Received-SPF: pass (google.com: domain of linux-kernel+bounces-189172-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Date: Sat, 25 May 2024 00:35:52 +0200
From: Brian Johannesmeyer <bjohannesmeyer@gmail.com>
To: Alexander Potapenko <glider@google.com>
Cc: Marco Elver <elver@google.com>, Dmitry Vyukov <dvyukov@google.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	kasan-dev@googlegroups.com, linux-mm@kvack.org,
	linux-kernel@vger.kernel.org, Thomas Gleixner <tglx@linutronix.de>,
	Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
	Dave Hansen <dave.hansen@linux.intel.com>, x86@kernel.org,
	"H . Peter Anvin" <hpa@zytor.com>
Subject: Re: [PATCH] x86: kmsan: Fix hook for unaligned accesses
Message-ID: <ZlEWSIcHXnh/BqbW@rex>
References: <20240523215029.4160518-1-bjohannesmeyer@gmail.com>
 <CAG_fn=XR6KVQ=DbKZW3kNXsCHgULm2J7i6GCm8CZUjpjuk-d2A@mail.gmail.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CAG_fn=XR6KVQ=DbKZW3kNXsCHgULm2J7i6GCm8CZUjpjuk-d2A@mail.gmail.com>

On Fri, May 24, 2024 at 10:28:05AM +0200, Alexander Potapenko wrote:
> Nice catch! Does it fix any known bugs?

Not that I know of. Based on my cursory testing, it seems that
string_memcpy_fromio() is rarely called with an unaligned `from`, so
this is a bit of an edge case.

On that note: I tried creating a unit test for this, to verify that
an unaligned memcpy_fromio() would yield uninitialized data without the
patch, and would yield initialized data with the patch. However, what I
found is that kmsan_unpoison_memory() seems to always unpoison an entire
4-byte word, even if called with a `size` of less than 4. However, this
issue is somewhat unrelated to the patch at hand, so I'll create a
separate patch to demonstrate what I mean.

Thanks,
Brian