login
Search
"It's time to sanitize prototypes of bdev ->open(), ->release() and ->ioctl()," Al Viro began in an RFC posted to the Linux Kernel mailing list, "this stuff had sat in 'need to fix' for a long time and there is a bunch of bugs hard to fix without dealing with it." Following a detailed explanation of how he intends to meet this goal, he added, "[the] resulting APIs will be a lot saner and [the] entire thing is reasonably easy to split into bisect-friendly chunks. It is an API breaker, of course, but we don't have a lot of affected modules and anything not converted will be (a) immediately caught by gcc and (b) trivial to convert."
Linux creator Linus Torvalds responded favorably to the proposal, "from your description, I have no objections - everything sounds good. My only concern is how painful the patch ends up being (and a worry about whether this will affect a metric truck-load of external modules? That said, I can't really see us worrying about those)". Al noted that he would begin with a series of preparatory patches, "I have the beginning of that series done and the rest mapped out in enough detail to implement it over this week. If somebody has objections, questions or comments - yell."
Continuing the debate over the right way to go about implementing some of the features found in the newly released Reiser4 [story], Hans Reiser asked Al Viro to clarify the issues he thinks could arise from the current implementation. The result was a brilliant explanation of what problems Al sees, specifically related to dentry aliasing, and how the current VFS architecture handles some of these problems.
Read on for Al's response and further clarification from Linus Torvalds. The interesting exchange provides some good insight into the Linux VFS layer.
Linus Torvalds released the 2.6.7-rc3 release candidate kernel which
contains a few cleanups. He notes:
"Ok, let's calm down for a while before the final 2.6.7.
-rc3 does a lot of sparse type cleanup, mainly thanks to Al Viro (but his
work ended up getting some other people involved too, since the list of
sparse warnings isn't as daunting any more). Some of that has unearthed
real bugs which Al fixed.But there are DRM, AGP, cpufreq, sparc64, and input updates there too."
Read on for the full changelog.
Linux creator Linus Torvalds has released the linux 2.6.0-test5 kernel, with the following comments:
"Lots of small stuff, as usual. I think the biggest "core" change is the Futex changes by Jamie and Hugh, and the dev_t preparations by Al Viro. But there are ARM and ppc updates here too, and a few drivers have bigger fixes (tg3 driver and the USB gadget interface stand out on diffstat). Watchdog driver updates etc. And Russell King fixed more PCMCIA issues."
Read on for the full changelog.
Additionally, if you followed my recent upgrade howto [story], are running a 2.6.0-test kernel, and are interested in upgrading to 2.6.0-test5, read on for a few simple tips on upgrading with incremental patches.
