Skip to contentSkip to conversation

BUG: isofs broken (2.2 and 2.4)

DISCUSSION Harald Koenig 1 participant 1 message Nov 17, 2000 1 message not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

#1 lore raw
Quoted 6 lines · On Nov 17, [email protected] wrote:
Quoted 5 lines

Review

Quoted 3 lines
(quoted diff) +2 -0
+               if (cpnt)
+                       kfree(cpnt);

this seems to make things much worse

Quoted 8 lines

Yes, I meant

		if (cpnt) {
			kfree(cpnt);
			cpnt = NULL;
		}

at that place, otherwise things will be freed multiple times.

_MUCH_ better now!!! no lockups anymore, no memory leak(s).

BUT: there is still this small performace and memory usage issue:

each of these CDs contains >80k files in ~110 directories each (the full db consists of 18 CDs!) and running "find" or "du" on one CDROM (or 4MB isofs loop mount from hard disk) takes a huge amount of time (real and system cpu) with cold cache:

	time find /cdrom
	0.610u 97.250s 1:40.58 97.2%    0+0k 0+0io 98pf+0w

flush cache...

	time du /cdrom
	0.470u 97.220s 1:40.29 97.4%    0+0k 0+0io 112pf+0w

whereas with hot cache (takes ~45MB memory off the value
for "free + buffer/cache" for a 4MB isofs image!) gives (PPro200, 128MB):

	time find /cdrom
	0.460u 1.280s 0:01.79 97.2%     0+0k 0+0io 102pf+0w

	time du /cdrom
	0.270u 1.260s 0:01.54 99.3%     0+0k 0+0io 108pf+0w

so it seems to work (data still not checked, can do this only next week) but performace really sucks :(

anyway, thanks a lot for your help and quick patch ! now at least we can copy all the data to some hard disk and use it that way.

a patch for 2.2.x (the real production machine can't run 2.4.x yet) and/or fixes for the bad performace would be appreciated anyway ;^)

thanks!

Harald

Signature · 12 lines
All SCSI disks will from now on                     ___       _____
be required to send an email notice                0--,|    /OOOOOOO\
24 hours prior to complete hardware failure!      <_/  /  /OOOOOOOOOOO\
                                                    \  \/OOOOOOOOOOOOOOO\
                                                      \ OOOOOOOOOOOOOOOOO|//
Harald Koenig,                                         \/\/\/\/\/\/\/\/\/
Inst.f.Theoret.Astrophysik                              //  /     \\  \
[email protected]                     ^^^^^       ^^^^^
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
Please read the FAQ at http://www.tux.org/lkml/