On Mon, Apr 22, 2024 at 07:12:21PM -0400, Sasha Levin wrote: > On Mon, Apr 22, 2024 at 11:54:33AM -0700, Nathan Chancellor wrote: > > On Sun, Apr 21, 2024 at 01:11:19PM -0400, Sasha Levin wrote: > > > This is a note to let you know that I've just added the patch titled > > > > > > configs/hardening: Fix disabling UBSAN configurations > > > > > > to the 6.8-stable tree which can be found at: > > > http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary > > > > > > The filename of the patch is: > > > configs-hardening-fix-disabling-ubsan-configurations.patch > > > and it can be found in the queue-6.8 subdirectory. > > > > > > If you, or anyone else, feels it should not be added to the stable tree, > > > please let <stable@xxxxxxxxxxxxxxx> know about it. > > > > > > > > > > > > commit a54fba0bb1f52707b423c908e153d6429d08db58 > > > Author: Nathan Chancellor <nathan@xxxxxxxxxx> > > > Date: Thu Apr 11 11:11:06 2024 -0700 > > > > > > configs/hardening: Fix disabling UBSAN configurations > > > > > > [ Upstream commit e048d668f2969cf2b76e0fa21882a1b3bb323eca ] > > > > While I think backporting this makes sense, I don't know that > > backporting 918327e9b7ff ("ubsan: Remove CONFIG_UBSAN_SANITIZE_ALL") to > > resolve the conflict with 6.8 is entirely necessary (or beneficial, I > > don't know how Kees feels about it though). I've attached a version that > > applies cleanly to 6.8, in case it is desirable. > > I usually wouldn't do it, but 918327e9b7ff ("ubsan: Remove > CONFIG_UBSAN_SANITIZE_ALL") indicated that it's mostly a noop rather > than a change in behavior for existing config files. It is a change in behavior for architectures that did not select CONFIG_ARCH_HAS_UBSAN_SANITIZE_ALL because CONFIG_UBSAN_SANITIZE_ALL depends on that (for example, LoongArch). That change actually helped expose an issue in LoongArch's LLVM backend because UBSAN was now getting enabled on files in allmodconfig, which is good, but that shows the change is not risk free. However, given that it makes CONFIG_UBSAN perhaps work more expectedly (which could be considered a fix on its own), we can probably keep this as is and just back out of it if it is too disruptive to people. Cheers, Nathan
