LWN: Comments on "Improving iov_iter" https://lwn.net/Articles/1023891/ This is a special feed containing comments posted to the individual LWN article titled "Improving iov_iter". en-us Mon, 29 Sep 2025 22:59:24 +0000 Mon, 29 Sep 2025 22:59:24 +0000 https://www.rssboard.org/rss-specification lwn@lwn.net address representation and KASAN https://lwn.net/Articles/1025140/ https://lwn.net/Articles/1025140/ TheJH <div class="FormattedComment"> <span class="QuotedText">&gt; Most people seem to think that representation should be pairs of physical addresses and lengths, perhaps with a flag, he said.</span><br> <p> For what it's worth, from the perpective of address-tagging-based KASAN and the (not yet upstream) SLUB-virtual security mitigation, it would be nice if the kernel (in particular on 64-bit) stayed in the kernel-virtual-address representation as much as possible and only used physical addresses when actually initiating DMA operations or such.<br> <p> For example, some configurations of KASAN rely on tag bits in the upper part of an address being preserved for detection of memory safety bugs.<br> </div> Thu, 12 Jun 2025 13:12:45 +0000