--- title: AMD Posts Massive Patch Series For Enabling UALink In... description: AMD engineers submitted a large patch series to the Linux kernel adding UALink support for the AMDGPU driver. UALink is an open standard for high-speed... canonical: https://daily.dev/posts/amd-posts-massive-patch-series-for-enabling-ualink-in-the-linux-kernel-zr1srneqw twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: AMD Posts Massive Patch Series For Enabling UALink In The Linux Kernel | daily.dev og:description: AMD engineers submitted a large patch series to the Linux kernel adding UALink support for the AMDGPU driver. UALink is an open standard for high-speed... og:url: https://daily.dev/posts/amd-posts-massive-patch-series-for-enabling-ualink-in-the-linux-kernel-zr1srneqw og:image: https://api.daily.dev/og/posts/zr1Srneqw.png og:image:alt: AMD Posts Massive Patch Series For Enabling UALink In The Linux Kernel og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # AMD Posts Massive Patch Series For Enabling UALink In The Linux Kernel **[Phoronix](https://daily.dev/sources/phoronix)** · 2 min read · 0 upvotes · 0 comments ## Summary AMD engineers submitted a large patch series to the Linux kernel adding UALink support for the AMDGPU driver. UALink is an open standard for high-speed GPU-to-GPU interconnects in data center scale-up pods, supporting up to 1,024 accelerators in a single domain. The first series contains 95 patches (~7k lines) covering core infrastructure and AMDGPU integration, enabling direct remote memory access between GPUs without copying, using a shared Network Physical Address space. A second, smaller series of three patches adds a new IOCTL interface and documentation for memory import/export handling. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is UALink and how does it work for connecting GPUs in a data center? UALink (Ultra Accelerator Link) is an open standard, high-speed networking protocol for connecting GPUs and AI accelerators within scale-up pods, supporting up to 1,024 accelerators in a single domain. Unlike RDMA, it enables direct loads and stores over the fabric without memory copies, using a shared Network Physical Address (NPA) space partitioned across GPUs, with page tables similar to GPUVM managing each GPU's exported memory segment. _daily.dev surfaces kernel-level hardware interconnect developments like UALink for engineers tracking GPU infrastructure._ ### How large is AMD's initial Linux kernel patch series adding UALink support to the AMDGPU driver? The first patch series submitted by AMD engineers contains 95 patches totaling about 7,000 lines of code, adding core UALink infrastructure and integration into the AMDGPU kernel graphics driver. A follow-up second series of three patches adds a new IOCTL interface and documentation for UALink memory import and export handling. _Kernel contributors follow patch series like this on daily.dev to stay ahead of AMDGPU driver changes._ ### How does UALink handle memory access when a GPU node goes offline or is reset in a shared address space? When a node goes offline or is reset, its peer GPUs clean up any remaining references that are lost as a result. Exported memory in UALink is not pinned, so the exporter may need to move or invalidate it; this is managed through remote TLB shootdowns during TTM evictions or MMU notifier events, and remote presence checks with retry on the importer side. _Engineers evaluating GPU interconnect reliability can track UALink kernel integration details on daily.dev._ ## Similar posts on daily.dev - [AMD Posts Massive 109 Patch Series For GFX 12.1 RAS Support On Friday Evening](https://daily.dev/posts/amd-posts-massive-109-patch-series-for-gfx-12-1-ras-support-on-friday-evening-duhzbu67d) · Phoronix · 0 upvotes · 0 comments - [AMD Sends In More New GPU Hardware Enablement For Linux 7.4, Bug Fixes](https://daily.dev/posts/amd-sends-in-more-new-gpu-hardware-enablement-for-linux-7-4-bug-fixes-43cihok7b) · Phoronix · 0 upvotes · 0 comments - [AMD Begins Posting Open-Source Linux Patches For Their Next-Gen GPU IP](https://daily.dev/posts/amd-begins-posting-open-source-linux-patches-for-their-next-gen-gpu-ip-ibv4ix7eh) · Phoronix · 0 upvotes · 0 comments - [New AMD Linux Driver Patches Posted For Batch Userptr Allocation Support](https://daily.dev/posts/new-amd-linux-driver-patches-posted-for-batch-userptr-allocation-support-do5l9qjaj) · Phoronix · 0 upvotes · 0 comments --- Tags: [#linux](https://daily.dev/tags/linux), [#gpu](https://daily.dev/tags/gpu), [#amd](https://daily.dev/tags/amd) [View this post on daily.dev](https://daily.dev/posts/amd-posts-massive-patch-series-for-enabling-ualink-in-the-linux-kernel-zr1srneqw) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"AMD Posts Massive Patch Series For Enabling UALink In The Linux Kernel","url":"https://daily.dev/posts/amd-posts-massive-patch-series-for-enabling-ualink-in-the-linux-kernel-zr1srneqw","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/amd-posts-massive-patch-series-for-enabling-ualink-in-the-linux-kernel-zr1srneqw"},"datePublished":"2026-08-21T20:19:01.193Z","dateModified":"2026-08-21T20:19:26.251Z","description":"AMD engineers submitted a large patch series to the Linux kernel adding UALink support for the AMDGPU driver. UALink is an open standard for high-speed...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/907f41f520baebc6436eb8432fa08301?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/907f41f520baebc6436eb8432fa08301?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Phoronix","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Phoronix","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/6a26ba379dd64932830aa0c9803dd7ee","url":"https://daily.dev/sources/phoronix"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/amd-posts-massive-patch-series-for-enabling-ualink-in-the-linux-kernel-zr1srneqw","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"linux,gpu,amd","timeRequired":"PT2M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Phoronix","item":"https://daily.dev/sources/phoronix"},{"@type":"ListItem","position":3,"name":"AMD Posts Massive Patch Series For Enabling UALink In The Linux Kernel"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/amd-posts-massive-patch-series-for-enabling-ualink-in-the-linux-kernel-zr1srneqw#faq","mainEntity":[{"@type":"Question","name":"What is UALink and how does it work for connecting GPUs in a data center?","acceptedAnswer":{"@type":"Answer","text":"UALink (Ultra Accelerator Link) is an open standard, high-speed networking protocol for connecting GPUs and AI accelerators within scale-up pods, supporting up to 1,024 accelerators in a single domain. Unlike RDMA, it enables direct loads and stores over the fabric without memory copies, using a shared Network Physical Address (NPA) space partitioned across GPUs, with page tables similar to GPUVM managing each GPU's exported memory segment. daily.dev surfaces kernel-level hardware interconnect developments like UALink for engineers tracking GPU infrastructure."}},{"@type":"Question","name":"How large is AMD's initial Linux kernel patch series adding UALink support to the AMDGPU driver?","acceptedAnswer":{"@type":"Answer","text":"The first patch series submitted by AMD engineers contains 95 patches totaling about 7,000 lines of code, adding core UALink infrastructure and integration into the AMDGPU kernel graphics driver. A follow-up second series of three patches adds a new IOCTL interface and documentation for UALink memory import and export handling. Kernel contributors follow patch series like this on daily.dev to stay ahead of AMDGPU driver changes."}},{"@type":"Question","name":"How does UALink handle memory access when a GPU node goes offline or is reset in a shared address space?","acceptedAnswer":{"@type":"Answer","text":"When a node goes offline or is reset, its peer GPUs clean up any remaining references that are lost as a result. Exported memory in UALink is not pinned, so the exporter may need to move or invalidate it; this is managed through remote TLB shootdowns during TTM evictions or MMU notifier events, and remote presence checks with retry on the importer side. Engineers evaluating GPU interconnect reliability can track UALink kernel integration details on daily.dev."}}]} ```