fritzcapi does not compile anymore for kernels >= 2.6.20 because of http://lwn.net/Articles/213149/ The attached patch fixes that. It should only be applied for kernels >= 2.6.20 though. Reproducible: Always Steps to Reproduce:
Created attachment 109289 [details, diff] workqueue changes kernel 2.6.20
thanks, fixed!
patch is wrong and will break. please drop it for now. i'll post further explanation later today.
The work functions (e.g. closing_worker) currently take a void* pointer argument. On 2.6.19, this points to the user specified data element, passed in as the third argument to INIT_WORK. On 2.6.20, the argument to the work function is always the work struct, i.e. a pointer to closing_work in the above example. As the old code has this as void*, the compiler doesn't complain, but infact the code is interpreting that work_struct as if it is the previous user data pointer (capi_card) which is an entirely different area of memory. The solution is to make the work structs a member variable of the capi_card struct (the user data pointer), and then use an inverse lookup (container_of macro) inside the work function to find out which capi_card contains the pointer to that work item.
>>> Emerging (1 of 1) net-dialup/fritzcapi-2.6.43 to / * km_fritzcapi-2.6-43.x86_64.rpm MD5 ;-) ... [ ok ] * km_fritzcapi-2.6-43.x86_64.rpm RMD160 ;-) ... [ ok ] * km_fritzcapi-2.6-43.x86_64.rpm SHA1 ;-) ... [ ok ] * km_fritzcapi-2.6-43.x86_64.rpm SHA256 ;-) ... [ ok ] * km_fritzcapi-2.6-43.x86_64.rpm size ;-) ... [ ok ] * checking ebuild checksums ;-) ... [ ok ] * checking auxfile checksums ;-) ... [ ok ] * checking miscfile checksums ;-) ... [ ok ] * checking km_fritzcapi-2.6-43.x86_64.rpm ;-) ... [ ok ] * Determining the location of the kernel source code * Found kernel source directory: * /usr/src/linux * Found sources for kernel version: * 2.6.20-gentoo * * You can control the modules which are built with the variable * FRITZCAPI_CARDS which should contain a blank separated list * of a selection from the following cards: * fcpci fcpcmcia fcusb2 * * I give you the chance of hitting Ctrl-C and make the necessary * adjustments in /etc/make.conf. * Selected cards: fcpci fcusb2 >>> Unpacking source... * Converting ./Makefile to use M= instead of SUBDIRS= ... [ ok ] * Converting ./fritz.pci/Makefile to use M= instead of SUBDIRS= ... [ ok ] * Converting ./fritz.pci/src/Makefile to use M= instead of SUBDIRS= ... [ ok ] * Converting ./fritz.pcmcia/Makefile to use M= instead of SUBDIRS= ... [ ok ] * Converting ./fritz.pcmcia/src/Makefile to use M= instead of SUBDIRS= ... [ ok ] * Converting ./fritz.usb2/Makefile to use M= instead of SUBDIRS= ... [ ok ] * Converting ./fritz.usb2/src/Makefile to use M= instead of SUBDIRS= ... [ ok ] * Applying fcusb2-2.6.19.patch ... [ ok ] * Applying 2.6.43-linux-2.6.19-irq_handler.patch ... * Failed Patch: 2.6.43-linux-2.6.19-irq_handler.patch ! * ( /usr/portage/net-dialup/fritzcapi/files/2.6.43-linux-2.6.19-irq_handler.patch ) * * Include in your bugreport the contents of: * * /var/tmp/portage/net-dialup/fritzcapi-2.6.43/temp/2.6.43-linux-2.6.19-irq_handler.patch-10168.out !!! ERROR: net-dialup/fritzcapi-2.6.43 failed. Call stack: ebuild.sh, line 1614: Called dyn_unpack ebuild.sh, line 751: Called qa_call 'src_unpack' environment, line 4118: Called src_unpack fritzcapi-2.6.43.ebuild, line 114: Called epatch '/usr/portage/net-dialup/fritzcapi/files/2.6.43-linux-2.6.19-irq_handler.patch' eutils.eclass, line 341: Called die !!! Failed Patch: 2.6.43-linux-2.6.19-irq_handler.patch! !!! If you need support, post the topmost build error, and the call stack if relevant. !!! A complete build log is located at '/var/tmp/portage/net-dialup/fritzcapi-2.6.43/temp/build.log'.
Created attachment 110153 [details] There is log from patching of package, as described in output from emerge
Comment on attachment 110153 [details] There is log from patching of package, as described in output from emerge Sorry, this should be applied to bug #163366
*** This bug has been marked as a duplicate of bug 163366 ***
