[Bug] Kernel panic on Intel Mac with AMD dGPU during heavy parallel Claude Code CLI usage - WindowServer watchdog timeout
Bug Description
Title: Intel MacBook Pro (AMD dGPU) kernel panics correlating with heavy/parallel Claude Code CLI usage — WindowServer watchdog timeout
Environment: MacBook Pro 16" 2019, Intel Core i9-9980HK, 32GB RAM, macOS 26.5.2 (25F84), AMD Radeon Pro 5500M + Intel UHD 630 (auto-switching), iTerm2, Claude Code CLI.
Observed: 2 full kernel panics within ~7 hours (2026-07-01 13:34:58 UTC, 2026-07-01 20:38:37 UTC), both panic(...): userspace watchdog timeout: no successful checkins from WindowServer (2 induced crashes) in 120 seconds. Followed by 20 AMD GPU Reset events over the next ~7.5 hours.
One panic occurred 55 seconds after invoking the Workflow tool under Fable 5 + /effort ultracode (multi-agent parallel orchestration mode). The other panic occurred during Opus 4.8/Sonnet 5 usage without Fable 5 active, so this is not exclusively tied to one model — but it suggests sustained heavy/parallel CLI tool-use load (multiple concurrent sessions, high-frequency terminal output, Workflow multi-agent spawning) stresses WindowServer/AMD GPU driver stability on Intel Macs with discrete GPUs, not observed on Apple Silicon.
Ask: Any guidance on reducing terminal-rendering / concurrent-process load from Claude Code on Intel + dGPU Macs, or whether Workflow's concurrency cap should be more conservative on such hardware.
Environment Info
- Platform: darwin
- Terminal: iTerm.app
- Version: 2.1.201
- Feedback ID: 3e45af88-e1d4-4781-a046-4feb9d5c75b1
Errors
3 Comments
<h2>English Comment (Ready to Post)</h2><p>Same machine, same macOS version, same GPU — but I'm <strong>not</strong> using Claude Code at all. This confirms the root cause is macOS 26.5.2's AMD GPU driver, not any specific application.</p><p><strong>Machine:</strong> MacBook Pro 16" 2019 (MacBookPro16,1), Intel Core i9, 64GB RAM
<strong>GPU:</strong> AMD Radeon Pro 5500M + Intel UHD 630
<strong>macOS:</strong> 26.5.2 (25F84)</p><p><strong>Timeline — directly correlates with the 26.5.2 update:</strong></p><ul><li>June 2: Installed macOS 26.5.1 (25F80)</li></ul><ul><li>June 2 – June 29: <strong>28 days, zero crashes</strong></li></ul><ul><li>June 30: macOS 26.5.2 (25F84) released → <strong>first GPU Reset at 13:34, first shutdown stall at 17:52, same day</strong></li></ul><ul><li>June 30 – July 7: 9 shutdown stalls + 20+ GPU Resets, all on AMD 5500M</li></ul><p><strong>Controlled experiments (gpuswitch):</strong></p>
Setting | Result
-- | --
gpuswitch 2 (auto, default) | Random crashes during normal use and sleep/wake
gpuswitch 1 (forced dGPU) | 11 GPU Resets in 12 minutes, WindowServer crash — AMD GPU is fundamentally unstable under 26.5.2
gpuswitch 0 (forced iGPU) | No crashes during normal use, but sleep/wake still triggers kernel panic (PCIe power state recovery fails)
gpuswitch 0 + disablesleep 1 | Stable — GPU Resets still occur occasionally but self-recover without system crash
<p><strong>Key findings:</strong></p><ol><li><strong>Not application-specific:</strong> I don't use Claude Code. Crashes happen with WindowServer, mediaanalysisd, and other system processes — whatever touches the AMD GPU.</li></ol><ol><li><strong>Sleep/wake is a major trigger:</strong> Panic logs show <code>PCIeUp link state: 0x1614</code> on every crash. The AMD GPU fails to recover its PCIe link after wake, causing the watchdog timeout. This matches <code>Signature: 2, Restart Channel: 18 VMPT</code> in GPU Reset reports.</li></ol><ol><li><strong>Started exactly with 26.5.2:</strong> 28 days of zero crashes on 26.5.1, then immediate failures on 26.5.2 release day. The update clearly changed AMD GPU power management / PCIe link recovery behavior.</li></ol><ol><li><strong>Workaround:</strong> <code>sudo pmset -a gpuswitch 0</code> + <code>sudo pmset -a disablesleep 1</code> keeps the machine stable, but disabling sleep is not a real fix.</li></ol><p><strong>Panic log excerpt:</strong></p><pre><code></code></pre><p>This is a macOS 26.5.2 regression affecting all MacBookPro16,1 with AMD 5500M, independent of workload. Would appreciate Apple addressing this in 26.5.3 or 26.6.</p><sc-clipboard block-data="[{"blockId":"d4A5FlN65Bxdn1X8ifnRSl","blockSubtree":{"block":{"d4A5FlN65Bxdn1X8ifnRSl":{"role":"editor","value":{"id":"d4A5FlN65Bxdn1X8ifnRSl","parentId":"","type":"header2","props":{"title":[["English Comment (Ready to Post)",null]]},"children":[],"parentTable":"block","enabled":true}}}}},{"blockId":"SXaX4RUoCbLVQKa3GoxD0n","blockSubtree":{"block":{"SXaX4RUoCbLVQKa3GoxD0n":{"role":"editor","value":{"id":"SXaX4RUoCbLVQKa3GoxD0n","parentId":"","type":"text","props":{"title":[["Same machine, same macOS version, same GPU — but I'm ",null],["not",[["b"]]],[" using Claude Code at all. This confirms the root cause is macOS 26.5.2's AMD GPU driver, not any specific application.",null]]},"children":[],"parentTable":"block","enabled":true}}}}},{"blockId":"aRBFBdNojiahEeiwHpCiw3","blockSubtree":{"block":{"aRBFBdNojiahEeiwHpCiw3":{"role":"editor","value":{"id":"aRBFBdNojiahEeiwHpCiw3","parentId":"","type":"text","props":{"title":[["Machine:",[["b"]]],[" MacBook Pro 16\" 2019 (MacBookPro16,1), Intel Core i9, 64GB RAM\n",null],["GPU:",[["b"]]],[" AMD Radeon Pro 5500M + Intel UHD 630\n",null],["macOS:",[["b"]]],[" 26.5.2 (25F84)",null]]},"children":[],"parentTable":"block","enabled":true}}}}},{"blockId":"AkIor6f6etvZNSDTCAcrmC","blockSubtree":{"block":{"AkIor6f6etvZNSDTCAcrmC":{"role":"editor","value":{"id":"AkIor6f6etvZNSDTCAcrmC","parentId":"","type":"text","props":{"title":[["Timeline — directly correlates with the 26.5.2 update:",[["b"]]]]},"children":[],"parentTable":"block","enabled":true}}}}},{"blockId":"4IE84Ad7lUP1bpo4lSk4kk","blockSubtree":{"block":{"4IE84Ad7lUP1bpo4lSk4kk":{"role":"editor","value":{"id":"4IE84Ad7lUP1bpo4lSk4kk","parentId":"","type":"bulleted_list","props":{"title":[["June 2: Installed macOS 26.5.1 (25F80)",null]]},"children":[],"parentTable":"block","enabled":true}}}}},{"blockId":"ipS9BH3FzGrcN4aOuTmyRW","blockSubtree":{"block":{"ipS9BH3FzGrcN4aOuTmyRW":{"role":"editor","value":{"id":"ipS9BH3FzGrcN4aOuTmyRW","parentId":"","type":"bulleted_list","props":{"title":[["June 2 – June 29: ",null],["28 days, zero crashes",[["b"]]]]},"children":[],"parentTable":"block","enabled":true}}}}},{"blockId":"EOGLKO7kKQ23fiA4LbFrlE","blockSubtree":{"block":{"EOGLKO7kKQ23fiA4LbFrlE":{"role":"editor","value":{"id":"EOGLKO7kKQ23fiA4LbFrlE","parentId":"","type":"bulleted_list","props":{"title":[["June 30: macOS 26.5.2 (25F84) released → ",null],["first GPU Reset at 13:34, first shutdown stall at 17:52, same day",[["b"]]]]},"children":[],"parentTable":"block","enabled":true}}}}},{"blockId":"kkLOoX5gj5urTSDmRpl2Y2","blockSubtree":{"block":{"kkLOoX5gj5urTSDmRpl2Y2":{"role":"editor","value":{"id":"kkLOoX5gj5urTSDmRpl2Y2","parentId":"","type":"bulleted_list","props":{"title":[["June 30 – July 7: 9 shutdown stalls + 20+ GPU Resets, all on AMD 5500M",null]]},"children":[],"parentTable":"block","enabled":true}}}}},{"blockId":"Y54zP122nz8TZhULk8RLM1","blockSubtree":{"block":{"Y54zP122nz8TZhULk8RLM1":{"role":"editor","value":{"id":"Y54zP122nz8TZhULk8RLM1","parentId":"","type":"text","props":{"title":[["Controlled experiments (gpuswitch):",[["b"]]]]},"children":[],"parentTable":"block","enabled":true}}}}},{"blockId":"5dcod8EfIwP7NsltVx0NfA","blockSubtree":{"block":{"5dcod8EfIwP7NsltVx0NfA":{"role":"editor","value":{"id":"5dcod8EfIwP7NsltVx0NfA","parentId":"","type":"simple_table","props":{"tableColumnOrder":["4vrXXszI","qdvsChcS"],"tableColumnWidthMap":{"4vrXXszI":120,"qdvsChcS":120},"tableType":"fit"},"children":["cT8HhmEpAnBop1zyjc911d","O72GoxfW09o2xwJUkw69JN","5fo7NTnGDx6jBSrMWPbf0H","1l48j6ucH92JKhLswB0Bqb","oivfpLJqlCMwJjwfEcqATq"],"parentTable":"block","enabled":true}},"cT8HhmEpAnBop1zyjc911d":{"role":"editor","value":{"id":"cT8HhmEpAnBop1zyjc911d","parentId":"5dcod8EfIwP7NsltVx0NfA","type":"simple_table_row","props":{},"children":["GkdJqNx3p2ZvMmvsS6doV2","UIDlWKeBKPXXKTmh3BgUXP"],"parentTable":"block","enabled":true}},"GkdJqNx3p2ZvMmvsS6doV2":{"role":"editor","value":{"id":"GkdJqNx3p2ZvMmvsS6doV2","parentId":"cT8HhmEpAnBop1zyjc911d","type":"simple_table_cell","props":{"columnKey":"4vrXXszI"},"children":["h5CafPpcSXzAHPwH7HLHxa"],"parentTable":"block","enabled":true}},"h5CafPpcSXzAHPwH7HLHxa":{"role":"editor","value":{"id":"h5CafPpcSXzAHPwH7HLHxa","parentId":"GkdJqNx3p2ZvMmvsS6doV2","type":"text","props":{"title":[["Setting",null]]},"children":[],"parentTable":"block","enabled":true}},"UIDlWKeBKPXXKTmh3BgUXP":{"role":"editor","value":{"id":"UIDlWKeBKPXXKTmh3BgUXP","parentId":"cT8HhmEpAnBop1zyjc911d","type":"simple_table_cell","props":{"columnKey":"qdvsChcS"},"children":["cFkP3ODu3yC2NDcNrBfSbI"],"parentTable":"block","enabled":true}},"cFkP3ODu3yC2NDcNrBfSbI":{"role":"editor","value":{"id":"cFkP3ODu3yC2NDcNrBfSbI","parentId":"UIDlWKeBKPXXKTmh3BgUXP","type":"text","props":{"title":[["Result",null]]},"children":[],"parentTable":"block","enabled":true}},"O72GoxfW09o2xwJUkw69JN":{"role":"editor","value":{"id":"O72GoxfW09o2xwJUkw69JN","parentId":"5dcod8EfIwP7NsltVx0NfA","type":"simple_table_row","props":{},"children":["NDRVwyxqoGW52uJjUKIJ8E","eRMgGWpPPgZm6o0e16QwEe"],"parentTable":"block","enabled":true}},"NDRVwyxqoGW52uJjUKIJ8E":{"role":"editor","value":{"id":"NDRVwyxqoGW52uJjUKIJ8E","parentId":"O72GoxfW09o2xwJUkw69JN","type":"simple_table_cell","props":{"columnKey":"4vrXXszI"},"children":["0TszH6eBgYvLBTG0elsuPV"],"parentTable":"block","enabled":true}},"0TszH6eBgYvLBTG0elsuPV":{"role":"editor","value":{"id":"0TszH6eBgYvLBTG0elsuPV","parentId":"NDRVwyxqoGW52uJjUKIJ8E","type":"text","props":{"title":[["gpuswitch 2",[["c"]]],[" (auto, default)",null]]},"children":[],"parentTable":"block","enabled":true}},"eRMgGWpPPgZm6o0e16QwEe":{"role":"editor","value":{"id":"eRMgGWpPPgZm6o0e16QwEe","parentId":"O72GoxfW09o2xwJUkw69JN","type":"simple_table_cell","props":{"columnKey":"qdvsChcS"},"children":["c7ReQ26rpka922K05LV2Gv"],"parentTable":"block","enabled":true}},"c7ReQ26rpka922K05LV2Gv":{"role":"editor","value":{"id":"c7ReQ26rpka922K05LV2Gv","parentId":"eRMgGWpPPgZm6o0e16QwEe","type":"text","props":{"title":[["Random crashes during normal use and sleep/wake",null]]},"children":[],"parentTable":"block","enabled":true}},"5fo7NTnGDx6jBSrMWPbf0H":{"role":"editor","value":{"id":"5fo7NTnGDx6jBSrMWPbf0H","parentId":"5dcod8EfIwP7NsltVx0NfA","type":"simple_table_row","props":{},"children":["PYL2rOSGe0UXZxrQaIJSj4","WHtK9o7S8T6wuB91u1KIar"],"parentTable":"block","enabled":true}},"PYL2rOSGe0UXZxrQaIJSj4":{"role":"editor","value":{"id":"PYL2rOSGe0UXZxrQaIJSj4","parentId":"5fo7NTnGDx6jBSrMWPbf0H","type":"simple_table_cell","props":{"columnKey":"4vrXXszI"},"children":["MAc20oTsoWqOTcdMv8Sq2u"],"parentTable":"block","enabled":true}},"MAc20oTsoWqOTcdMv8Sq2u":{"role":"editor","value":{"id":"MAc20oTsoWqOTcdMv8Sq2u","parentId":"PYL2rOSGe0UXZxrQaIJSj4","type":"text","props":{"title":[["gpuswitch 1",[["c"]]],[" (forced dGPU)",null]]},"children":[],"parentTable":"block","enabled":true}},"WHtK9o7S8T6wuB91u1KIar":{"role":"editor","value":{"id":"WHtK9o7S8T6wuB91u1KIar","parentId":"5fo7NTnGDx6jBSrMWPbf0H","type":"simple_table_cell","props":{"columnKey":"qdvsChcS"},"children":["6JhB3qXfwbx5EyKcLJ4G0b"],"parentTable":"block","enabled":true}},"6JhB3qXfwbx5EyKcLJ4G0b":{"role":"editor","value":{"id":"6JhB3qXfwbx5EyKcLJ4G0b","parentId":"WHtK9o7S8T6wuB91u1KIar","type":"text","props":{"title":[["11 GPU Resets in 12 minutes",[["b"]]],[", WindowServer crash — AMD GPU is fundamentally unstable under 26.5.2",null]]},"children":[],"parentTable":"block","enabled":true}},"1l48j6ucH92JKhLswB0Bqb":{"role":"editor","value":{"id":"1l48j6ucH92JKhLswB0Bqb","parentId":"5dcod8EfIwP7NsltVx0NfA","type":"simple_table_row","props":{},"children":["0t4ssf3S27wr7CSM6PfHAl","2YcsThYR9fArdrqTw1myKr"],"parentTable":"block","enabled":true}},"0t4ssf3S27wr7CSM6PfHAl":{"role":"editor","value":{"id":"0t4ssf3S27wr7CSM6PfHAl","parentId":"1l48j6ucH92JKhLswB0Bqb","type":"simple_table_cell","props":{"columnKey":"4vrXXszI"},"children":["rxGoh4ZELfiH6ljZpHcFps"],"parentTable":"block","enabled":true}},"rxGoh4ZELfiH6ljZpHcFps":{"role":"editor","value":{"id":"rxGoh4ZELfiH6ljZpHcFps","parentId":"0t4ssf3S27wr7CSM6PfHAl","type":"text","props":{"title":[["gpuswitch 0",[["c"]]],[" (forced iGPU)",null]]},"children":[],"parentTable":"block","enabled":true}},"2YcsThYR9fArdrqTw1myKr":{"role":"editor","value":{"id":"2YcsThYR9fArdrqTw1myKr","parentId":"1l48j6ucH92JKhLswB0Bqb","type":"simple_table_cell","props":{"columnKey":"qdvsChcS"},"children":["JV8vCCi1RyTDGcglHM9eP9"],"parentTable":"block","enabled":true}},"JV8vCCi1RyTDGcglHM9eP9":{"role":"editor","value":{"id":"JV8vCCi1RyTDGcglHM9eP9","parentId":"2YcsThYR9fArdrqTw1myKr","type":"text","props":{"title":[["No crashes during normal use, ",null],["but sleep/wake still triggers kernel panic",[["b"]]],[" (PCIe power state recovery fails)",null]]},"children":[],"parentTable":"block","enabled":true}},"oivfpLJqlCMwJjwfEcqATq":{"role":"editor","value":{"id":"oivfpLJqlCMwJjwfEcqATq","parentId":"5dcod8EfIwP7NsltVx0NfA","type":"simple_table_row","props":{},"children":["1gCpvZ5APmopVL6Pefg8AQ","IeKTo8W2NYJCu6uYegfPTJ"],"parentTable":"block","enabled":true}},"1gCpvZ5APmopVL6Pefg8AQ":{"role":"editor","value":{"id":"1gCpvZ5APmopVL6Pefg8AQ","parentId":"oivfpLJqlCMwJjwfEcqATq","type":"simple_table_cell","props":{"columnKey":"4vrXXszI"},"children":["hu5NhRzOOqjXz3UNgR24r9"],"parentTable":"block","enabled":true}},"hu5NhRzOOqjXz3UNgR24r9":{"role":"editor","value":{"id":"hu5NhRzOOqjXz3UNgR24r9","parentId":"1gCpvZ5APmopVL6Pefg8AQ","type":"text","props":{"title":[["gpuswitch 0",[["c"]]],[" + ",null],["disablesleep 1",[["c"]]]]},"children":[],"parentTable":"block","enabled":true}},"IeKTo8W2NYJCu6uYegfPTJ":{"role":"editor","value":{"id":"IeKTo8W2NYJCu6uYegfPTJ","parentId":"oivfpLJqlCMwJjwfEcqATq","type":"simple_table_cell","props":{"columnKey":"qdvsChcS"},"children":["XAsEENY8b30DUpMCTXCvos"],"parentTable":"block","enabled":true}},"XAsEENY8b30DUpMCTXCvos":{"role":"editor","value":{"id":"XAsEENY8b30DUpMCTXCvos","parentId":"IeKTo8W2NYJCu6uYegfPTJ","type":"text","props":{"title":[["Stable — GPU Resets still occur occasionally but self-recover without system crash",null]]},"children":[],"parentTable":"block","enabled":true}}}}},{"blockId":"7uNiazfdnu5KsTFW4xtmSr","blockSubtree":{"block":{"7uNiazfdnu5KsTFW4xtmSr":{"role":"editor","value":{"id":"7uNiazfdnu5KsTFW4xtmSr","parentId":"","type":"text","props":{"title":[["Key findings:",[["b"]]]]},"children":[],"parentTable":"block","enabled":true}}}}},{"blockId":"psSrMP8QyH4rBv9ckQOLs5","blockSubtree":{"block":{"psSrMP8QyH4rBv9ckQOLs5":{"role":"editor","value":{"id":"psSrMP8QyH4rBv9ckQOLs5","parentId":"","type":"numbered_list","props":{"title":[["Not application-specific:",[["b"]]],[" I don't use Claude Code. Crashes happen with WindowServer, mediaanalysisd, and other system processes — whatever touches the AMD GPU.",null]]},"children":[],"parentTable":"block","enabled":true}}}}},{"blockId":"tneCFPPBHX7Zt0UVkrqF7M","blockSubtree":{"block":{"tneCFPPBHX7Zt0UVkrqF7M":{"role":"editor","value":{"id":"tneCFPPBHX7Zt0UVkrqF7M","parentId":"","type":"numbered_list","props":{"title":[["Sleep/wake is a major trigger:",[["b"]]],[" Panic logs show ",null],["PCIeUp link state: 0x1614",[["c"]]],[" on every crash. The AMD GPU fails to recover its PCIe link after wake, causing the watchdog timeout. This matches ",null],["Signature: 2, Restart Channel: 18 VMPT",[["c"]]],[" in GPU Reset reports.",null]]},"children":[],"parentTable":"block","enabled":true}}}}},{"blockId":"vN7g1DXdZPBfVeXry1RNKG","blockSubtree":{"block":{"vN7g1DXdZPBfVeXry1RNKG":{"role":"editor","value":{"id":"vN7g1DXdZPBfVeXry1RNKG","parentId":"","type":"numbered_list","props":{"title":[["Started exactly with 26.5.2:",[["b"]]],[" 28 days of zero crashes on 26.5.1, then immediate failures on 26.5.2 release day. The update clearly changed AMD GPU power management / PCIe link recovery behavior.",null]]},"children":[],"parentTable":"block","enabled":true}}}}},{"blockId":"GwPSbwBKPW30FNjgfBHcP5","blockSubtree":{"block":{"GwPSbwBKPW30FNjgfBHcP5":{"role":"editor","value":{"id":"GwPSbwBKPW30FNjgfBHcP5","parentId":"","type":"numbered_list","props":{"title":[["Workaround:",[["b"]]],[" ",null],["sudo pmset -a gpuswitch 0",[["c"]]],[" + ",null],["sudo pmset -a disablesleep 1",[["c"]]],[" keeps the machine stable, but disabling sleep is not a real fix.",null]]},"children":[],"parentTable":"block","enabled":true}}}}},{"blockId":"cwgsST2CqbObYKm9YwI2sH","blockSubtree":{"block":{"cwgsST2CqbObYKm9YwI2sH":{"role":"editor","value":{"id":"cwgsST2CqbObYKm9YwI2sH","parentId":"","type":"text","props":{"title":[["Panic log excerpt:",[["b"]]]]},"children":[],"parentTable":"block","enabled":true}}}}},{"blockId":"YwaRwgA2ACpxqcSHY80xYc","blockSubtree":{"block":{"YwaRwgA2ACpxqcSHY80xYc":{"role":"editor","value":{"id":"YwaRwgA2ACpxqcSHY80xYc","parentId":"","type":"code","props":{"title":[["panic(cpu 1 caller 0x...): macOS watchdog detected\nPCIeUp link state: 0x1614\nBoot → Sleep → Wake → panic (within minutes)"]]},"children":[],"parentTable":"block","enabled":true}}}}},{"blockId":"IjVq3rjPcFU7tjlTOl9tzr","blockSubtree":{"block":{"IjVq3rjPcFU7tjlTOl9tzr":{"role":"editor","value":{"id":"IjVq3rjPcFU7tjlTOl9tzr","parentId":"","type":"text","props":{"title":[["This is a macOS 26.5.2 regression affecting all MacBookPro16,1 with AMD 5500M, independent of workload. Would appreciate Apple addressing this in 26.5.3 or 26.6.",null]]},"children":[],"parentTable":"block","enabled":true}}}}}]"></sc-clipboard>
@kikalex Thanks for the detailed
gpuswitchA/B testing — that pins down the actual cause.I need to correct my own report: the "Workflow tool 55s before panic" correlation I flagged was likely coincidental. My other panic that same day happened during plain Opus/Sonnet use, no Fable/Workflow involved — same
panic_stringboth times. So this looks like your macOS 26.5.2 AMD dGPU PCIe link regression, not Claude-Code-specific.I've applied
gpuswitch 0+disablesleep 1a few days ago (same MacBookPro16,1 / 5500M / 26.5.2). No panics/GPU Resets so far, but too early to call it resolved — will report back.Follow-up as promised: the workaround (
gpuswitch 0+disablesleep 1) held for 12 days, then panicked again this morning — samepanic_string(WindowServer watchdog timeout).Notable details from the logs:
Application:field empty, all command processors IDLE (CPGraphics= IDLE CPCompute= IDLE), 130 RSMU timeout events — so this fires even with zero workload, further confirming it's the 26.5.2 driver's power management itselfAMDRadeonX6000*frames — even undergpuswitch 0it keeps a handle to the AMD Metal driver, so a dGPU firmware death still takes it downSo
gpuswitch 0+disablesleep 1reduces frequency (0 resets in 12 days vs. daily storms before) but doesn't fully prevent it. Still on 26.5.2 — no fix available viasoftwareupdateas of today.