:I took the changes out, and I'm running a stock kernel, and it seems to
:be ok. So it's gotta be something going on. Maybe i should just use
:lockmgr locks, and forget the intr/serializer stuff?
:
:--Peter
It's up to you. lockmgr will work. The advantage of using the
serializer is that the interrupt subsystem can determine whether the
interlock will block *before* trying to call the interrupt procedure,
which can be beneficial when multiple devices share the same irq.
I'd say... get it working and stable first, then worry about making
it efficient.
-Matt
Matthew Dillon
<dillon@backplane.com>
| Torsten Kaiser | Re: 2.6.23-mm1 |
| debian developer | Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3 |
| Davide Libenzi | [patch 7/8] fdmap v2 - implement sys_socket2 |
| Luca Tettamanti | [BUG] Oops with KVM-27 |
git: | |
| Rene Herman | "git pull . <branch>" versus "git merge <branch>" |
| Wink Saville | Resolving conflicts |
| Matthieu Moy | git push to a non-bare repository |
| Tim Ansell | Whats happening with git-notes? |
| Pieter Verberne | pkg_add keeps giving errors |
| Matthew Dempsky | Re: Real men don't attack straw men |
| GVG GVG | ssh_exchange_identification: Connection closed by remote host |
| Ivo Chutkin | problem installing some packages on 4.2 |
| David Miller | TCP event tracking via netlink... |
| Michael Grollman | Re: 8169 Intermittent ifup Failure Issue With RTL8102E Chipset in Intel's New D945... |
| Rafael J. Wysocki | Re: [bug?] tg3: Failed to load firmware "tigon/tg3_tso.bin" |
| Jon Smirl | e1000 tcp checksum incorrect, x86 64b |
