I'm trying to pick a kernel for ext3.
Looking at the web page (http://www.zip.com.au/~akpm/linux/ext3/), I see
little that clues me in to the fact that "most" of ext3 is already in
(some|most|all) of the 2.4 series kernels.
Am I correct that most of the patches listed on the web page now are
actually bug fixes to the existing ext3 code, as opposed to previous
patches that actually added ext3 functionality to the stock kernels?
New arrivals to the page will want to know that, especially given all
the out-of-date documentation floating around the Net saying you have
to patch to add ext3 to the kernel.
As best I understand what I see on the web page, patches get put there
because of the time lag in getting them into the actual kernel. A patch
put on the web page for kernel X.Y is eventually rolled into kernel X.Z
(for some Z > Y) and the patch is unneccessary for anyone running kernels
X.P for all P >= Z.
What is *not* at all clear is how newcomers can determine the value of
"Z". Is the patch ext3-2.4-0.9.16-2417p2.gz against 2.4.17-pre2
already
incorported into the next kernel version 2.4.17-pre3, or did it hang
around until 2.4.17-pre4, or -pre5, or -pre999? Did it make it into
the kernel at all? How can we know?
Could the web page be updated to contain the information about in
which kernel version the listed patches have actually been applied?
Merci and thank you to all doing this excellent work.
--
-IAN! Ian! D. Allen Ottawa, Ontario, Canada idallen@ncf.ca
Home Page on the Ottawa FreeNet: http://www.ncf.ca/~aa610/
College professor at: http://www.algonquincollege.com/~alleni/
Board Member, TeleCommunities CANADA http://www.tc.ca/
"Ian! D. Allen [NCFreeNet]" wrote:> > I'm trying to pick a kernel for ext3. > > Looking at the web page (http://www.zip.com.au/~akpm/linux/ext3/), I see > little that clues me in to the fact that "most" of ext3 is already in > (some|most|all) of the 2.4 series kernels. > > Am I correct that most of the patches listed on the web page now are > actually bug fixes to the existing ext3 code, as opposed to previous > patches that actually added ext3 functionality to the stock kernels? > New arrivals to the page will want to know that, especially given all > the out-of-date documentation floating around the Net saying you have > to patch to add ext3 to the kernel. > > As best I understand what I see on the web page, patches get put there > because of the time lag in getting them into the actual kernel. A patch > put on the web page for kernel X.Y is eventually rolled into kernel X.Z > (for some Z > Y) and the patch is unneccessary for anyone running kernels > X.P for all P >= Z.Yup. Before ext3 was integrated, the patches were of course whole-filesystem additions. After Alan (and later Linus) merged the filesystem, the web page became a staging area for patches, for people to test prior to submitting the patch for mainstream inclusion. Nowadays, the lag between the posting of a patch and its inclusion at kernel.org is about one week. It's been suggested that I add "this was integrated in x.y.z" to the changelog. I need to do that... If you're looking for a kernel, I'd suggest 2.4.18-pre7. It works OK and has current ext3.
begin Ian! D. Allen [NCFreeNet] quotation:> I'm trying to pick a kernel for ext3. > > Looking at the web page (http://www.zip.com.au/~akpm/linux/ext3/), I see > little that clues me in to the fact that "most" of ext3 is already in > (some|most|all) of the 2.4 series kernels.Correct, starting at about 2.4.13 or so (I forget exactly which release).> Am I correct that most of the patches listed on the web page now are > actually bug fixes to the existing ext3 code, as opposed to previous > patches that actually added ext3 functionality to the stock kernels? > New arrivals to the page will want to know that, especially given all > the out-of-date documentation floating around the Net saying you have > to patch to add ext3 to the kernel.Well, you can certainly assume that if a patch says that it's for 2.4.x, it is newer than the ext3 code in the official 2.4.x sources, for whatever value of x. However, it may be that the only changes are for very unlikely error cases that may not even be relevant to your hardware. So far, I have seen no need to patch kernel 2.4.17 with newer ext3 code.> As best I understand what I see on the web page, patches get put there > because of the time lag in getting them into the actual kernel. A patch > put on the web page for kernel X.Y is eventually rolled into kernel X.Z > (for some Z > Y) and the patch is unneccessary for anyone running kernels > X.P for all P >= Z.This is what I believe also.> What is *not* at all clear is how newcomers can determine the value of > "Z". Is the patch ext3-2.4-0.9.16-2417p2.gz against 2.4.17-pre2 already > incorported into the next kernel version 2.4.17-pre3, or did it hang > around until 2.4.17-pre4, or -pre5, or -pre999? Did it make it into > the kernel at all? How can we know?You could consult the kernel change logs (which are online at kernel.org). My guess is that if the ext3 web site has a patch to be applied to, say, 2.4.18pre2, and the kernel change log for some subsequent 2.4.18pre says that it includes "ext3 updates", then the pre2 patch is probably no longer needed. To be doubly certain, you could compare the ext3-2.4.18pre2 patch to the kernel-2.4.18pre3 patch (or whatever version contains "ext3 updates") to see if it appears to be included.> Could the web page be updated to contain the information about in > which kernel version the listed patches have actually been applied?That's obviously a little extra work, but if the ext3 developers can spare the time for it, it would be nice. Craig