Date: Sun, 24 Feb 2008 17:44:43 +0100 From: Guido Demmenie <g.p.demmenie@student.tudelft.nl> To: FreeBSD Users Questions <freebsd-questions@freebsd.org> Subject: Harddisks DMA errors Message-ID: <DD2E0C8E-3A7C-40C8-9A34-E6535632D6AB@student.tudelft.nl>
next in thread | raw e-mail | index | archive | help
Hello, After having problems with DMA errors on my old PATA 200GB disks because of a faulty harddisk, I switched to SATA 500GB disks. I used fdisk to create slices and newfs them, then copied everything using dump & restore to my new harddisks. Now I have problems running into the next error for both my harddiscs: ---------------- ad4: WARNING - READ_DMA UDMA ICRC error (retrying request) LBA=268435454 ad4: WARNING - SETFEATURES SET TRANSER MODE taskqueue timeout - completing request directly ad4: WARNING - SETFEATURES SET TRANSER MODE taskqueue timeout - completing request directly ad4: WARNING - SETFEATURES SET RCHACHE MODE taskqueue timeout - completing request directly ad4: WARNING - SETFEATURES SET WCHACHE MODE taskqueue timeout - completing request directly ad4: WARNING - SET_MULTI taskqueue timeout - completing request directly ad4: TIMEOUT - READ_DMA retrying (0 retries left) LBA=268435454 ad4: FAILURE = READ_DMA status=ff<BUSY,READY,DMA_READY,DSC,DRQ,CORRECTABLE,INDEX,ERROR> error = ff < ICRC ,UNCORRECTABLE ,MEDIA_CHANGED ,NID_NOT_FOUND,MEDIA_CHANGE_REQUEST,ABORTED,NO_MEDIA,ILLEGAL_LENGTH> LBA=268435454 ---------------- I am able to reproduce this error at will if I issue the following command: # dd if=/dev/ad4 of=/dev/null bs=1024 What surprises me is that it the following information is provided by dd: ---------------- dd: /dev/ad4: Input/output error 134217727+0 records in 134217727+0 records out 137438952448 bytes transferred in 9732.012957 secs (14122356 bytes/sec) ---------------- This number of _bytes transfered_ looks an awful lot like 137 GB which normal LBA can handle. And it probably isn't coincidence that "normal" LBA can only handle 268435455 locations. Large LBA should work better I suppose. Problem is that I have used fdisk to newfs and slice my disk and they were fully accessible then. My Promise SATA300 TX2Plus does have support for LargeLBA so there again it should work. I'm quite sure the disks aren't faulty because I get the problem on both disks and the errors are exactly the same. I'm not sure what the SETFEATURE error has to do with this all, maybe someone can explain. So it ends up in two questions: 1. How do I get those drives working? 2. If it is of relevance what does that SETFEATURES warning mean? Anyone that can help me? -- Guido Please reply to this address, because the address that was subscribed doesn't work now because if this server failure.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?DD2E0C8E-3A7C-40C8-9A34-E6535632D6AB>