From: Mark Lord <liml@rtr.ca>
To: Justin Piszcz <jpiszcz@lucidpixels.com>
Cc: Jody Belka <lists-lkml@pimb.org>,
linux-ide@vger.kernel.org, Jeff Garzik <jgarzik@pobox.com>
Subject: Re: marvell sata status?
Date: Sat, 03 Jan 2009 09:46:59 -0500 [thread overview]
Message-ID: <495F7A63.9010509@rtr.ca> (raw)
In-Reply-To: <495F79B4.2050309@rtr.ca>
Mark Lord wrote:
> Justin Piszcz wrote:
>>
> ..
>> One would think if the controller overwrites bits on the device when
>> mdadm assembles the raid device it would kick it out of the array for
>> being unclean or find a lot of mismatches on the array or is that the
>> wrong idea?
> ..
>
> No, there's nothing in mdadm that I know of which will detect this.
> You'd basically need to run a job in the background, reading/comparing
> all of the mirrors with each other.
>
> But even that mightn't work, since the BIOS would have corrupted the
> exact same sector on every drive attached to it, quite possibly with
> the same contents, so the mirror-compare wouldn't see anything different.
>
> The only way to know for sure is to test it manually,
> by zeroing the entire drive, power/reboot, scan the entire
> drive for non-zero. Then repeat with a different pattern
> just in case "zero" is what the BIOS wrote.
>
> Very very messy, at least for the Highpoint RAID cards.
> But we really don't know about what the others do,
> apart from some historical data in the dm code.
..
Oh, and from the existing sata_mv.c driver code:
if (pdev->vendor == PCI_VENDOR_ID_TTI &&
(pdev->device == 0x2300 || pdev->device == 0x2310))
{
/*
* Highpoint RocketRAID PCIe 23xx series cards:
*
* Unconfigured drives are treated as "Legacy"
* by the BIOS, and it overwrites sector 8 with
* a "Lgcy" metadata block prior to Linux boot.
*
* Configured drives (RAID or JBOD) leave sector 8
* alone, but instead overwrite a high numbered
* sector for the RAID metadata. This sector can
* be determined exactly, by truncating the physical
* drive capacity to a nice even GB value.
*
* RAID metadata is at: (dev->n_sectors & ~0xfffff)
*
* Warn the user, lest they think we're just buggy.
*/
printk(KERN_WARNING DRV_NAME ": Highpoint RocketRAID"
" BIOS CORRUPTS DATA on all attached drives,"
" regardless of if/how they are configured."
" BEWARE!\n");
printk(KERN_WARNING DRV_NAME ": For data safety, do not"
" use sectors 8-9 on \"Legacy\" drives,"
" and avoid the final two gigabytes on"
" all RocketRAID BIOS initialized drives.\n");
}
next prev parent reply other threads:[~2009-01-03 14:45 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-01-01 5:10 marvell sata status? Jody Belka
2009-01-01 9:54 ` Justin Piszcz
2009-01-01 14:07 ` Mark Lord
2009-01-01 18:44 ` Jody Belka
2009-01-01 19:11 ` Justin Piszcz
2009-01-01 19:43 ` Jody Belka
2009-01-02 16:09 ` Mark Lord
2009-01-02 16:10 ` Mark Lord
2009-01-02 18:03 ` Justin Piszcz
2009-01-02 20:18 ` Mark Lord
2009-01-02 20:18 ` Justin Piszcz
2009-01-03 5:04 ` Mark Lord
2009-01-03 9:13 ` Justin Piszcz
2009-01-03 14:44 ` Mark Lord
2009-01-03 14:46 ` Mark Lord [this message]
2009-01-21 1:33 ` Jody Belka
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=495F7A63.9010509@rtr.ca \
--to=liml@rtr.ca \
--cc=jgarzik@pobox.com \
--cc=jpiszcz@lucidpixels.com \
--cc=linux-ide@vger.kernel.org \
--cc=lists-lkml@pimb.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).