Prev: How to send mouse inputs to a specific window via implementation of driver?
Next: about NDIS_WDM
From: sven.kreamer on 3 Apr 2008 05:24
I have a SCSI attached removable disk drive. In the XP device manager
Policies tab, the options are grayed out and "Optimize for
performance" is selected. Is there a way to default the drive to
"Optimize for quick removal" instead?
From: Ben Voigt [C++ MVP] on 3 Apr 2008 10:07
sven.kreamer(a)gmail.com wrote:
> I have a SCSI attached removable disk drive. In the XP device manager
> Policies tab, the options are grayed out and "Optimize for
> performance" is selected. Is there a way to default the drive to
> "Optimize for quick removal" instead?

Perhaps you have a caching SCSI controller, in that case disabling the
Windows write cache wouldn't actually help with surprise removal.


From: sven.kreamer on 4 Apr 2008 00:02
> Perhaps you have a caching SCSI controller, in that case disabling the
> Windows write cache wouldn't actually help with surprise removal.

The controller is not caching type. Even if it were, that is no
problem at all because the miniport would set CachesData=TRUE and
Windows would send SRB_FUNCTION_FLUSH whenever it wishes to
synchronously commit all data from the adapter cache to the media.

The problem with "Optimize for performance" is some stuff never gets
written to the disk until shutdown. Users cannot be expected to go
through the safely remove hardware wizard so when the user pulls the
disk out it is in a bad state. Isn't that what "Optimize for quick
removal" is for? Removable disk drives on USB get "Optimize for quick
removal" automatically. I need to find a way to achieve this same
default for SCSI. I don't know where to begin. Is it a change needed
in the miniport itself? An INF file is needed for the disk? A filter
driver?
From: Ben Voigt [C++ MVP] on 7 Apr 2008 09:39
sven.kreamer(a)gmail.com wrote:
>> Perhaps you have a caching SCSI controller, in that case disabling
>> the Windows write cache wouldn't actually help with surprise removal.
>
> The controller is not caching type. Even if it were, that is no
> problem at all because the miniport would set CachesData=TRUE and
> Windows would send SRB_FUNCTION_FLUSH whenever it wishes to
> synchronously commit all data from the adapter cache to the media.
>
> The problem with "Optimize for performance" is some stuff never gets
> written to the disk until shutdown. Users cannot be expected to go
> through the safely remove hardware wizard so when the user pulls the
> disk out it is in a bad state. Isn't that what "Optimize for quick
> removal" is for? Removable disk drives on USB get "Optimize for quick
> removal" automatically. I need to find a way to achieve this same
> default for SCSI. I don't know where to begin. Is it a change needed
> in the miniport itself? An INF file is needed for the disk? A filter
> driver?

Well, surprise removal on USB drives doesn't work either, in XP you can get
data corruption really easily. Vista includes a fix (and also automatically
runs chkdsk on any drive that was surprise removed in XP).


From: Pavel A. on 7 Apr 2008 15:30
Recently an usb disk corruption on Vista sp1 after successful safe removal
has been reported in .vista.hardware_devices.
So it seems that the story will be continued.

--PA


"Ben Voigt [C++ MVP]" <rbv(a)nospam.nospam> wrote in message
news:uf8MsRLmIHA.4504(a)TK2MSFTNGP06.phx.gbl...
> sven.kreamer(a)gmail.com wrote:
>>> Perhaps you have a caching SCSI controller, in that case disabling
>>> the Windows write cache wouldn't actually help with surprise removal.
>>
>> The controller is not caching type. Even if it were, that is no
>> problem at all because the miniport would set CachesData=TRUE and
>> Windows would send SRB_FUNCTION_FLUSH whenever it wishes to
>> synchronously commit all data from the adapter cache to the media.
>>
>> The problem with "Optimize for performance" is some stuff never gets
>> written to the disk until shutdown. Users cannot be expected to go
>> through the safely remove hardware wizard so when the user pulls the
>> disk out it is in a bad state. Isn't that what "Optimize for quick
>> removal" is for? Removable disk drives on USB get "Optimize for quick
>> removal" automatically. I need to find a way to achieve this same
>> default for SCSI. I don't know where to begin. Is it a change needed
>> in the miniport itself? An INF file is needed for the disk? A filter
>> driver?
>
> Well, surprise removal on USB drives doesn't work either, in XP you can
> get data corruption really easily. Vista includes a fix (and also
> automatically runs chkdsk on any drive that was surprise removed in XP).

 |  Next  |  Last
Pages: 1 2
Prev: How to send mouse inputs to a specific window via implementation of driver?
Next: about NDIS_WDM