[Bug 211823] [patch] fsck_ffs incorrectly failing in check_cgmagic

[bugzilla-noreply at freebsd.org]

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=211823

--- Comment #6 from Christian Heckendorf <heckendorfc at gmail.com> ---
Both variables hold the same value in their internal representation (0xcc00)
but cg_old_niblk overflows into negative values due to type limitations that
fs_ipg doesn't have. Clearly check_cgmagic() thinks the fs data is invalid yet
somehow OpenBSD fsck thinks it works and the filesystem in general seems to
work fine on both OS, at least from a basic usage standpoint (I haven't looked
into anything more than basic I/O and whether fsck says clean or dirty). Is
this check in check_cgmagic meant to compare internal representations or their
signed/unsigned integer values?

Is general compatibility between UFS in OpenBSD and FreeBSD purely accidental?
If so, I'll drop the issue.

-- 
You are receiving this mail because:
You are the assignee for the bug.