Re: aesni(?) corrupts data on 8.2-BETA1

2010-12-12 Thread Kostik Belousov
On Sat, Dec 11, 2010 at 07:37:51PM -0500, Mike Tancsa wrote:
 On 12/11/2010 6:22 PM, Kostik Belousov wrote:
  On Sat, Dec 11, 2010 at 06:08:08PM -0500, Mike Tancsa wrote:
  On 12/11/2010 11:01 AM, Kostik Belousov wrote:
 
  I have no access to AESNI hardware. For start, you may use
  src/tools/tools/crypto/cryptotest
  to somewhat verify the sanity of the driver.
 
  I doesnt happen every time, but one out of 5 or so 
 
  First, which arch is it, amd64 or i386 ?
  
  Also, please revert r216162 and do the same tests.
 
 Hi,
   Its AMD64, but i386 seems to be impacted too. I am not sure how to
 revert to a specific commit, but for now I csup'd with a date tag of
 
 *date=2010.12.02.23.00.00
 
 which is a day before
 http://lists.freebsd.org/pipermail/svn-src-stable-8/2010-December/004338.html
 
 
 And that seems to fix it!
 
 I  have been running
 cryptotest -c -z -t 10
 in a loop for the past 10min and not one error.

Please try this patch on the latest HEAD or RELENG_8.

diff --git a/sys/amd64/amd64/fpu.c b/sys/amd64/amd64/fpu.c
index 482b5da..1b493b4 100644
--- a/sys/amd64/amd64/fpu.c
+++ b/sys/amd64/amd64/fpu.c
@@ -426,7 +426,9 @@ fpudna(void)
fxrstor(fpu_initialstate);
if (pcb-pcb_initial_fpucw != __INITIAL_FPUCW__)
fldcw(pcb-pcb_initial_fpucw);
-   fpuuserinited(curthread);
+   pcb-pcb_flags |= PCB_FPUINITDONE;
+   if (PCB_USER_FPU(pcb))
+   pcb-pcb_flags |= PCB_USERFPUINITDONE;
} else
fxrstor(pcb-pcb_save);
critical_exit();
diff --git a/sys/i386/isa/npx.c b/sys/i386/isa/npx.c
index 9ec5d25..f314e44 100644
--- a/sys/i386/isa/npx.c
+++ b/sys/i386/isa/npx.c
@@ -684,7 +684,9 @@ npxdna(void)
fpurstor(npx_initialstate);
if (pcb-pcb_initial_npxcw != __INITIAL_NPXCW__)
fldcw(pcb-pcb_initial_npxcw);
-   npxuserinited(curthread);
+   pcb-pcb_flags |= PCB_NPXINITDONE;
+   if (PCB_USER_FPU(pcb))
+   pcb-pcb_flags |= PCB_NPXUSERINITDONE;
} else {
/*
 * The following fpurstor() may cause an IRQ13 when the


pgpA0gqcjE6TG.pgp
Description: PGP signature


Re: aesni(?) corrupts data on 8.2-BETA1

2010-12-12 Thread Mike Tancsa
On 12/12/2010 3:43 AM, Kostik Belousov wrote:
 On Sat, Dec 11, 2010 at 07:37:51PM -0500, Mike Tancsa wrote:
 On 12/11/2010 6:22 PM, Kostik Belousov wrote:
 *date=2010.12.02.23.00.00

 which is a day before
 http://lists.freebsd.org/pipermail/svn-src-stable-8/2010-December/004338.html


 And that seems to fix it!

 I  have been running
 cryptotest -c -z -t 10
 in a loop for the past 10min and not one error.
 
 Please try this patch on the latest HEAD or RELENG_8.
 

I tried both on i386 and amd64 and all looks good!  I did 1000
iterations of
cryptotest -c -z -t 10

without issue!  Thanks for the quick fix!

---Mike

0(backup3)# patch -p1  p
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--
|diff --git a/sys/amd64/amd64/fpu.c b/sys/amd64/amd64/fpu.c
|index 482b5da..1b493b4 100644
|--- a/sys/amd64/amd64/fpu.c
|+++ b/sys/amd64/amd64/fpu.c
--
Patching file sys/amd64/amd64/fpu.c using Plan A...
Hunk #1 succeeded at 426.
Hmm...  The next patch looks like a unified diff to me...
The text leading up to this was:
--
|diff --git a/sys/i386/isa/npx.c b/sys/i386/isa/npx.c
|index 9ec5d25..f314e44 100644
|--- a/sys/i386/isa/npx.c
|+++ b/sys/i386/isa/npx.c
--
Patching file sys/i386/isa/npx.c using Plan A...
Hunk #1 succeeded at 684.
done
0(backup3)#
___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org


8.2-BETA1 sysinstall: No USB devices found

2010-12-12 Thread Kim Culhan
Intalling from:
ftp://ftp.freebsd.org/pub/FreeBSD/ISO-IMAGES-amd64/8.2/FreeBSD-8.2-BETA1-amd64-memstick.img

From sysinstall choose installation media: 9 USB   Install from a USB drive

Returns:

No USB devices found!

Several usb slots on the motherboard were tried with the same result.

The motherboard is: Intel D945GCLF2 with the latest bios dated 04142010

Please let me know if there is any way I can help.

-kim
___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org


Re: 8.2-BETA1 sysinstall: No USB devices found

2010-12-12 Thread Bruce Cran
On Sun, 12 Dec 2010 09:38:59 -0500
Kim Culhan w8hd...@gmail.com wrote:

 No USB devices found!

Can you try re-scanning for devices (via the Options screen) -
sysinstall can fail to detect USB disks the first time around.

-- 
Bruce Cran
___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org


Re: 8.2-BETA1 sysinstall: No USB devices found

2010-12-12 Thread Kim Culhan
On Sun, Dec 12, 2010 at 10:17 AM, Bruce Cran br...@cran.org.uk wrote:

 On Sun, 12 Dec 2010 09:38:59 -0500
 Kim Culhan w8hd...@gmail.com wrote:

  No USB devices found!

 Can you try re-scanning for devices (via the Options screen) -
 sysinstall can fail to detect USB disks the first time around.

Thanks Bruce worked fine.

Hitting space on Re-scan devices didn't appear to have any effect but
hitting a space on
Media Type allowed choosing USB.

thanks
-kim
___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org


Re: 8.2-BETA1 sysinstall: No USB devices found

2010-12-12 Thread Brandon Gooch
On Sun, Dec 12, 2010 at 9:17 AM, Bruce Cran br...@cran.org.uk wrote:
 On Sun, 12 Dec 2010 09:38:59 -0500
 Kim Culhan w8hd...@gmail.com wrote:

 No USB devices found!

 Can you try re-scanning for devices (via the Options screen) -
 sysinstall can fail to detect USB disks the first time around.

 --
 Bruce Cran

How difficult would it be for the installer to automatically re-scan
for devices when USB is selected as install or Fix-it media?

-Brandon
___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org


Re: 8.2-BETA1 sysinstall: No USB devices found

2010-12-12 Thread Michael Voorhis

On 12/12/2010 11:28 AM, Brandon Gooch wrote:

How difficult would it be for the installer to automatically re-scan
for devices when USB is selected as install or Fix-it media?


Ditto this.  I do all my OS installs via a USB CD reader and it took me 
a little while to come up with the rescan-fix.


I don't order rack servers with optical disk readers that I'll only be 
using once to install the OS, so all the stuff I build is installed from 
USB disc readers.  My initial response was how can we not know there's 
a CD reader when I just booted and read the installer from it.


One of the 1st things sysinstall does is run a device scan 
deviceGetAll(), but that misses the USB device, apparently.  The 
deviceRescan() calls the same function, but preceeds it with deviceReset().


All this stuff is in src/usr.sbin/sysinstall/devices.c.  I'm not smart 
enough to make my own install disks right now; what would happen if the 
reset were inserted in front of the rescan (in main.c) at the beginning 
of sysinstall execution?  That seems to be the only difference between 
the initial scan and the option scan.





___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org


Re: 8.2-BETA1 sysinstall: No USB devices found

2010-12-12 Thread Ian Smith
On Sun, 12 Dec 2010, Michael Voorhis wrote:
  On 12/12/2010 11:28 AM, Brandon Gooch wrote:
   How difficult would it be for the installer to automatically re-scan
   for devices when USB is selected as install or Fix-it media?
  
  Ditto this.  I do all my OS installs via a USB CD reader and it took me a
  little while to come up with the rescan-fix.

This one goes back a while; I got involved in a discusion in questions@ 
where Randi provided a fair bit of info on some of the issues here:
 Re: 8.0-RELEASE-i386-memstick fixit - No USB devices found!
 http://lists.freebsd.org/pipermail/freebsd-questions/2009-December/208977.html

  I don't order rack servers with optical disk readers that I'll only be using
  once to install the OS, so all the stuff I build is installed from USB disc
  readers.  My initial response was how can we not know there's a CD reader
  when I just booted and read the installer from it.

Ah, perhaps it's not just older kit and/or some types of USB stick that 
have this issue then ..

  One of the 1st things sysinstall does is run a device scan deviceGetAll(),
  but that misses the USB device, apparently.  The deviceRescan() calls the
  same function, but preceeds it with deviceReset().
  
  All this stuff is in src/usr.sbin/sysinstall/devices.c.  I'm not smart enough
  to make my own install disks right now; what would happen if the reset were
  inserted in front of the rescan (in main.c) at the beginning of sysinstall
  execution?  That seems to be the only difference between the initial scan and
  the option scan.

There may be an issue with doing that too: it seems that reset also 
resets other things, in my case losing the disk slicing, labelling and 
choice of distributions that had already been done:
 Re: Installing 8.1-RELEASE from the memstick
 http://lists.freebsd.org/pipermail/freebsd-questions/2010-August/219690.html

I still don't undersand why memstick images need to be in DD mode (daXa) 
rather than allowing being properly sliced (daXsY), unless it's just the 
clash with sliced da devices indicating 'real' SCSI disks in devices.c?

I'd love to be able to put amd64 and i386 installers, plus say the 
packages from the DVD, plus somewhere to save dmesgs, on one stick.

cheers, Ian
___
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org


Re: 8.2-BETA1 sysinstall: No USB devices found

2010-12-12 Thread Jeremy Chadwick
On Mon, Dec 13, 2010 at 01:53:41PM +1100, Ian Smith wrote:
 On Sun, 12 Dec 2010, Michael Voorhis wrote:
   On 12/12/2010 11:28 AM, Brandon Gooch wrote:
How difficult would it be for the installer to automatically re-scan
for devices when USB is selected as install or Fix-it media?
   
   Ditto this.  I do all my OS installs via a USB CD reader and it took me a
   little while to come up with the rescan-fix.
 
 This one goes back a while; I got involved in a discusion in questions@ 
 where Randi provided a fair bit of info on some of the issues here:
  Re: 8.0-RELEASE-i386-memstick fixit - No USB devices found!
  
 http://lists.freebsd.org/pipermail/freebsd-questions/2009-December/208977.html
 
   I don't order rack servers with optical disk readers that I'll only be 
 using
   once to install the OS, so all the stuff I build is installed from USB disc
   readers.  My initial response was how can we not know there's a CD reader
   when I just booted and read the installer from it.
 
 Ah, perhaps it's not just older kit and/or some types of USB stick that 
 have this issue then ..
 
   One of the 1st things sysinstall does is run a device scan deviceGetAll(),
   but that misses the USB device, apparently.  The deviceRescan() calls the
   same function, but preceeds it with deviceReset().
   
   All this stuff is in src/usr.sbin/sysinstall/devices.c.  I'm not smart 
 enough
   to make my own install disks right now; what would happen if the reset were
   inserted in front of the rescan (in main.c) at the beginning of sysinstall
   execution?  That seems to be the only difference between the initial scan 
 and
   the option scan.
 
 There may be an issue with doing that too: it seems that reset also 
 resets other things, in my case losing the disk slicing, labelling and 
 choice of distributions that had already been done:
  Re: Installing 8.1-RELEASE from the memstick
  http://lists.freebsd.org/pipermail/freebsd-questions/2010-August/219690.html
 
 I still don't undersand why memstick images need to be in DD mode (daXa) 
 rather than allowing being properly sliced (daXsY), unless it's just the 
 clash with sliced da devices indicating 'real' SCSI disks in devices.c?
 
 I'd love to be able to put amd64 and i386 installers, plus say the 
 packages from the DVD, plus somewhere to save dmesgs, on one stick.

I've been trying to do this for quite some time -- I dedicated nearly 4
weeks to trying to accomplish something similar ~6 months ago.

My idea was to have multiple .iso files on a single USB memstick, which
GRUB/MEMLINUX would therefore boot.  Things like Kubuntu, FreeBSD 7.x
i386 and amd64, 8.x i386 and amd64, DBAN, memtest86+, Western Digital,
Seagate, Maxtor, Hitachi, etc. disk analysis/repair tools, and a couple
of other ISOs.

To clarify: literally I wanted a single FAT32 filesystem with .iso files
placed on it, with some GRUB/MEMLINUX boot semantics/config files put in
place, to provide a menuing system with information of what ISO file
represented what.  You'd select the OS/thing you wanted to boot (see
below website for a screenshot) and off it went.  Why .iso files on
FAT32 rather than a dedicated slice for each thing?  Because it makes
testing a new release or upgrading an ISO very simple; insert USB stick
in any OS, delete old file, copy over new one, edit the GRUB
configuration bits, sync and/or umount, and pull the USB stick.

The tool I was using was this (though it has been updated regularly.
Note that the latest version is 2 days old; I was working with a version
many months ago):

http://www.pendrivelinux.com/boot-multiple-iso-from-usb-multiboot-usb/

The result: I was able to get things like memtest86+ and DBAN to boot,
and very specific Linux distributions (with a lot of pain), in addition
to FreeBSD.  However, FreeBSD would always fail to find the necessary
installation packages/sources/etc. due to how the whole booting
process above works.  I had to spend a lot of time messing around with
MEMLINUX and the map --hook rootnoverify (0xff) parameters and similar
whatnots:

http://syslinux.zytor.com/wiki/index.php/MEMDISK

Another problem I ran into is described on that site under the Boot
Errors section, blabbing about how the drive has to be defragmented and
the files have to be in contiguous order (I assume this means every
block of data in the ISO has to be in linear order and cannot be spread
across non-linear sectors/blocks of the USB drive).

What I'm trying to say is that basically accomplishing things how I
described above is impossible given the limitations of the PC BIOS
(I have to assume INT 0x13), and that's really too bad given the need
for such a thing in this day and age.

What did I end up doing?  Buying multiple HP-brand 4GB USB memory sticks
and dedicating each of them to whatever tool was needed.  When visiting
the datacenter I haul with me a large tote of tools, cables, and of
course said sticks.

HTH.

-- 
| Jeremy Chadwick   

Re: 8.2-BETA1 sysinstall: No USB devices found

2010-12-12 Thread Daniel O'Connor

On 13/12/2010, at 14:06, Jeremy Chadwick wrote:
 The result: I was able to get things like memtest86+ and DBAN to boot,
 and very specific Linux distributions (with a lot of pain), in addition
 to FreeBSD.  However, FreeBSD would always fail to find the necessary
 installation packages/sources/etc. due to how the whole booting
 process above works.  I had to spend a lot of time messing around with
 MEMLINUX and the map --hook rootnoverify (0xff) parameters and similar
 whatnots:

For just FreeBSD I have used syslinux to boot an MFS and then have sysinstall 
read off the USB stick as a hard disk.

I submitted a few patches and they're committed now so you can probably do it 
out of the box.

The only pain is generating the MFS but that isn't _too_ tricky.

It would be nice if it was generated during the normal 'make release' though 
(hint hint)

Also, I tried a minimal MFS and then loading the kernel directly from FAT32 
using the loader but I could never get it to see the FAT32 even though by my 
reading of the code it should work..

--
Daniel O'Connor software and network engineer
for Genesis Software - http://www.gsoft.com.au
The nice thing about standards is that there
are so many of them to choose from.
  -- Andrew Tanenbaum
GPG Fingerprint - 5596 B766 97C0 0E94 4347 295E E593 DC20 7B3F CE8C