How is memory allocated between the different TCP queues ?

[Teto]

Hi,

I know there are different TCP queues (prequeue/backlog/out of order)
depending on the state of the application. I was wondering how the
values set in net.ipv4.tcp_rmem were "ventilated" (allocated) between
the different queues ? For instance can the out of order queue
represent 100% of tcp_rmem max ?

Regards

MAtt