Hi all, i see a critical bug in kernel version 2.4.0 and 2.4.1, look it:
i use Linux Slackware 7.1
My Hard Disk:
hda: IBM-DPTA-372730, ATA DISK drive
hda: 53464320 sectors (27374 MB) w/1961KiB Cache, CHS=3328/255/63, UDMA(33)
I have instaled in this hard disk 3 OS (OpenBSD, Linux Slack and Windows 2K
(NTFS).
==============
OpenBSD --> 2GB
Linux -----> 7GB
Win2k ----> 18GB
==============
Windows2K is automount in fstab. When i run command "df -h" the values of
capacity its REALLY wrong, look it:
[root@john /]:: df -h
Filesystem Size Used Avail Use% Mounted on
/dev/hda1 5.4G 3.7G 1.5G 71% /
/dev/hda2 143G 136G 6.8G 95% /mnt/hda2
[root@john /]::
When i had kernel 2.2.18 i did not have this problem....
Sorry for my really BAD English...
Att,
Marcel Silva e Sousa
[email protected]
On Sun, Feb 11, 2001 at 05:33:35AM -0200, Marcel Silva e Sousa wrote:
> Hi all, i see a critical bug in kernel version 2.4.0 and 2.4.1, look it:
>
> My Hard Disk:
> hda: IBM-DPTA-372730, ATA DISK drive
> hda: 53464320 sectors (27374 MB) w/1961KiB Cache, CHS=3328/255/63, UDMA(33)
>
> [root@john /]:: df -h
> Filesystem Size Used Avail Use% Mounted on
> /dev/hda1 5.4G 3.7G 1.5G 71% /
> /dev/hda2 143G 136G 6.8G 95% /mnt/hda2
> [root@john /]::
>
> When i had kernel 2.2.18 i did not have this problem....
Hmm. df is not the kernel.
But df does a statfs system call, and something might be broken,
especially if you have an unusual filesystem type.
You might report on (i) what filesystem type lives on /dev/hda2?
(ii) what does "strace -e statfs df" say? What df outputs in the
"Size" column is the value of f_blocks.
Of course, if you write garbage to the size field of some superblock,
then the kernel will happily report it - there need not be a bug.
But now that you say "2.2.18 did not have this problem":
what is the 2.2.18 df output?
What is the "dmesg | grep hda" on both 2.4 and 2.2?