Messages in this thread | | | Date | Wed, 27 Jan 1999 09:24:17 +0000 (GMT) | From | Bill Melotti <> | Subject | Re: RAMDISK Bug - 2.0.36 kernel (again) |
| |
John
Thanks for the reply. The only one from an obviously popular and busy list.
I guessed this might be the case when it allowed me to allocate one huge block of 24Meg, hence I tried mallocing small blocks and also using calloc to see what would happen. That's when I crashed the machine.
This does not happen when the machine is running from a hard disk or entirely from an NFS server (including root) I am sure this is a RAMDISK related bug.
Do you know anything about the RAMDISK internals ?
Bill Melotti Cognito Ltd UK
On Tue, 26 Jan 1999, John Alvord wrote:
> One aspect is that malloc() allocated storage does not consume (much) RAM > until the program actually uses it. john alvord > > > Music, Management, Poetry and more... > http://www.candlelist.org/kuilema > > Cheap CDs @ http://www.cruzio.com/~billpeet/MusicByCandlelight > > > On Tue, 26 Jan 1999, Bill Melotti wrote: > > > > > > > Hi > > > > Can anyone offer any suggestions on a possible RAMDISK bug I have seen. > > > > I wish to boot a system from a flash IDE drive into a 16Meg RAMDISK on a > > system with 32 Meg installed. I have modified drivers/block/rd.c and > > commented out the line that checks for a floppy as the root dev, this > > allows me to at least load a compressed filesystem image from an HD into > > RAMDISK. > > > > The system I have uses about 11 meg of the 16. You would imagine I have a > > little less than 16Meg available for application use. Actually I have > > a few processes running from boot and free reports only about 6Meg > > left. > > > > I tried running some programs that would allocate memory until failure to > > confirm what was really available (because I though 6 was a little low) > > > > I wrote a program that would malloc single large blocks. With this I can > > malloc 23 Meg before failure. (malloc returns NULL) > > > > Thinking there was a problem, I changed this to use calloc instead. If I > > tried to calloc more than 10 Meg, the machine locked up. Between 6 and 10 it > > would usually keep going for a while, but lock up later. I tried running > > these programs as nobody rather than root. Same result. If I were running any > > other programs (eg 'top') in another session, this would usually bomb with a > > bus error at the same time. The callocing app might also bomb with a bus > > error before the machine locked. I am doing this stuff over a telnet session, > > I haven't tried it on a console. > > > > It is as if the RAMDISK code as not marked the buffers it uses as > > unavailable for app code. I looked through the code and see it sets the > > BH_Protected flag in the file buffer structs. However I also see that > > there are 2 or 3 places (fs/buffers.c and somewhere in ext2 hierarchy) > > where these can be cleared. > > > > > > Can anyone offer any help/comments/suggestions ? > > > > Many thanks > > > > > > Bill Melotti > > Cognito Ltd UK > > > > > > - > > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in > > the body of a message to majordomo@vger.rutgers.edu > > Please read the FAQ at http://www.tux.org/lkml/ > > > > > > - > > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in > > the body of a message to majordomo@vger.rutgers.edu > > Please read the FAQ at http://www.tux.org/lkml/ > > >
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |