2001-02-25 15:54:41

by Nick Kurshev

[permalink] [raw]
Subject: Probably patch-2.4.1 is not complete

Hello!

I have downloaded a full tarball of linux-kernel-2.4.0 and patches: patch-2.4.1 patch-2.4.2.
But patch-2.4.1 imho it not complete. During linking a linker said about unresolved reference:
__buggy_fxsr_alignment
The patch-2.4.1 contain following records:
/* Enable FXSR and company _before_ testing for FP problems. */
-#if defined(CONFIG_X86_FXSR) || defined(CONFIG_X86_RUNTIME_FXSR)
/*
* Verify that the FXSAVE/FXRSTOR data will be 16-byte aligned.
*/
- if (offsetof(struct task_struct, thread.i387.fxsave) & 15)
- panic("Kernel compiled for PII/PIII+ with FXSR, data not 16-byte aligned!");
-
+ if (offsetof(struct task_struct, thread.i387.fxsave) & 15) {
+ extern void __buggy_fxsr_alignment(void);
+ __buggy_fxsr_alignment();
+ }
if (cpu_has_fxsr) {
printk(KERN_INFO "Enabling fast FPU save and restore... ");
set_in_cr4(X86_CR4_OSFXSR);
printk("done.\n");
}
-#endif

I.e. in version 2.4.0 of kernel this code was conditionally compiled and did not cause any problems.
In 2.4.1 coditional compilation was avoided. In 2.4.2 this place is not modified.
I did an attempt to find out implementation of __buggy_fxsr_alignment but have not found any source which
contains body of function.

I have K6 cpu and for me this piece of code is unnecessary. But already second patch I'm modifing by hand.
May be it would be better for me to download a full tarball of linux-2.4.2?

Any suggestions please!

Best regards! Nick



2001-02-25 18:21:46

by Siddharth Kashyap

[permalink] [raw]
Subject: Incorrect CD Drive speed


Hi all,

I have 52X ATAPI/IDE internel cdrom dive. I am using
2.2.16 kernel. At the boot time I get the message:

hdb: CDROM DRIVE 52X

Then two lines later, i get the following message:

hdb: 48X CD-ROM DRIVE w/128kb buffer

Can someone tell me why am I getting these two
different messages?

Thanks,
Siddharth.

__________________________________________________
Do You Yahoo!?
Get email at your own domain with Yahoo! Mail.
http://personal.mail.yahoo.com/

2001-02-25 19:08:22

by Mircea Ciocan

[permalink] [raw]
Subject: Re: Incorrect CD Drive speed

Simple, our friends at No-name OEM Corporation ;) named that drive
"CDROM DRIVE 52X" in his ID string, a perfect OEM name that does not
involves any brand and looks good in windoze where the l/user see it in
system configuration as above.
The hard reality is next, meaning the Linux driver determined "by
specific means" :) that the drive is at most "48X CD-ROM DRIVE w/128kb
buffer", THAT windozians will NOT see it.
It happens a lot here in Romania also with crap chinese/taiwanese
drivers that doesn't carry any brand name and are bulk solded.
Personally, I'll trust more a drive that say "Teac Corporation 32X"
than that no-name shit. IMMV.

Mircea C.


>

Siddharth Kashyap wrote:
>
> Hi all,
>
> I have 52X ATAPI/IDE internel cdrom dive. I am using
> 2.2.16 kernel. At the boot time I get the message:
>
> hdb: CDROM DRIVE 52X
>
> Then two lines later, i get the following message:
>
> hdb: 48X CD-ROM DRIVE w/128kb buffer
>
> Can someone tell me why am I getting these two
> different messages?
>
> Thanks,
> Siddharth.

2001-02-25 20:45:06

by J.A. Magallon

[permalink] [raw]
Subject: Re: Probably patch-2.4.1 is not complete


On 02.25 Nick Kurshev wrote:
> Hello!
>
> I have downloaded a full tarball of linux-kernel-2.4.0 and patches:
> patch-2.4.1 patch-2.4.2.
> But patch-2.4.1 imho it not complete. During linking a linker said about
> unresolved reference:
> __buggy_fxsr_alignment

Are you using pgcc ?
Which compiler do you build kernel with ?

--
J.A. Magallon $> cd pub
mailto:[email protected] $> more beer

Linux werewolf 2.4.2-ac3 #1 SMP Fri Feb 23 21:48:09 CET 2001 i686

2001-02-25 21:16:40

by Jens Axboe

[permalink] [raw]
Subject: Re: Incorrect CD Drive speed

On Sun, Feb 25 2001, Mircea Ciocan wrote:
> Simple, our friends at No-name OEM Corporation ;) named that drive
> "CDROM DRIVE 52X" in his ID string, a perfect OEM name that does not
> involves any brand and looks good in windoze where the l/user see it in
> system configuration as above.
> The hard reality is next, meaning the Linux driver determined "by
> specific means" :) that the drive is at most "48X CD-ROM DRIVE w/128kb
> buffer", THAT windozians will NOT see it.

None of the two are the hard reality, the latter is just what
the drive reports in the mode sense page while (as you correctly
identified) the former is the info from the inquiry command. Why
these don't match up is a question best asked at the manufacturerer :)

--
Jens Axboe