On 11/01/2008, Shreyansh <shreyansh.jain@xxxxxxxxx> wrote: > Greg Freemyer <greg.freemyer <at> gmail.com> writes: > ----8<---- > > If you want to have a high quality copy of the messages on disk, then > > make your own copy. ie. "dmesg > my-dmesg-copy" > > > > The negative with dmesg is that the kernel's buffer is a fifo. If for > > some reason it is too small for the amount of logging you need, then > > you can adjust the size and make as big as you need. > ---8<---- > > Hi Greg and list, > note: usually you should leave people you reply to on Cc: ( or To: ) . > Thanks, I will try with that as well. I understand the limitation of dmesg being > circluar log buffer, but I think that if I re-direct data into my own copy > (file) I should be able to tackle that problem. > > Nevertheless, I still would like to know if actually someone has observed log > data corruption in view of excessive printks (or it might be happening on my > machine due to some other yet-undetected problem). > Well, if you have the kernels log buffer set to a very small value and you generate a lot of log messages very fast, then sure the buffer can get "corrupted" by new messages overwriting old ones before they can be read. The easy solution, as Greg also pointed out, is to increase the size of the kernels log buffer so it is large enough to accomodate the amount of data you throw at it. It's simple, just adjust the CONFIG_LOG_BUF_SHIFT option in menuconfig (General Setup --> Kernel log buffer size) - just set it really high, like 18 or so, or higher - sure, you'll waste some memory, but that's usually not so important when debugging :) . -- Jesper Juhl <jesper.juhl@xxxxxxxxx> Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html Plain text mails only, please http://www.expita.com/nomime.html -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx Please read the FAQ at http://kernelnewbies.org/FAQ
