With ac-sources-2.4.21_rc1-r1 installed and the system is booted with a kernel built from those sources (`uname -r` = 2.4.21-rc1-ac1), and the /usr/src/linux symlink set to the correct kernel source directory, an emerge of alsa-driver-0.9.2 fails, but I coded up a patch for it. Reproducible: Always Steps to Reproduce: 1. emerge ac-sources (2.4.21_rc1-r1) 2. set the /usr/src/linux symlink 3. configure, build and install kernel 4. reboot 5. emerge alsa-driver (0.9.2) Actual Results: The emerge of alsa-driver fails while compiling 'serialmidi.c' line 159(-ish). Expected Results: It should have emerged correctly. The compilation of 'serialmidi.c' fails as 'tty->count' no longer being an 'int' in the kernel sources; the "ac" patch changes it to 'atomic_t'. I have a patch which treats 'tty->count' in serialmidi.c as an 'atomic_t' regardless of whether it should be an 'int' or an 'atomic_t'. That should be safe (I've checked all the asm-XXX/atomic.h files in the kernel sources for binary compatibility between 'int' and 'atomic_t'), but is a bit naughty, and is intended as a temporary workaround. (The patch should be attached this bug report shortly after submission.)
Created attachment 11132 [details, diff] Proposed fix
Personally, I don't feel like applying the proposed patch, because: 1. Only one version of kernel souces is affected, and it's an early RC one. 2. The patch is a bit too crude. Most people don't need it. If you could enclose the changes into the precompiler directives, limiting their scope to the specific kernel version, then I'd accept it. I'm closing this as WONTFIX. If you disagree with the above points, please feel free to reopen this bug. p.s. Is it still an issue with 0.9.3a?
In response to Arcady's comments: 1. It is still an issue with ac-sources-2.4.21_rc1-r4 and alsa-driver-0.9.3a (though I haven't built either yet). 2. I agree the patch is crude, but there are no sufficiently localised changes made by the "ac" patch that can be tested at the preprocessor stage.
I just wanted to add that I have also run into this problem. alsa-driver-0.9.2 linux-2.4.22-ac4 As you can imagine, this is a real pain to deal with since I need to re-emerge alsa-drivers after every kernel build.