1. 12 Apr, 2004 40 commits
    • Andrew Morton's avatar
      [PATCH] v850: make v850 dma-mapping.h header work when !CONFIG_PCI · ba5ff4c5
      Andrew Morton authored
      From: <miles@mcspd15.ucom.lsi.nec.co.jp> (Miles Bader)
      
      Is this something that should be done in <asm-generic/dma-mapping.h>?
      ba5ff4c5
    • Andrew Morton's avatar
      [PATCH] v850: use volatile qualifier on v850 test-n-bitop asm statements · d54f0b47
      Andrew Morton authored
      From: <miles@mcspd15.ucom.lsi.nec.co.jp> (Miles Bader)
      
      Otherwise the compiler can delete them (this is one of those "how on earth
      did it ever work before" moments).
      d54f0b47
    • Andrew Morton's avatar
      [PATCH] fix posix-timers to have proper per-process scope · 0e568881
      Andrew Morton authored
      From: Roland McGrath <roland@redhat.com>
      
      The posix-timers implementation associates timers with the creating thread
      and destroys timers when their creator thread dies.  POSIX clearly
      specifies that these timers are per-process, and a timer should not be torn
      down when the thread that created it exits.  I hope there won't be any
      controversy on what the correct semantics are here, since POSIX is clear
      and the Linux feature is called "posix-timers".
      
      The attached program built with NPTL -lrt -lpthread demonstrates the bug.
      The program is correct by POSIX, but fails on Linux.  Note that a until
      just the other day, NPTL had a trivial bug that always disabled its use of
      kernel timer syscalls (check strace for lack of timer_create/SYS_259).  So
      unless you have built your own NPTL libs very recently, you probably won't
      see the kernel calls actually used by this program.
      
      Also attached is my patch to fix this.  It (you guessed it) moves the
      posix_timers field from task_struct to signal_struct.  Access is now
      governed by the siglock instead of the task lock.  exit_itimers is called
      from __exit_signal, i.e.  only on the death of the last thread in the
      group, rather than from do_exit for every thread.  Timers' it_process
      fields store the group leader's pointer, which won't die.  For the case of
      SIGEV_THREAD_ID, I hold a ref on the task_struct for it_process to stay
      robust in case the target thread dies; the ref is released and the dangling
      pointer cleared when the timer fires and the target thread is dead.  (This
      should only come up in a buggy user program, so noone cares exactly how the
      kernel handles that case.  But I think what I did is robust and sensical.)
      
      /* Test for bogus per-thread deletion of timers.  */
      
      #include <stdio.h>
      #include <error.h>
      #include <time.h>
      #include <signal.h>
      #include <stdint.h>
      #include <sys/time.h>
      #include <sys/resource.h>
      #include <unistd.h>
      #include <pthread.h>
      
      /* Creating timers in another thread should work too.  */
      static void *do_timer_create(void *arg)
      {
      	struct sigevent *const sigev = arg;
      	timer_t *const timerId = sigev->sigev_value.sival_ptr;
      	if (timer_create(CLOCK_REALTIME, sigev, timerId) < 0) {
      		perror("timer_create");
      		return NULL;
      	}
      	return timerId;
      }
      
      int main(void)
      {
      	int i, res;
      	timer_t timerId;
      	struct itimerspec itval;
      	struct sigevent sigev;
      
      	itval.it_interval.tv_sec = 2;
      	itval.it_interval.tv_nsec = 0;
      	itval.it_value.tv_sec = 2;
      	itval.it_value.tv_nsec = 0;
      
      	sigev.sigev_notify = SIGEV_SIGNAL;
      	sigev.sigev_signo = SIGALRM;
      	sigev.sigev_value.sival_ptr = (void *)&timerId;
      
      	for (i = 0; i < 100; i++) {
      		printf("cnt = %d\n", i);
      
      		pthread_t thr;
      		res = pthread_create(&thr, NULL, &do_timer_create, &sigev);
      		if (res) {
      			error(0, res, "pthread_create");
      			continue;
      		}
      		void *val;
      		res = pthread_join(thr, &val);
      		if (res) {
      			error(0, res, "pthread_join");
      			continue;
      		}
      		if (val == NULL)
      			continue;
      
      		res = timer_settime(timerId, 0, &itval, NULL);
      		if (res < 0)
      			perror("timer_settime");
      
      		res = timer_delete(timerId);
      		if (res < 0)
      			perror("timer_delete");
      	}
      
      	return 0;
      }
      0e568881
    • Andrew Morton's avatar
      [PATCH] sh-sci compile error fix patch · 2aa53d18
      Andrew Morton authored
      From: Yoshinori Sato <ysato@users.sourceforge.jp>
      
      - add Kconfig depends H8300
      - H8/300 support compile error fixed.
      2aa53d18
    • Andrew Morton's avatar
      [PATCH] H8/300 support update · 8d94c528
      Andrew Morton authored
      From: Yoshinori Sato <ysato@users.sourceforge.jp>
      
      - fix any error/warning
      - fix {request,freee}_irq interrupt control fix
      - add dump_stack
      - fix show_trace_task
      - fix typo
      8d94c528
    • Andrew Morton's avatar
      [PATCH] H8/300 support update (3/3) - others · 092ed951
      Andrew Morton authored
      From: Yoshinori Sato <ysato@users.sourceforge.jp>
      
      - use new serial driver (drivers/serial/sh-sci.[ch])
      - typo fix
      - add message level
      092ed951
    • Andrew Morton's avatar
      [PATCH] H8/300 support update (2/3) - entry.S cleanup · 4280119e
      Andrew Morton authored
      From: Yoshinori Sato <ysato@users.sourceforge.jp>
      
      - cleanup define
      4280119e
    • Andrew Morton's avatar
      [PATCH] H8/300 support update (1/3) - ptrace fix · 4f9ad28f
      Andrew Morton authored
      From: Yoshinori Sato <ysato@users.sourceforge.jp>
      
      - fix PTRACE_SIGLESTEP bug.
      - separate to CPU depend.
      4f9ad28f
    • Andrew Morton's avatar
      [PATCH] use EFLAGS #defines instead of inline constants · 219c7ea1
      Andrew Morton authored
      From: "Randy.Dunlap" <rddunlap@osdl.org>
      
      Use x86 EFLAGS defines in place of hardwired constants.
      219c7ea1
    • Andrew Morton's avatar
      [PATCH] stack reduction: ISDN · 76e62eba
      Andrew Morton authored
      From: Arjan van de Ven <arjanv@redhat.com>
      
      isdn: dynamically allocate big structures
      76e62eba
    • Andrew Morton's avatar
      [PATCH] stack reductions: ide · 1a92fb13
      Andrew Morton authored
      From: Arjan van de Ven <arjanv@redhat.com>
      
      ide.c: constant array of strings can be static
      1a92fb13
    • Andrew Morton's avatar
      [PATCH] stack reduction: ide-cd · ff94e4ae
      Andrew Morton authored
      From: Arjan van de Ven <arjanv@redhat.com>
      
      ide-cd: a few 512 byte scratch buffers can be static; they are just for
      putting "padding" sectors in that aren't used.
      
      (acked by Jens)
      ff94e4ae
    • Andrew Morton's avatar
      [PATCH] binfmt_elf.c fix for 32-bit apps with large bss · 9b2bc421
      Andrew Morton authored
      From: Julie DeWandel <jdewand@redhat.com>
      
      A problem exists where a 32-bit application can have a huge bss, one that
      is so large that an overflow of the TASK_SIZE happens.  But in this case,
      the overflow is not detected in load_elf_binary().  Instead, because
      arithmetic is being done using 32-bit containers, a truncation occurs and
      the program gets loaded when it shouldn't have been.  Subsequent execution
      yields unpredictable results.
      
      The attached patch fixes this problem by checking for the overflow
      condition and sending a SIGKILL to the application if the overflow is
      detected.  This problem can in theory exist when loading the elf
      interpreter as well, so a similar check was added there.
      9b2bc421
    • Andrew Morton's avatar
      [PATCH] es1688 Definition redundancy · a804dbaf
      Andrew Morton authored
      From: Fabian Frederick <Fabian.Frederick@skynet.be>
      
      Here's a trivial patch to avoid definition redundancy in es1688.
      a804dbaf
    • Andrew Morton's avatar
      [PATCH] intermezzo leak fixes · 7804ee5a
      Andrew Morton authored
      - Don't leak a pathname ref on error
      
      - Don't do putname() on a nameidata.
      7804ee5a
    • Andrew Morton's avatar
      [PATCH] more i386 head.S cleanups · 8ae81ff5
      Andrew Morton authored
      From: Brian Gerst <bgerst@didntduck.org>
      
      - Move empty_zero_page and swapper_pg_dir to BSS.  This requires that BSS
        is cleared earlier, but reclaims over 3k that was lost due to page
        alignment.
      
      - Move stack_start, ready, and int_msg, boot_gdt_descr, idt_descr, and
        cpu_gdt_descr to .data.  They were interfering with disassembly while in
        .text.
      8ae81ff5
    • Andrew Morton's avatar
      [PATCH] unmap_vmas latency improvement · f0d53a52
      Andrew Morton authored
      unmap_vmas() will cause scheduling latency when tearing down really big vmas
      on !CONFIG_PREEMPT.  That's a bit unkind to the non-preempt case, so let's do
      a cond_resched() after zapping 1024 pages.
      f0d53a52
    • Andrew Morton's avatar
      [PATCH] CONFIG_SND_MIXART doesn't compile · b3548dd4
      Andrew Morton authored
      From: Bernhard Rosenkraenzer <bero@arklinux.org>
      
      mixart.h uses tasklet_struct without including linux/interrupt.h -- fix
      attached.
      b3548dd4
    • Andrew Morton's avatar
      [PATCH] selinux: remove ratelimit from avc · 1eb3edb2
      Andrew Morton authored
      From: Stephen Smalley <sds@epoch.ncsc.mil>
      
      This patch drops the ratelimit code from the SELinux avc, as this can now
      be handled by the audit framework.  Enabling and setting the ratelimit is
      then left to userspace.
      1eb3edb2
    • Andrew Morton's avatar
      [PATCH] selinux: Audit compute_sid errors · df25ad33
      Andrew Morton authored
      From: Stephen Smalley <sds@epoch.ncsc.mil>
      
      This patch changes an error message printk'd by security_compute_sid to use
      the audit framework instead.  These errors reflect situations where a
      security transition would normally occur due to policy, but the resulting
      security context is not valid.  The patch also changes the code to always
      call the audit framework rather than only doing so when permissive as this
      was causing problems with testing policy, and does some code cleanup.
      df25ad33
    • Andrew Morton's avatar
      [PATCH] selinux: make IPv6 code work with audit framework · 7787c5a4
      Andrew Morton authored
      From: James Morris <jmorris@redhat.com>
      
      This patch makes the IPv6 code work with the audit framework, following the
      merge of both.
      7787c5a4
    • Andrew Morton's avatar
      [PATCH] Light-weight Auditing Framework · f85a96f6
      Andrew Morton authored
      From: Rik Faith <faith@redhat.com>
      
      This patch provides a low-overhead system-call auditing framework for Linux
      that is usable by LSM components (e.g., SELinux).  This is an update of the
      patch discussed in this thread:
      
          http://marc.theaimsgroup.com/?t=107815888100001&r=1&w=2
      
      In brief, it provides for netlink-based logging of audit records that have
      been generated in other parts of the kernel (e.g., SELinux) as well as the
      ability to audit system calls, either independently (using simple
      filtering) or as a compliment to the audit record that another part of the
      kernel generated.
      
      The main goals were to provide system call auditing with 1) as low overhead
      as possible, and 2) without duplicating functionality that is already
      provided by SELinux (and/or other security infrastructures).  This
      framework will work "stand-alone", but is not designed to provide, e.g.,
      CAPP functionality without another security component in place.
      
      This updated patch includes changes from feedback I have received,
      including the ability to compile without CONFIG_NET (and better use of
      tabs, so use -w if you diff against the older patch).
      
      Please see http://people.redhat.com/faith/audit/ for an early example
      user-space client (auditd-0.4.tar.gz) and instructions on how to try it.
      
      My future intentions at the kernel level include improving filtering (e.g.,
      syscall personality/exit codes) and syscall support for more architectures.
       First, though, I'm going to work on documentation, a (real) audit daemon,
      and patches for other user-space tools so that people can play with the
      framework and understand how it can be used with and without SELinux.
      
      
      Update:
      
      Light-weight Auditing Framework receive filter fixes
      From: Rik Faith <faith@redhat.com>
      
      Since audit_receive_filter() is only called with audit_netlink_sem held, it
      cannot race with either audit_del_rule() or audit_add_rule(), so the
      list_for_each_entry_rcu()s may be replaced by list_for_each_entry()s, and
      the rcu_read_{un,}lock()s removed.  A fix for this is part of the attached
      patch.
      
      Other features of the attached patch are:
      
      1) generalized the ability to test for inequality
      
      2) added syscall exit status reporting and testing
      
      3) added ability to report and test first 4 syscall arguments (this adds
         a large amount of flexibility for little cost; not implemented or tested
         on ppc64)
      
      4) added ability to report and test personality
      
      User-space demo program enhanced for new fields and inequality testing:
      http://people.redhat.com/faith/audit/auditd-0.5.tar.gz
      f85a96f6
    • Andrew Morton's avatar
      [PATCH] From: James Morris <jmorris@redhat.com> · 0e8e57e3
      Andrew Morton authored
      This patch removes a harmless duplicate assignment from the IPv6 code.
      0e8e57e3
    • Andrew Morton's avatar
      [PATCH] selinux: add IPv6 support · 5e752b7e
      Andrew Morton authored
      From: James Morris <jmorris@redhat.com>
      
      The patch below adds explicit IPv6 support to SELinux.
      
      Brief description of changes:
      
      o IPv6 networking is now subject to the same controls as IPv4 (in
        addition to the generic socket permissions which cover all protocols),
        namely: bind to local node address; bind to local port; send & receive
        TCP/UDP and raw IP packets based on local network interface and remote
        node address.
      
      o Packet parsing has been extended to IPv6 packets for logging and
        control, and simplified for IPv4.
      
      o Support for logging of IPv6 addresses has also been added.
      
      o The kernel policy database code has been modified to support IPv6, and
        reworked to provide generic security policy version handling so that
        older policy versions will still work, making upgrading simpler.
      
      Corresponding userspace patches are available at
      <http://people.redhat.com/jmorris/selinux/ipv6/>, although current
      userspace tools will continue to function normally (but without explicit
      IPv6 support).
      
      For more details at the security management level, see
      <http://marc.theaimsgroup.com/?l=selinux&m=108068187630948&w=2>
      
      This code has been under testing and review for several weeks.
      5e752b7e
    • Andrew Morton's avatar
      [PATCH] reiserfs writepage race with data=ordered · bcf506bd
      Andrew Morton authored
      From: Chris Mason <mason@suse.com>
      
      reiserfs-writepage-ordered-race needs a minor update to include your latest
      __block_write_full_page fixes for the direct_read_under bug Daniel was
      hitting.
      bcf506bd
    • Andrew Morton's avatar
      [PATCH] reiserfs_kfree warning fix · b566678f
      Andrew Morton authored
      fs/reiserfs/journal.c: In function `reiserfs_end_persistent_transaction':
      fs/reiserfs/journal.c:2616: warning: unused variable `s'
      
      Make the functions static inline so that typechecking is enabled if
      !CONFIG_REISERFS_CHECK.
      b566678f
    • Andrew Morton's avatar
      [PATCH] reiserfs: fix dirty-buffer warnings · 77ae13bc
      Andrew Morton authored
      From: Chris Mason <mason@suse.com>
      
      block_write_full_page() might see and lock clean metadata buffers, which leads
      to journal-1777 messages.  Change the message to ignore bh locked.
      77ae13bc
    • Andrew Morton's avatar
      [PATCH] reiserfs: scheduling latency improvements · 6f2085c0
      Andrew Morton authored
      From: Chris Mason <mason@suse.com>
      
      Some latency improvements for the reiserfs data=ordered code from Takashi.
      6f2085c0
    • Andrew Morton's avatar
      [PATCH] reiserfs: truncate leak fix · e2de7edb
      Andrew Morton authored
      From: Chris Mason <mason@suse.com>
      
      reiserfs_unmap_buffer should clean and wait on all buffers.  This fixes a
      leak under fsx workloads.
      e2de7edb
    • Andrew Morton's avatar
      [PATCH] reiserfs: laptop-mode support · 008c597b
      Andrew Morton authored
      From: Chris Mason <mason@suse.com>
      
      Add reiserfs support for laptop mode.
      008c597b
    • Andrew Morton's avatar
      [PATCH] reiserfs: sparse file handling fix · 34cd7802
      Andrew Morton authored
      From: Chris Mason <mason@suse.com>
      
      reiserfs_file_write makes a hole one block too large if it is the first thing
      in the file.
      34cd7802
    • Andrew Morton's avatar
      [PATCH] reiserfs: fix race with writepage · 7a06b83d
      Andrew Morton authored
      From: Chris Mason <mason@suse.com>
      
      Fix reiserfs_writepage so it doesn't race with data=ordered writes.  This
      still has a pending fix to redirty the page when it finds a locked buffer. 
      Waiting for Andrew to finish sorting that out on ext3 first.
      7a06b83d
    • Andrew Morton's avatar
      [PATCH] reiserfs: tail repacking fix · fb1b3b04
      Andrew Morton authored
      From: Chris Mason <mason@suse.com>
      
      Repacking a tail might leave a journal handle attached to an unmapped buffer.
       If that buffer gets dirtied again (via mmap for example), the reiserfs
      data=ordered code might try to write the dirty unmapped buffer to disk.
      
      The fix is to make sure we remove the journal handle when we unmap buffers.
      fb1b3b04
    • Andrew Morton's avatar
      [PATCH] reiserfs: preallocation support · d12e3392
      Andrew Morton authored
      From: Chris Mason <mason@suse.com>
      
      Enable preallocation for reiserfs_file_write when the write size is smaller
      than the default preallocation size.
      d12e3392
    • Andrew Morton's avatar
      [PATCH] reiserfs: locking fix · ec6f9553
      Andrew Morton authored
      From: Chris Mason <mason@suse.com>
      
      Make sure to hold the BKL while ending a transaction in the error path or
      reiserfs_prepare_write.
      ec6f9553
    • Andrew Morton's avatar
      [PATCH] reiserfs: data=ordered support · bb0d9672
      Andrew Morton authored
      From: Chris Mason <mason@suse.com>
      
      reiserfs data=ordered support.
      bb0d9672
    • Andrew Morton's avatar
      [PATCH] reiserfs: logging rework · 7c563ced
      Andrew Morton authored
      From: Chris Mason <mason@suse.com>
      
      reiserfs logging rework, making things much faster for small transactions. 
      metadata buffers are dirtied when they are safe to write, so normal kernel
      mechanisms can contribute to log cleaning.
      7c563ced
    • Andrew Morton's avatar
      [PATCH] reiserfs: cleanups · 8f576882
      Andrew Morton authored
      From: Chris Mason <mason@suse.com>
      
      reiserfs cleanup, get rid of old debugging code.
      8f576882
    • Andrew Morton's avatar
      [PATCH] reiserfs: support for nested transactions · 7a10b433
      Andrew Morton authored
      From: Chris Mason <mason@suse.com>
      
      reiserfs support for nested transactions.  This originally came from Peter
      Braam for 2.4.x and was ported forward by Jeff Mahoney.
      7a10b433
    • Andrew Morton's avatar
      [PATCH] Fix ext3 transaction batching · 545b0840
      Andrew Morton authored
      ext3 transaction batching has been ineffective since the scheduler changes
      forced us to replace the yield() with a schedule().
      
      Using schedule_timeout(1) fixes it up again.  Benchmarking is positive with
      wither a 1 or 10 millisecond delay in there, so there appears to be no need
      to play around with HZ.
      545b0840