EDD still failing on some systems with 2.6.23-rc6-git2

EDD still failing on some systems with 2.6.23-rc6-git2

[Chuck Ebbert]

I added debugging code to print a letter for each step in i386 setup,
and it gets to 'J', then it hangs:

        /* Query EDD information */
#if defined(CONFIG_EDD) || defined(CONFIG_EDD_MODULE)
        putchar('J');
        query_edd();
#endif

Multiple reports of this happening on 2.6.23-rc6 can be found at
the end of:

https://bugzilla.redhat.com/show_bug.cgi?id=251724
https://bugzilla.redhat.com/show_bug.cgi?id=239585

Re: EDD still failing on some systems with 2.6.23-rc6-git2

[H. Peter Anvin]

Chuck Ebbert wrote:
> I added debugging code to print a letter for each step in i386 setup,
> and it gets to 'J', then it hangs:
> 
>         /* Query EDD information */
> #if defined(CONFIG_EDD) || defined(CONFIG_EDD_MODULE)
>         putchar('J');
>         query_edd();
> #endif
> 
> Multiple reports of this happening on 2.6.23-rc6 can be found at
> the end of:
> 
> https://bugzilla.redhat.com/show_bug.cgi?id=251724
> https://bugzilla.redhat.com/show_bug.cgi?id=239585

I just attached a test patch to the bug report shown above.  If someone
with access to the relevant hardware could test out the patch, I would
greatly appreciate it.

	-hpa

Re: EDD still failing on some systems with 2.6.23-rc6-git2

[Chuck Ebbert]

On 09/14/2007 12:06 PM, H. Peter Anvin wrote:
> Chuck Ebbert wrote:
>> I added debugging code to print a letter for each step in i386 setup,
>> and it gets to 'J', then it hangs:
>>
>>         /* Query EDD information */
>> #if defined(CONFIG_EDD) || defined(CONFIG_EDD_MODULE)
>>         putchar('J');
>>         query_edd();
>> #endif
>>
>> Multiple reports of this happening on 2.6.23-rc6 can be found at
>> the end of:
>>
>> https://bugzilla.redhat.com/show_bug.cgi?id=251724
>> https://bugzilla.redhat.com/show_bug.cgi?id=239585
> 
> I just attached a test patch to the bug report shown above.  If someone
> with access to the relevant hardware could test out the patch, I would
> greatly appreciate it.
> 

Still fails. And apparently fails on disk 0, because it hangs right
after printing a zero:

--- linux-2.6.22.noarch.orig/arch/i386/boot/edd.c
+++ linux-2.6.22.noarch/arch/i386/boot/edd.c
@@ -151,6 +151,7 @@ void query_edd(void)
                 * Scan the BIOS-supported hard disks and query EDD
                 * information...
                 */
+               putchar('0' + (devno - 0x80) % 10);
                get_edd_info(devno, &ei);

                if (boot_params.eddbuf_entries < EDDMAXNR) {

Re: EDD still failing on some systems with 2.6.23-rc6-git2

[H. Peter Anvin]

Chuck Ebbert wrote:
> 
> Still fails. And apparently fails on disk 0, because it hangs right
> after printing a zero:
> 
> --- linux-2.6.22.noarch.orig/arch/i386/boot/edd.c
> +++ linux-2.6.22.noarch/arch/i386/boot/edd.c
> @@ -151,6 +151,7 @@ void query_edd(void)
>                  * Scan the BIOS-supported hard disks and query EDD
>                  * information...
>                  */
> +               putchar('0' + (devno - 0x80) % 10);
>                 get_edd_info(devno, &ei);
> 
>                 if (boot_params.eddbuf_entries < EDDMAXNR) {

You know... we actually do have printf()... you don't have to play this
kind of games; you can just printf("Drive %x...\n", devno);

So does this happen ONLY when booting from CD-ROM, or is there any other
case when it fails, too?

	-hpa