From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Stephens, Mike" Date: Wed, 19 Jul 2000 16:44:08 +0000 Subject: RE: [Linux-ia64] kernel modules Message-Id: List-Id: References: In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-ia64@vger.kernel.org -----Original Message----- > From: Christian Groessler [mailto:cpg@aladdin.de] > Sent: Wednesday, July 19, 2000 6:45 AM > To: linux-ia64@linuxia64.org > Subject: [Linux-ia64] kernel modules > > Hi, > > I'm asking, whether loadable modules support is working and if not, > what's still missing. Yes modules can be made to work. However there are still a work in progress, i.e. there are still bugs that need to be fix. > > I tried the current RH6.2 beta, and compiled the parport code > as module. (which compiled after I copied include/asm-i386/parport.h > to include/asm-ia64). > > But it crashed at load time. The crash seems to be happening at > a function call itself, as a printk exactly before the call is > displayed and a printk right at the beginning of the called > function isn't. Seems like the module hasn't been linked > correctly. Unless you make changes to the Makefile they don't get linked correctly. The current IA64 version modutils can't load a module if it (or the kernel) was compiled with the -mconstant-gp flag, which is part of the CFLAGS define in arch/ia64/Makefile. Using a 2.4.0-test1 + 0609 patch (or newer kernel) and compile without the constant-gp flag about 95% of the modules I have tried work. > > I upgraded the kernel then to 2.4.0-test4, but no change. > > Then I upgraded modutils to 2.3.12, and now when I do lsmod > it aborts with "get_kernel_syms: Function not implemented". This is caused by a problems in the configure and makefile that cause the modutils to get link wrong for IA64. The quick fix is add USE_SYSCALL=n to the top of util/Makefile. I hope to have a patch to fix this and a couple of other bugs out some time this week. > > As far as I can tell, browsing thru the kernel code, there is > a handler for this function and also in the sys_call_table. > But I couldn't find the place where this error is generated. > > Any insights would be greatly appreciated. > > regards, > chris regards Mike _______________________ mike.stephens@intel.com