Re: Inheritance

From: cristalink (cristalink_at_nospam.nospam)
Date: 02/16/05 

Date: Thu, 17 Feb 2005 11:46:14 +1300

I remember getting some prize from OSR bugbash campaign a few years ago.
Apart from that, I made a few attempts to post bug reports into this
newsgroup. Now I have a feeling that MS is not really interested in bug
reports for whatever reasons. I understand they don't want to hear end user
complaints. But they could possible provide an email for bug reports from
professional developers. I would be happy to submit a report if I received
an automatic reply saying MS is looking into it, even if they trashed the
report straight away. The problem is that to my knowledge there is no such
an email address, for kernel bugs in particular. There's a site for
reporting bugs in VS 2005 beta. There's no site for reporting bugs in VS
2003. I understand that nobody will fix 2003. All updates will come with a
new version several years later. The same is probably true for all the rest.

I've got a few fresh dumps of crashes in Win2k3 atapi.sys code called from
ISR. Should I post them here? Do you think somebody will actually be looking
into this, especially because they have SP1 RC?

IEEE-1394 stack is extremely buggy. It should be just rewritten. I don't
believe it will be. Moreover, MS seems to gradually remove support for 1394.

CLASSPNP's handling of media change notifications and events is a mess.
There's no reliable way of getting the MCN status of a device, unless
polling it yourself with IOCTL_STORAGE_CHECK_VERIFY. I am talking about a
lots of options in different places which affect MCN. There's no way to
consistently receive eject notifications when a user pressed a button on a
drive. If someone is sending GET EVENT STATUSes to a device then this screws
up the notifications. I've just recalled a problem with MS Media Player 9
that didn't want to update the content of an audio CD when I had AutoRun=0.
With a new CD in the drive WMP displayed the content of the old CD. When I
clicked on any title, WMP began playing with the LBA remembered for the old
disk. As a result, WMP played new tracks from some random positions in the
middle. I haven't found a workaround except of restarting WMP.

The most recent problem that made me angry is in CDROM.SYS.
IOCTL_DISK_GET_DRIVE_LAYOUT and IOCTL_DISK_GET_PARTITION_INFO return the
CDROM capacity different from IOCTL_GET_DISK_GEOMETRY due to the following
nice piece of code in src\storage\class\cdrom\ioctl.c:

    //
    // we've defaulted to 32/64 forever. don't want to change this now...
    //
    fdoExtension->DiskGeometry.TracksPerCylinder = 0x40;
    fdoExtension->DiskGeometry.SectorsPerTrack = 0x20;
    fdoExtension->DiskGeometry.Cylinders.QuadPart = (LONGLONG)((lastSector +
1) / (32 * 64));
    commonExtension->PartitionLength.QuadPart =
        (commonExtension->PartitionLength.QuadPart <<
fdoExtension->SectorShift);

They rounded the number of sectors to (32*64) for DISK_GEOMETRY, while
returning the correct capacity for DRIVE_LAYOUT. This effectively prevents
certain filesystems on a CD/DVD from being mounted.

The bottom line. I need fixes from MS now. The best I can expect is to have
them in a few months, if I am lucky enough. So I count on myself only. For
my last projects, I spent more time finding workarounds for bugs in MS code
rather than fixing my ones. Why should I bother trying to find an
appropriate contact at Microsoft if I won't receive a reply anyway, or the
reply I expect from them - "there's a small chance it will be fixed in some
future versions". 

-- 
http://www.firestreamer.com - NTBackup to DVD and DV
"Don Burn" <burn@stopspam.acm.org> wrote in message 
news:a_OQd.12587$9J.2626@fe06.lga...
> Well if you see a bug are you reporting it?  Microsoft has been extremely 
> good about fixing problems with the samples, which is why you should 
> always be using the latest DDK.  Reporting it can be as simple as posting 
> here or on ntdev.
>
>
> -- 
> Don Burn (MVP, Windows DDK)
> Windows 2k/XP/2k3 Filesystem and Driver Consulting
> Remove StopSpam from the email to reply
>
>
>
>
>
> "cristalink" <cristalink@nospam.nospam> wrote in message 
> news:OHtJV2GFFHA.3492@TK2MSFTNGP12.phx.gbl...
>>
>> "Maxim S. Shatskih" <maxim@storagecraft.com> wrote in message 
>> news:%23Tb5ksGFFHA.1408@TK2MSFTNGP10.phx.gbl...
>>>> which is actually nothing more than heaps of rubbish. Frankly, I have 
>>>> not
>>>> seen any, let alone C++, good code at all, except of my one.
>>>
>>> I'm also not free from such attitude :-), though MS's OS-level sources 
>>> from DDK
>>> samples are rated high by me :-)
>>
>> Really? Well, I was hesitant about including a remark about DDK samples 
>> in my previous post... Here it goes. They look a bit messy and buggy 
>> IMHO. No wonder Windows kernel programming is so hard.
>>
>>> But the question is another. Dirty C++/OO code is by far worse then 
>>> dirty C :-)
>>
>> Agree.
>>
>> -- 
>> http://www.firestreamer.com - NTBackup to DVD and DV
>>
>>
>>>
>>> -- 
>>> Maxim Shatskih, Windows DDK MVP
>>> StorageCraft Corporation
>>> maxim@storagecraft.com
>>> http://www.storagecraft.com
>>>
>>>
>>
>>
>>
>
>
Quantcast