[Mageia-discuss] partitioning bug

Thu Jul 19 03:46:46 CEST 2012

Pascal Terjan wrote:
+
+> On Fri, Jul 6, 2012 at 8:37 AM, blind Pete
+> <0123peter at gmail.com> wrote:
+>> David W. Hodgins wrote:
+>>
+>>> On Thu, 05 Jul 2012 05:10:30 -0400, blind Pete
+>>> <0123peter at gmail.com> wrote:
+>>>
+>>>> Cfdisk is very unhappy about something.
+>>>>
+>>>>     FATAL ERROR: Bad primary partition 3: Partition begins after
+>>>>     end-of-disk Press any key to exit cfdisk
+>>>>
+>>>> It looks like cfdisk's error message is wrong.  Partition 3 is mounted
+>>>> and in use.
+>>>
+>>>> Disk /dev/sda: 121601 cylinders, 255 heads, 63 sectors/track
+>>>> Warning: extended partition does not start at a cylinder boundary.
+>>>> DOS and Linux will interpret the contents differently.
+>>>> Units = sectors of 512 bytes, counting from 0
+>>> 121601x255x63=1953520065
+>>
+>> CHS has not been reliable since zones were introduced decades ago.
+>> They are just a general hint that it is a big drive.  LBA or
+>> sector number is all that matters.  What happens inside the HDD
+>> case is an unknowable mystery.
+> 
+> But cfdisk is still based on CHS (or was few years ago).
+
+Ancient history:  
+
+Before the introduction of the ATA hard drive standard, in about 1990, 
+the normal way to address a sector was with a CHS tuple.  With ATA 
+zone formatting became standard practice.  There are fewer sectors in 
+a cylinder near the spindle than in a cylinder near the rim.  The 
+Integrated Drive Electronics then "translated" the hidden real, but 
+variable, geometry into a pretend geometry with a constant number of 
+sectors per cylinder, so that existing operating systems like DOS and 
+possibly BIOSes would not fail miserably by being told that 1,1,20 
+does not exist, but 20,1,30 is fine.
+
+Aligning things with physical cylinders was sometimes a good idea, but 
+now that information is hidden inside the drive.  Aligning things with 
+pretend cylinders is pointless, because they probably don't correspond 
+to physical cylinders.  
+
+> For example https://bugzilla.redhat.com/show_bug.cgi?id=630340#c1 says
+> that 2.18 (the one in mageia 1) expected the size of the device to be
+> a multiple of the cylinder size.
+
+There was a bug in cfdisk two years ago.  It seems to have been fixed.  
+Mageia 2's cfdisk does not complain about my hard drive.  
+
+>> Even if partition four were past the end of the disk it would not
+>> excuse cfdisk for complaining about partition three.
+> 
+> Yes I don't know why it complains about 3 and not 4
+
+The old version does some invalid maths - and gets the wrong answer.  
+
+-- 
+blind Pete
+Sig goes here...  
+
+