Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 22 Jan 2025 20:04:54 +0000
From:      bugzilla-noreply@freebsd.org
To:        fs@FreeBSD.org
Subject:   [Bug 281402] zfs: kmsan reports Uninitialized UMA memory from zio_data_buf_alloc
Message-ID:  <bug-281402-3630-GUTV1D6Hna@https.bugs.freebsd.org/bugzilla/>
In-Reply-To: <bug-281402-3630@https.bugs.freebsd.org/bugzilla/>
References:  <bug-281402-3630@https.bugs.freebsd.org/bugzilla/>

next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D281402

--- Comment #6 from Mark Johnston <markj@FreeBSD.org> ---
(In reply to Alan Somers from comment #5)
Oh, yes, that's quite possible if the accelerated routines are implemented =
in
assembly or use intrinsics, or there's some bug in the build which causes s=
ome
file(s) to be compiled without -fsanitize=3Dkernel-memory.  I believe LLVM =
has
some ability to instrument inline asm blocks based on the input and output
operand descriptions, not sure if that's relevant here.

It should be possible to use kmsan_mark(buf, sz, KMSAN_STATE_INITED) to
manually mark a region as initialized.  You might also want to use
kmsan_check() to ensure that the inputs are initialized.  I presume that
Linux's KMSAN has similar interfaces, but I'm not sure what those are.

--=20
You are receiving this mail because:
You are the assignee for the bug.=



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-281402-3630-GUTV1D6Hna>