home bbs files messages ]

Forums before death by AOL, social media and spammers... "We can't have nice things"

   alt.os.linux      Getting to be as bloated as Windows!      107,822 messages   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]

   Message 107,061 of 107,822   
   Paul to J.O. Aho   
   Re: Hard disk error (Error probing devic   
   07 Apr 25 15:00:29   
   
   From: nospam@needed.invalid   
      
   On Mon, 4/7/2025 10:37 AM, J.O. Aho wrote:   
   > On 07/04/2025 14.07, Carlos E.R. wrote:   
   >> On 2025-04-07 11:42, Simon wrote:   
   >   
   >>> I found a fix to this some time ago, maybe it will help still now?   
   https://blog.simonj.eu/blog/jms578-based-adaptor-on-linux   
   >>   
   >> What does this do? The link doesn't explain.   
   >>   
   >> Google says that this changes usb-disk driver "UAS" to "usb-storage". Why   
   would I need that, what is the advantage?   
   >   
   > The UAS is a newer implementation for mass storage devices, this uses a SCSI   
   based protocol instead of the Bulk-Only Transport of the USB. Main benefit   
   with UAS is higher transfer speeds from/to usb connected mass storage units.   
   >   
   > As the device you use seems to have issues with UAS (at least the Linux   
   implementation), so the option seems to be to relay on the old USB   
   functionality and the benefit for you would be something that works.   
   >   
   > The following numbers based on tests on a Banana Pi:   
   > Seq Write: ~10% slower on USB vs UAS   
   > Seq Read: ~14% slower on USB vs UAS   
   >   
   >   
   > // Aho   
      
   For this particular device, the anecdotal evidence is the   
   regular USB storage seems to work, whereas the UAS implementation   
   had rough edges. The UAS supports TRIM, which would be a better   
   fit for SSDs, but I would not put SSDs in that enclosure.   
   A hard drive may survive any brutal treatment that firmware   
   hands out. We don't really know what that firmware   
   does to drives. If a firmware isn't remotely "compliant",   
   it remains a science experiment.   
      
   The next issue, is benching it for performance, and seeing   
   what kind of write rate it can manage. None of the chit-chat   
   about it I've seen so far, mentions a bench and whether the   
   internal processor limits it. It may be examining each ATA   
   command set command, as it goes by.   
      
      Paul   
      
   --- SoupGate-DOS v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]


(c) 1994,  bbs@darkrealms.ca