Tuesday, August 19th 2025
Phison Responds to Windows 11 24H2 Update Crashing SSDs
Yesterday we reported on Windows 11 24H2 security update as part of the August 2025 Patch Tuesday cycle to counter the Lumma stealer and similar malware and exploits. The cumulative package KB5063878 included in the update caused SSDs to fail, possibly corrupting data. Community testing has since identified a reproducible pattern in which certain SSDs, when running sustained sequential writes of approximately 50 GB, become inaccessible and display unreadable SMART telemetry. Additionally, under 60% or higher controller load, most of the SSDs failed. This indicates that Microsoft's patch and some controller firmware may be disagreeing in the logic of operations. However, Phison, as one of the makers of the impacted controllers, has decided to be vocal about it and informed Tom's Hardware that the company is actively looking into the issue.
Source:
Tom's Hardware
Phison for Tom's HardwarePhison has recently been made aware of the industry-wide effects of the 'KB5063878' and 'KB5062660' updates on Windows 11 that potentially impacted several storage devices, including some supported by Phison. We understand the disruption this may have caused and promptly engaged industry stakeholders. We are steadfast in our commitment to product integrity and the success of our partners and end users. At this time, the controllers that may have been affected are under review and we are working with partners. We will continue to provide updates and advisories to partners who may have been impacted to provide support and ensure any applicable remediation.

97 Comments on Phison Responds to Windows 11 24H2 Update Crashing SSDs
in the meantime, what are people to do without an avenue to replace bricks?
and oh, even disabling updates on a hardware firewall level doesn't block a lot of updates sadly. I dual boot on 3 laptops w Linux and w11 and I'm afraid to go into w11 now. Watch MS throw Phison under the bus like they always blame the hardware manufacturers when it's their friggin screwups!
Not just Phison there's a megathread on reddit where samsung users have this going on too and AFAIK samsung uses their own controllers and there's a few innogrit based drives also failing due to this.
WHAT A FRICKIN MESS!!!
Hope this lawsuit grows legs and kicks them in the teeth: www.techradar.com/computing/windows/microsoft-is-being-sued-for-killing-windows-10-here-are-5-reasons-why-the-lawsuit-could-actually-force-its-hand-to-extend-support-further?utm_source=chatgpt.com
Apologies for going off-topic - is MS keeping Extended Security Updates (ESU) strictly for business/enterprise or can regular users still get in? With certainty, not long ago they opened it up for personal users too. Thing is, I’ve seen the consumer ESU option appear and vanish from MS's own support page a couple of times now. Have the silly beans finally decided who’s actually allowed to subscribe or are they still faffing about?
Windows 7 is, as the kids say, the GOAT.
Samsung too? 😰
I’ll admit when I’m wrong and this time, I’m definitely wrong. This is going to be a mess—a bigger mess than the Crowdstrike mess.
Unfortunately Phison is probably going to get the blame for Microsoft's mess, but Samsung SSD's are failing too? Does it affect all Samsung SSD's or just the DRAM-less ones?
Real guestion is why there isn't class action lawsuit ripping MS new one for all the untested/known to be buggy updates forced on users.
I mean if you lose control of your car and crash into someones garden/house etc, or do it on purpose, you're going to pay for damages/the usual punishments for breaking law!
Now MS and Nadelle have been knowingly ***** up people's PCs for years with no repercussions.After that UI started going more backwards than forward with that Kool-Aid drinker of Ballmer.
And Nadelle is plain Kool-Aid factory.
We need Gates come out of retirement.
I assume MS is testing their updates with AI now given how many of them are causing issues.
This is really going to blow up on Microsoft’s face and if you ask me, it needs to. Maybe (I hope) that this will be the wake-up call Microsoft needs to finally fix their shit.
There have been a few Samsung and WDC in-house controller SSDs mentioned as well.
There is a lot of ai involved in the original and the translation but here:
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
Landmine Pattern 1: DRAM-less NVMe relying on HMB
In 24H2, the HMB allocation jumped from 64 MB to 200 MB → DMA fault → Controller went silent → OS detected "Surprise Removal" and SMART could no longer be read.
2. Old Phison FW
The E12 generation is prone to timeouts when rewriting the flash map, causing Storport to drop the entire bus → NG Lv.1.
3. SATA write cache outbursts. WD Blue SA510 2TB failed step 3. The drive experienced an NCQ timeout and went offline while the OS was forcing a cache flush.
In short, it's a trinity of "controller + presence/absence of DRAM + firmware generation."
Samsung Hynix is fine because they fixed the issue with their own firmware, while the cheaper DRAM-less ones are left behind by the Windows spec changes -- it's a blunt picture. (・ω・`)
Solution: Disable HMB or set it to 64 MB. HKLM\SYSTEM\CurrentControlSet\Services\stornvme\Parameters HostMemoryBufferDisable=1 or HostMemoryBufferMax=40000
Many reports of BSOD and disk evaporation stopped on SN770/SN580
WD has improved this in the 2024/10 version and later, and the Phison E12/E16 series also has reduced timeouts in the 2025/05 version.
Currently, "low-cost" controllers such as Phison E19/E21, SanDisk G2, SM2269XT, and older Phison E12/E16/E18 series (older with DRAM)
There was an opinion that this was a high risk.
Also, it turns out that Nekoru is in favor of calling Samsung Electronics "Kanson" (cold village) (・ω・`)
Well, rather than saying that the SSD manufacturer is to blame, it seems like this mysterious explosion is the result of Windows messing around with the cache specifications.
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
Does someone here knows if it makes any sense?
Can someone explain the 40000 setting?
Wouldn’t that make for a 40 MB buffer where it should be 64 MBs?
It is hard to believe it is an accident with the MS"s resources.