TODO

From Linux PARISC Wiki
(Difference between revisions)
Jump to: navigation, search
m (→‎Review all the todo entries below!)
m (→‎PA-RISC Linux TODO (actually this is Helge's TODO list))
Line 1: Line 1:
 
= PA-RISC Linux TODO (actually this is Helge's TODO list) =
 
= PA-RISC Linux TODO (actually this is Helge's TODO list) =
  
−
== Status & TODO 2016-02-14 ==
+
== Status & TODO 2016-03-17 ==
−
* fix up __ro_after_init: http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=d09e356ad06a8b6f5cceabf7c6cf05fdb62b46e5
+
* add wiki page about kernel cross-compiling
  
−
* ALPHA: Add https://lkml.org/lkml/2007/8/14/259 aboot patch
+
* kernel work
−
* ALPHA: Kernel patches: http://marc.info/?t=140610652200001&r=1&w=2, The proposed fix is in: http://marc.info/?t=140675670100010&r=1&w=2, http://marc.info/?t=142327510200001&r=1&w=2,  
+
** STARTED: fix up __ro_after_init: http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=d09e356ad06a8b6f5cceabf7c6cf05fdb62b46e5
−
* DONE: kernel timerfd_settime/timerfd_gettime issue, test with debian package:  liblinux-fd-perl -> https://patchwork.kernel.org/patch/7951661/, backported to all kernels
+
** STARTED: Fix exception table entries to be of type int (32bit), even on 64bit kernel -> saves space. - https://patchwork.kernel.org/patch/3033801/ - partly done.
−
* Implement kernel VDSO support
+
** further implement seccomp syscal, e.g. more modes than just: SECCOMP_MODE_STRICT
−
* kernel ptrace bug: http://thread.gmane.org/gmane.linux.ports.parisc/26273, possible fix: https://patchwork.kernel.org/patch/8063301/
+
** ALPHA processor port:
−
* DONE: implement HUGE pages
+
*** Add https://lkml.org/lkml/2007/8/14/259 aboot patch
−
* implement transparent huge pages (TODO?)
+
*** Kernel patches: http://marc.info/?t=140610652200001&r=1&w=2, The proposed fix is in: http://marc.info/?t=140675670100010&r=1&w=2, http://marc.info/?t=142327510200001&r=1&w=2,  
−
* Add KASLR (Kernel Address Layout Randomization) - like arm64
+
** Implement kernel VDSO support
−
* Debian packages:
+
** Add KASLR (Kernel Address Layout Randomization, ARCH_HAS_ELF_RANDOMIZE, arch_mmap_rnd() and such) - like arm64
−
# elfutils -- testsuite error
+
** convert irq handling to use CONFIG_IRQ_DOMAIN_HIERARCHY (serial port problem, action handler msising when registering irq)
−
# glib2.0 -- testsuite error
+
** Due to very bad performance running multiple copies of dot program (Dave):
−
# gtk+3.0 succeeds by adding these two lines to debian/rules:
+
*** Another issue that I have been wondering about is the performance of our kernel spinlock code.  PA 8800 and PA 8900 processors have a cache line length of 128 bytes.  I believe thereare some places where we assume 64 bytes.  We also allocate spinlocks on 16 byte boundaries. We might get better performance if we allocated on 32 (PA 1.X) and 128 (PA 2.0).  Helge: change: __PA_LDCW_ALIGNMENT
−
## touch testsuite/reftests/color-transition.ui.known_fail
+
*** I've also seen that better performance may be obtained by doing a byte store to spinlock prior to ldcw.  This dirties the cache line and allows operation in cache in certain situations. I think this mainly applies to PA 1.X but I'm not sure.  I have some example code somewhere. Helge:  change struct arch_spinlock_t; in spinlock_types.h
−
## touch testsuite/reftests/linear-gradient-transition-to-other.ui.known_fail
+
−
# libgadu - testsuite fails probably due to missing pthread cancellation in glibc.
+
−
# glibc pthread cancellation: the main issue is the target specific code needed to handle stubs and signal frames.  See linux-unwind.h and hpux-unwind.h in gcc's libgcc/config/pa directory.  The function pa32_fallback_frame_state does the hard work for linux.  Hopefully, the signal tramp hasn't changed. All the unwind tests in gcc pass, so I think the implementation there is okay. It may be we need something similar to pa32_fallback_frame_state in glibc to fix cancellation
+
−
# heimdal suceeds by increasing timeout for iprop test to 120 seconds.
+
−
# cupt
+
−
# libqb - testsuite error
+
−
# libsoup2.4 -- testsuite error
+
−
# DONE: vlc - linker does not drop dependencies to unneeded shared libs, https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=770741, problem arises due to __gmon_start__ symbol which then wrongly gets resolved e.g. by other shared libs. Fix for glibc is reported: https://sourceware.org/bugzilla/show_bug.cgi?id=19170
+
−
# DONE: nss - dladdr from glibc returns wrong names on hppa, http://article.gmane.org/gmane.linux.ports.parisc/26290, Solution: https://sourceware.org/bugzilla/show_bug.cgi?id=19415
+
−
# DONE: capnproto - testsuite was wrong: PARISC uses mips-floatingpoint NANs - https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=781787
+
−
 
+
−
* convert irq handling to use CONFIG_IRQ_DOMAIN_HIERARCHY (serial port problem, action handler msising when registering irq)
+
−
* Due to very bad performance running multiple copies of dot program (Dave):
+
−
** Another issue that I have been wondering about is the performance of our kernel spinlock code.  PA 8800 and PA 8900 processors have a cache line length of 128 bytes.  I believe thereare some places where we assume 64 bytes.  We also allocate spinlocks on 16 byte boundaries. We might get better performance if we allocated on 32 (PA 1.X) and 128 (PA 2.0).  Helge: change: __PA_LDCW_ALIGNMENT
+
−
** I've also seen that better performance may be obtained by doing a byte store to spinlock prior to ldcw.  This dirties the cache line and allows operation in cache in certain situations. I think this mainly applies to PA 1.X but I'm not sure.  I have some example code somewhere. Helge:  change struct arch_spinlock_t; in spinlock_types.h
+
−
* MOSTLY DONE: push gcc patch for variable CAS (gcc atomic ops, __sync_XXX) upstream (Kernel 3.17, http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=89206491201cbd1571009b36292af781cef74c1b) - packages needing rebuild: libatomic-ops (gauche and libgc have copies of libatomic-ops included), x265, ceph, lizardfs, firebird2.5, gauche, idba, mapsembler2, padevchooser, filezilla, idba, http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=785654
+
−
** __sync_val_compare_and_swap_8 missing in: openvdb, idba, lizardfs, ...
+
−
** gcc-4.9 commit: https://gcc.gnu.org/viewcvs/gcc?view=revision&revision=227914
+
−
** gcc-5.0 commit: https://gcc.gnu.org/viewcvs/gcc?view=revision&revision=227913
+
−
 
+
−
* kernel
+
 
** add asm/io.h file, needed for some debian packages, e.g. crrcsim
 
** add asm/io.h file, needed for some debian packages, e.g. crrcsim
−
** implement ARCH_HAS_ELF_RANDOMIZE (arch_mmap_rnd() and such)
 
 
** use CONFIG_COMPAT_BINFMT_ELF and get rid of arch/parisc/kernel/binfmt_elf32.c
 
** use CONFIG_COMPAT_BINFMT_ELF and get rid of arch/parisc/kernel/binfmt_elf32.c
 
** implement kexec syscall
 
** implement kexec syscall
 
** Implement ltrace on hppa
 
** Implement ltrace on hppa
−
** further implement seccomp syscal, e.g. more modes than just: SECCOMP_MODE_STRICT
 
 
** SLAB breaks, SLUB works, why? Mail from Guy Martin.
 
** SLAB breaks, SLUB works, why? Mail from Guy Martin.
 
** sparsemem: http://lwn.net/Articles/439472/ -> convert to sparsemem (http://lwn.net/Articles/440056/), patch available from James
 
** sparsemem: http://lwn.net/Articles/439472/ -> convert to sparsemem (http://lwn.net/Articles/440056/), patch available from James
Line 53: Line 30:
 
** better implementation to get stack pointer (in return_address(), in stack checker in irq.c, ...)
 
** better implementation to get stack pointer (in return_address(), in stack checker in irq.c, ...)
 
** physically broken memory in Dave's machine - utilize PDT (http://ftp.parisc-linux.org/docs/arch/pdc20-v1.1-Ch4-pdc.pdf (page #46)) - https://patchwork.kernel.org/patch/2541561/
 
** physically broken memory in Dave's machine - utilize PDT (http://ftp.parisc-linux.org/docs/arch/pdc20-v1.1-Ch4-pdc.pdf (page #46)) - https://patchwork.kernel.org/patch/2541561/
−
 
−
* userspace
 
−
** add/port debian openmpi package (done: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=776730)
 
−
** teach "blkid" about palo partition
 
−
** STARTED (by Dave): fix haskell (ghc) hangs: futex(0xea00d08e, FUTEX_WAIT_PRIVATE, 2, NULL) = -1 EINVAL (Invalid argument). Solution: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=756260#15
 
−
** STARTED: fix linker -fpie support, https://sourceware.org/bugzilla/show_bug.cgi?id=18421
 
−
*** fixes debian package iceweasel and others, see: http://www.openbsd.org/papers/nycbsdcon08-pie/
 
−
*** TODO: Fix kernel to load -fpie compiled programs not at 0x400000
 
−
** STARTED: why is GNU AS so slow, e.g. when building freeorion, quantlib-swig or yade packages, testcases: openturns
 
−
*** https://sourceware.org/bugzilla/show_bug.cgi?id=18427
 
−
*** http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=787192
 
−
** glibc: add backtrace() function, problems when building elfutils package: http://buildd.debian-ports.org/status/package.php?p=elfutils&suite=sid
 
−
** gcc: Add _builtin_thread_pointer() and _builtin_set_thread_pointer(), e.g.: https://gcc.gnu.org/ml/gcc-patches/2012-07/msg00428.html  for glibc: ports/sysdeps/hppa/nptl/tls.h __set_cr27()
 
−
** provide libphobos packages - gcc-defaults - http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=763103
 
−
** why does acl2 build hangs process ?
 
−
** gdb crashes Kernel - http://permalink.gmane.org/gmane.linux.ports.parisc/5406
 
−
** debian-cd to use xorriso
 
−
** debian-cd: do not compress kernel.
 
−
 
−
* add wiki page about kernel cross-compiling
 
  
 
* Kernel: Hard to solve outstanding issues
 
* Kernel: Hard to solve outstanding issues
Line 79: Line 36:
  
 
* Kernel: still missing Linux kernel development (no patches available yet):
 
* Kernel: still missing Linux kernel development (no patches available yet):
 +
** Fix kernel to load -fpie compiled programs not at 0x400000
 +
** maybe implement transparent huge pages (due to cache coherency technically only possible on PA8500-PA8700 CPUs, not possible on PA8800-PA8900 CPUs)
 
** Fix FTRACE: https://patchwork.kernel.org/patch/3554391/
 
** Fix FTRACE: https://patchwork.kernel.org/patch/3554391/
 
** optimize flush_dcache_page() implementation like sparc. See Documentation/cachetlb.txt, e.g. used in aio subsystem
 
** optimize flush_dcache_page() implementation like sparc. See Documentation/cachetlb.txt, e.g. used in aio subsystem
Line 88: Line 47:
 
** use pdc_add_valid() in kern_addr_valid(),  protect against crashable acesses (like for superio-pc-drivers-detection in lp), add pdc_add_valid() to create a probe_kernel_read() implementation.
 
** use pdc_add_valid() in kern_addr_valid(),  protect against crashable acesses (like for superio-pc-drivers-detection in lp), add pdc_add_valid() to create a probe_kernel_read() implementation.
 
** arch/parisc/kernel/sys_parisc.c -> use SYSCALL_ALIAS() -> maybe
 
** arch/parisc/kernel/sys_parisc.c -> use SYSCALL_ALIAS() -> maybe
−
** on C8000, report 1 socket with 2 cores.
+
** implement libnuma kernel support to report on C8000 one socket with 2 cores.
 
** Review gettimeofday and clock_gettime() implementation, possibly use a light-weight-syscall and rely on cr16 and cpu speed for more accurate timing? similiar to commit http://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/?id=18ad51dd342a7eb09dbcd059d0b451b616d4dafc  
 
** Review gettimeofday and clock_gettime() implementation, possibly use a light-weight-syscall and rely on cr16 and cpu speed for more accurate timing? similiar to commit http://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/?id=18ad51dd342a7eb09dbcd059d0b451b616d4dafc  
  
 
* Kernel: Low priority patches needs further work  
 
* Kernel: Low priority patches needs further work  
−
** Fix exception table entries to be of type int (32bit), even on 64bit kernel -> saves space. - https://patchwork.kernel.org/patch/3033801/ - partly done.
 
 
** disable FP-register save/restore at kernel entry? - https://patchwork.kernel.org/patch/3975291/
 
** disable FP-register save/restore at kernel entry? - https://patchwork.kernel.org/patch/3975291/
 
** enable 16k/64k kernel page sizes (still cacheflushing-issues with SCSI driver, additional work in glibc needed to deal with >4k page size, e.g. ld-loader alignments of data/code segments)
 
** enable 16k/64k kernel page sizes (still cacheflushing-issues with SCSI driver, additional work in glibc needed to deal with >4k page size, e.g. ld-loader alignments of data/code segments)
 
** Superdome support - can someone test? Should work better now with latest kernels.
 
** Superdome support - can someone test? Should work better now with latest kernels.
−
** task_struct/thread_info split -- task_struct should not be visible in entry.S, we need to move some items into thread_info -- this includes pt_regs and maybe some of the flags (ptrace, etc)
 
 
** Visualize FX support (will probably never happen!)
 
** Visualize FX support (will probably never happen!)
−
* PROBABLY NOT POSSIBLE: implement BUG() and WARN() so that interrupt happens in delay branch of "or,<> %r0,cond,%r0". Not possible, because this gives "control reaches end of non-void function" warnings/errors.
+
** implement BUG() and WARN() with inline assembly so that interrupt happens in delay branch of "or,<> %r0,cond,%r0". Probably not possible, because gcc will report "control reaches end of non-void function" warnings/errors.
  
 
* Debian packages / Userspace work
 
* Debian packages / Userspace work
Line 105: Line 62:
 
** 64-bit userspace, 64-bit binutils needs to be fixed to get multiple stub section support.
 
** 64-bit userspace, 64-bit binutils needs to be fixed to get multiple stub section support.
 
** gcc -> join 32- and 64bit compiler, make "-m64" work, get rid of hppa64-linux-gnu-gcc
 
** gcc -> join 32- and 64bit compiler, make "-m64" work, get rid of hppa64-linux-gnu-gcc
 +
* userspace
 +
** elfutils -- testsuite error
 +
** glib2.0 -- testsuite error
 +
** gtk+3.0 succeeds by adding this to debian/rules: touch testsuite/reftests/color-transition.ui.known_fail testsuite/reftests/linear-gradient-transition-to-other.ui.known_fail
 +
** libgadu - testsuite fails probably due to missing pthread cancellation in glibc.
 +
** glibc pthread cancellation: the main issue is the target specific code needed to handle stubs and signal frames.  See linux-unwind.h and hpux-unwind.h in gcc's libgcc/config/pa directory.  The function pa32_fallback_frame_state does the hard work for linux.  Hopefully, the signal tramp hasn't changed. All the unwind tests in gcc pass, so I think the implementation there is okay. It may be we need something similar to pa32_fallback_frame_state in glibc to fix cancellation
 +
** heimdal suceeds by increasing timeout for iprop test to 120 seconds.
 +
** cupt - build errors
 +
** libqb - testsuite error
 +
** libsoup2.4 -- testsuite error
 +
** DONE: capnproto - testsuite was wrong: PARISC uses mips-floatingpoint NANs - https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=781787
 +
** teach "blkid" about palo partition
 +
** STARTED: why is GNU AS so slow, e.g. when building freeorion, quantlib-swig or yade packages, testcases: openturns
 +
*** https://sourceware.org/bugzilla/show_bug.cgi?id=18427
 +
*** http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=787192
 +
** glibc: add backtrace() function, problems when building elfutils package: http://buildd.debian-ports.org/status/package.php?p=elfutils&suite=sid
 +
** gcc: Add _builtin_thread_pointer() and _builtin_set_thread_pointer(), e.g.: https://gcc.gnu.org/ml/gcc-patches/2012-07/msg00428.html  for glibc: ports/sysdeps/hppa/nptl/tls.h __set_cr27()
 +
** provide libphobos packages - gcc-defaults - http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=763103
 +
** gdb crashes Kernel - http://permalink.gmane.org/gmane.linux.ports.parisc/5406
 +
** debian-cd to use xorriso
 +
** debian-cd: do not compress kernel.
 
    
 
    
 
* palo (PA-RISC Boot Loader):
 
* palo (PA-RISC Boot Loader):
Line 110: Line 88:
 
** move decompression of Kernel to Linux kernel code instead of palo
 
** move decompression of Kernel to Linux kernel code instead of palo
  
−
== Kernel patches ==
+
== Upstream PA-RISC-related kernel patch history ==
 
* Kernel 3.9:
 
* Kernel 3.9:
 
** Integrate Dave Anglin's latest cacheflush patchset (fixes unusual userspace crashes) https://patchwork.kernel.org/patch/2480891/ - Dave  
 
** Integrate Dave Anglin's latest cacheflush patchset (fixes unusual userspace crashes) https://patchwork.kernel.org/patch/2480891/ - Dave  
Line 277: Line 255:
 
** DONE: in interactive mode, add reboot option command
 
** DONE: in interactive mode, add reboot option command
  
−
* Debian kernel
+
* Debian kernel / Generic Kernel bugs fixed
 +
** DONE: Fixed 64bit SMP kernel & sticon on machines with more than 4GB RAM (Kernel 3.12)
 +
** DONE: Lots of syscall fixes regarding compat kernel
 
** DONE: send patches to remove UP from debian kernels, use sil680 instead of siimage, see http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=721191 -> fixed in 3.12 / 3.13
 
** DONE: send patches to remove UP from debian kernels, use sil680 instead of siimage, see http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=721191 -> fixed in 3.12 / 3.13
 
** DONE: Enable BMC in c8000 default kernel config - ipmi driver (CONFIG_IPMI_HANDLER=y) - http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=747482
 
** DONE: Enable BMC in c8000 default kernel config - ipmi driver (CONFIG_IPMI_HANDLER=y) - http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=747482
Line 285: Line 265:
 
** DONE: Kernel build problem: a) disable -mfast-indirect-calls in Makefile, b) gcc compiler bug in > gcc-4.8.3-3
 
** DONE: Kernel build problem: a) disable -mfast-indirect-calls in Makefile, b) gcc compiler bug in > gcc-4.8.3-3
 
** DONE: Drop HPUX-support from kernel (Kernel 4.0)
 
** DONE: Drop HPUX-support from kernel (Kernel 4.0)
 +
** DONE: implemented HUGE pages
 +
** DONE: kernel timerfd_settime/timerfd_gettime issue, test with debian package:  liblinux-fd-perl -> https://patchwork.kernel.org/patch/7951661/, backported to all kernels
 +
** DONE: kernel ptrace bug: http://thread.gmane.org/gmane.linux.ports.parisc/26273, possible fix: https://patchwork.kernel.org/patch/8063301/
  
 
* Debian installer:
 
* Debian installer:
Line 293: Line 276:
  
 
* binutils:
 
* binutils:
−
** DONE: linker -fpie support, https://sourceware.org/bugzilla/show_bug.cgi?id=18421
+
** DONE: Add linker -fpie support, https://sourceware.org/bugzilla/show_bug.cgi?id=18421, debian package iceweasel and others
 +
** DONE: Using "gcc -o /dev/null" gives "ld: final link failed: File truncated" error: https://sourceware.org/bugzilla/show_bug.cgi?id=19526
  
 
* gcc:
 
* gcc:
Line 302: Line 286:
 
* glibc:
 
* glibc:
 
** DONE: Problem building libcap-ng...The problem is the call to __tls_get_addr clobbers the value loaded into %r26 for the printf call. -> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59777  
 
** DONE: Problem building libcap-ng...The problem is the call to __tls_get_addr clobbers the value loaded into %r26 for the printf call. -> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59777  
 +
** DONE: O_SYNC wrongly defined in glibc (typo in hppa __O_SYNC define in <bits/fcntl.h>)
 
** DONE: eglibc mmap bug with localedef (see debian bugzilla http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=741243)
 
** DONE: eglibc mmap bug with localedef (see debian bugzilla http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=741243)
 +
** DONE: Wrong __gmon_start__ symbol, showed in vlc - linker does not drop dependencies to unneeded shared libs, https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=770741. glibc bug report: https://sourceware.org/bugzilla/show_bug.cgi?id=19170
 +
** DONE: nss - dladdr from glibc returns wrong names on hppa, http://article.gmane.org/gmane.linux.ports.parisc/26290, Solution: https://sourceware.org/bugzilla/show_bug.cgi?id=19415
 +
** DONE: gcc userspace support for variable CAS (gcc atomic ops, __sync_XXX), initial support in kernel 3.17, http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=89206491201cbd1571009b36292af781cef74c1b)
 +
*** gcc-4.9 commit: https://gcc.gnu.org/viewcvs/gcc?view=revision&revision=227914
 +
*** gcc-5.0 commit: https://gcc.gnu.org/viewcvs/gcc?view=revision&revision=227913
 +
*** packages needing rebuild: libatomic-ops (gauche and libgc have copies of libatomic-ops included), x265, ceph, lizardfs, firebird2.5, gauche, idba, mapsembler2, padevchooser, filezilla, idba, http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=785654
 +
*** __sync_val_compare_and_swap_8 was missing in: openvdb, idba, lizardfs, ...
  
 
* Debian userspace:
 
* Debian userspace:
 
** DONE: qt4-x11: Improve atomic support on parisc (https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=741190)
 
** DONE: qt4-x11: Improve atomic support on parisc (https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=741190)
 +
** DONE: haskell (ghc) hangs: futex(0xea00d08e, FUTEX_WAIT_PRIVATE, 2, NULL) = -1 EINVAL (Invalid argument). Solution: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=756260#15
 +
** DONE: port openmpi package: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=776730
  
 
== ABI breakages (HP-UX support) ==
 
== ABI breakages (HP-UX support) ==
Line 337: Line 331:
  
 
== Updates for next newsletter ==
 
== Updates for next newsletter ==
−
* fixed 64bit SMP kernel & sticon on machines with more than 4GB RAM
 
−
* Dave found O_SYNC bug in glibc (Typo in hppa __O_SYNC define in <bits/fcntl.h>)
 
−
* lots of syscall fixes regarding compat kernel
 
 
* Migration parisc-linux.org
 
* Migration parisc-linux.org
 
* debian-ports.org access
 
* debian-ports.org access
−
* 3.13 with all patches in (coherency)
 
 
* buildds running
 
* buildds running
 
* wiki migration finished (05 Feb 2014)
 
* wiki migration finished (05 Feb 2014)

Revision as of 16:02, 17 March 2016

Contents

PA-RISC Linux TODO (actually this is Helge's TODO list)

Status & TODO 2016-03-17

  • add wiki page about kernel cross-compiling
  • Kernel: still missing Linux kernel development (no patches available yet):
    • Fix kernel to load -fpie compiled programs not at 0x400000
    • maybe implement transparent huge pages (due to cache coherency technically only possible on PA8500-PA8700 CPUs, not possible on PA8800-PA8900 CPUs)
    • Fix FTRACE: https://patchwork.kernel.org/patch/3554391/
    • optimize flush_dcache_page() implementation like sparc. See Documentation/cachetlb.txt, e.g. used in aio subsystem
    • Implement CONFIG_DEBUG_STACK_USAGE ?
    • fix wrong _init notation in kernel
    • NO_IRQ -> should be converted to (0) instead of (-1) -> Helge (check assembly, check in drivers/parisc/*)
    • implement alternative coding for UP/SMP. See arch/x86/include/asm/alternative.h
    • check assembly in drivers/parisc/*
    • use pdc_add_valid() in kern_addr_valid(), protect against crashable acesses (like for superio-pc-drivers-detection in lp), add pdc_add_valid() to create a probe_kernel_read() implementation.
    • arch/parisc/kernel/sys_parisc.c -> use SYSCALL_ALIAS() -> maybe
    • implement libnuma kernel support to report on C8000 one socket with 2 cores.
    • Review gettimeofday and clock_gettime() implementation, possibly use a light-weight-syscall and rely on cr16 and cpu speed for more accurate timing? similiar to commit http://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/?id=18ad51dd342a7eb09dbcd059d0b451b616d4dafc
  • Kernel: Low priority patches needs further work
    • disable FP-register save/restore at kernel entry? - https://patchwork.kernel.org/patch/3975291/
    • enable 16k/64k kernel page sizes (still cacheflushing-issues with SCSI driver, additional work in glibc needed to deal with >4k page size, e.g. ld-loader alignments of data/code segments)
    • Superdome support - can someone test? Should work better now with latest kernels.
    • Visualize FX support (will probably never happen!)
    • implement BUG() and WARN() with inline assembly so that interrupt happens in delay branch of "or,<> %r0,cond,%r0". Probably not possible, because gcc will report "control reaches end of non-void function" warnings/errors.
  • Debian packages / Userspace work
    • IMPORTANT: fix OpenJDK - fixes required in zeroport (stack grows up!).
    • Fix qemu
    • 64-bit userspace, 64-bit binutils needs to be fixed to get multiple stub section support.
    • gcc -> join 32- and 64bit compiler, make "-m64" work, get rid of hppa64-linux-gnu-gcc
  • userspace
  • palo (PA-RISC Boot Loader):
    • Modify (if possible), that palo can read (sequential) tftpbootable-vmlinuxes as well as the current boot-vmlinux images (simplifies build/handling)
    • move decompression of Kernel to Linux kernel code instead of palo

Upstream PA-RISC-related kernel patch history

Debian kernel and userspace programs

  • Debian kernel / Generic Kernel bugs fixed
    • DONE: Fixed 64bit SMP kernel & sticon on machines with more than 4GB RAM (Kernel 3.12)
    • DONE: Lots of syscall fixes regarding compat kernel
    • DONE: send patches to remove UP from debian kernels, use sil680 instead of siimage, see http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=721191 -> fixed in 3.12 / 3.13
    • DONE: Enable BMC in c8000 default kernel config - ipmi driver (CONFIG_IPMI_HANDLER=y) - http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=747482
    • DONE: remove HIL RTC from debian kernel - it creates a backtrace at boot because we have already another rtc0 active.
    • DONE: add ppdev driver, needed by CUPS, systemd reports missing module load (3.17 & 3.16 kernel)
    • DONE: Add crc-modules, event, isofs, jfs, mouse, sata, sound, squashfs, udf, uinput, scsi-common, scsi-extra, loop, serial and usb-serial debian-installer packages
    • DONE: Kernel build problem: a) disable -mfast-indirect-calls in Makefile, b) gcc compiler bug in > gcc-4.8.3-3
    • DONE: Drop HPUX-support from kernel (Kernel 4.0)
    • DONE: implemented HUGE pages
    • DONE: kernel timerfd_settime/timerfd_gettime issue, test with debian package: liblinux-fd-perl -> https://patchwork.kernel.org/patch/7951661/, backported to all kernels
    • DONE: kernel ptrace bug: http://thread.gmane.org/gmane.linux.ports.parisc/26273, possible fix: https://patchwork.kernel.org/patch/8063301/

ABI breakages (HP-UX support)

parisc-linux.org Homepage and Debian build server infrastructure work

Updates for next newsletter

  • Migration parisc-linux.org
  • debian-ports.org access
  • buildds running
  • wiki migration finished (05 Feb 2014)
  • old parisc-linux kernel mailinglist imported into gmane.org - http://dir.gmane.org/gmane.linux.ports.parisc - 23 Nov 2014
  • parisc-linux domain is now owned by Helge (August 2014)
  • Debian 8.0 installer images available.

Testsuites, kernel and userspace testing tools

  • GLIBC issues: https://sourceware.org/glibc/wiki/PortStatus
  • We also have a problem with poll in rpcbind on 32-bit kernels. Runs forever. Not analyzed yet.
  • glib2.0 atomic bug: Dave: I see the problem building various packages -- apt is is one. It occurs when multiple dot processes are started by graph when package documentation is being generated. The number of dot processes spawned depends on the number of processors on the build system. However, the faults occur even on a UP system. I debugged one of these segvs enough to see that the problem occurred in a glib2.0 library. It's using the GCC "sync" interface.
  • From Dave: gcc thread issues including thread cancellation, glib2.0 problem, maybe __tls_get_addr clobber too?

TODO Drivers

  • IPMI support on rp-class / Support graceful shutdown from BMC on rp servers (`PC -GRACEFUL`)
  • write Lasi floppy driver
  • write Suckyio floppy driver
  • write spifi driver (rbrad)
  • modify ncr53c8xx driver for Outfield (735 & 755)
  • write GSC FDDI driver
  • write Timi ASIC (74x) support
  • EISA DMA support

Review all the todo entries below!

  • ns87415 dma doesn't work reliably on suckyio-systems
Retrieved from "https://parisc.wiki.kernel.org/index.php?title=TODO&oldid=8069"
Personal tools