Received: from mail.lichtvoll.de (luna.lichtvoll.de [194.150.191.11]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 49C6623AE for ; Sat, 30 Dec 2023 11:00:18 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=lichtvoll.de Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=lichtvoll.de Received: from 127.0.0.1 (localhost [127.0.0.1]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature ECDSA (secp384r1) server-digest SHA384) (No client certificate requested) by mail.lichtvoll.de (Postfix) with ESMTPSA id 464B285AD37; Sat, 30 Dec 2023 12:00:16 +0100 (CET) Authentication-Results: mail.lichtvoll.de; auth=pass smtp.auth=martin smtp.mailfrom=martin@lichtvoll.de From: Martin Steigerwald To: linux-bcachefs@vger.kernel.org Cc: Kent Overstreet Subject: Testing xxhash Date: Sat, 30 Dec 2023 12:00:15 +0100 Message-ID: <1885915.tdWV9SEqCh@lichtvoll.de> Precedence: bulk X-Mailing-List: linux-bcachefs@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: 7Bit Content-Type: text/plain; charset="us-ascii" Hi Kent, hi everyone, While still evaluating BCacheFS I thought I could give xxhash checksumming a try: metadata_checksum: none crc32c crc64 [xxhash] data_checksum: none crc32c crc64 [xxhash] compression: lz4 background_compression: none str_hash: crc32c crc64 [siphash] Anything to watch out for? So far everything seems to work just fine. As for now no scrubbing available or is there? Does BCacheFS verify checksums on reads? I am still not sure whether anything more than crc32c is really needed. But on the other hand even with already a bit older processors like AMD Ryzen 4750U with 8 cores I bet checksumming with xxhash should not be CPU bound. Actually on devices with fast enough devices I use xxhash with BTRFS since more than a year already and scrubbing goes between 1,6 and 2,6 GiB/s with NVME SSDs. And regarding hash collisions I'd rather be safe than sorry. Best, -- Martin