsyscall_darwin_amd64.go Wed Oct 16 00:41:03 2013 -0400
@@ -594,8 +594,6 @@
var _p0 unsafe.Pointer
if len(buf) > 0 {
_p0 = unsafe.Pointer(&buf[0])
- } else {
- _p0 = unsafe.Pointer(&_zero)
}
r0, _, e1 := Syscall(SYS_GETFSSTAT64, uintptr(_p0), uintptr(len(buf)), uintptr(flags))
n = int(r0)
the kernel is not checking nr, but whether _p0 is nil or not to determine whether to
return
the length. as can be seen from the above code snippet, we're always passing a non-nil
pointer to getfsstat.
I'm still not sure about the 2nd problem of returning all zero struct.
This is not a very big deal though, as the user could workaround by using syscall.Syscall
directly (just make sure to use SYS_GETFSSTAT64). Defer to Go 1.3.
Another reason is that although I gave a simple patch above, but there is still a long
way
to go before I can make a CL (we need to fix the program that generates the zsyscall
file,
and i think we'd need a general mechanism to annotate that certain parameter is allowed
to be nil, and then double check all syscall wrappers for the same problem; it will be a
somewhat big project).
Labels changed: added priority-later, go1.3, removed priority-triage. |
From getfsstat(2): int getfsstat(struct statfs *buf, long bufsize, int flags); [...] Note, the bufsize argument is the number of bytes that buf can hold, not the count of statfs structures it will hold. [...] syscall.Getfsstat() passes len(buf) as bufsize which is probably < sizeof(struct statfs) which explains why it doesn't touch buf at all. |
Passing unsafe.Sizeof(syscall.Statfs_t{}) * uintptr(len(buf)) as bufsize works as
expected (freebsd/386).
http://play.golang.org/p/ypiGCqFTn0
|
|
Comment 13 by pj@isomero.us: Could we just create a different function definition (and exclude it from being automatically generated) instead of modifying the script to handle this case? |
|
Comment 14 by preetam@vividcortex.com: Could we just create a different function definition (and exclude it from being automatically generated) instead of modifying the script to handle this case? |
CL https://golang.org/cl/84830043 references this issue. |
|
This issue was closed by revision 4abbd4a. Status changed to Fixed. |
This issue was closed.
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
by preetam@vividcortex.com:
go version go1.1.2 freebsd/amd64 The following program prints out structs filled with 0s, which shouldn't be the case. __________________________________________________________ package main import ( "fmt" "os" "syscall" ) func main() { len, err := syscall.Getfsstat(nil, 1) if err != nil { fmt.Println(err) os.Exit(1) } fmt.Println("Length:", len) data := make([]syscall.Statfs_t, len) _, err = syscall.Getfsstat(data, 1) if err != nil { fmt.Println(err) os.Exit(1) } fmt.Println("Data:", data) } __________________________________________________________The text was updated successfully, but these errors were encountered: