ChangeSet@1.1514, 2005-05-06 19:02:44-03:00, khali@linux-fr.org
  [PATCH] I2C updates: Fix an iteration bug in the handling of i2c client module parameters.
  
  Fix an iteration bug in the handling of i2c client module parameters.
  The "force" module parameter is a list of adapter, address pairs, not
  triplets. The current code would only handle the first, fourth, seventh
  etc. pairs properly. I guess that nobody ever needed more than one pair,
  or the bug would have been noticed way earlier. This bug was originally
  fixed by myself in Linux 2.6.
  
  http://marc.theaimsgroup.com/?l=linux-kernel&m=111231616107325
  http://linux.bkbits.net:8080/linux-2.5/diffs/drivers/i2c/i2c-core.c@1.66

ChangeSet@1.1513, 2005-05-06 19:02:01-03:00, khali@linux-fr.org
  [PATCH] I2C updates: Fix I2C_FUNC_* defines in i2c.h.
  
  Fix I2C_FUNC_* defines in i2c.h.
  
  These defines lack surroundings:
  parentheses, which might cause unexpected results where used. These were
  originally fixed by Corey Minyard in Linux 2.6.
  
  http://marc.theaimsgroup.com/?l=linux-kernel&m=110928784629301
  http://linux.bkbits.net:8080/linux-2.5/diffs/include/linux/i2c.h@1.48

ChangeSet@1.1512, 2005-05-06 18:59:41-03:00, khali@linux-fr.org
  [PATCH] I2C updates: Fix typo in a comment in i2c.h.
  
  Fix a typo in a comment in i2c.h. It was originally fixed by Corey
  Minyard in Linux 2.6.
  
  http://linux.bkbits.net:8080/linux-2.5/diffs/include/linux/i2c.h@1.48

ChangeSet@1.1511, 2005-04-30 11:05:23-03:00, David.Monniaux@ens.fr
  [PATCH] fix moxa crash with more than one 1 board
  
  SYMPTOMS:
  The current Moxa Intellio driver (moxa.c) panics when using > 1 board.
  
  BACKGROUND:
  The Moxa board needs a firmware download (see
  http://www.moxa.com/drivers/C320T/Linux/v5.4/MXDRV.TGZ, command
  moxaload -y) prior to usage.
  
  Unfortunately, the current Linux kernel code fails during this download if
  more than one board are installed.
  
  EXPLANATION AND FIX:
  The MoxaDriverPoll function does:
  [...]
          for (card = 0; card < MAX_BOARDS; card++) {
                  if ((ports = moxa_boards[card].numPorts) == 0)
                          continue;
                  if (readb(moxaIntPend[card]) == 0xff) {
  [...]
  
  Unfortunately, with multiple boards, there exists a point where
  MoxaDriverPoll() will be called when moxa_boards[card].numPorts != 0 but
  moxaIntPend[card] is still NULL. Result: kernel panic.
  
  Fix: use instead
                  if ((ports = moxa_boards[card].numPorts) == 0
                      || moxaIntPend[card] == 0)
                          continue;
  
  If someone who understands the code better than me proposes a better
  patch, I'd be delighted. [For the little story, the above patch was
  written after finding the bug in a remote location without internet access
  using a serial console for getting the panic trace...]

ChangeSet@1.1510, 2005-04-30 10:33:00-03:00, zaitcev@redhat.com
  [PATCH] USB: Add HX type pl2303
  
  This adds so-called "HX" with a backport from 2.6. We ship this with RHEL 3,
  so I thought I would post it for great justice, in case someone needs it.
  The failure symptom is that adapter gets recognized fine, receive works fine,
  but transmit doesn't.

ChangeSet@1.1507, 2005-04-29 10:57:21-03:00, ebs@ebshome.net
  [PATCH] ppc32: backport Book-E decrementer handling fix from 2.6
  
  Marcelo,
  
  this is backport of Matt Porter's patch for Book-E decrementer
  handling in timer_interrupt.
  
  The fix has been in 2.6 from August but never made it to 2.4, and I
  re-discovered this fix last week :)
  
  Original Matt's post to linuxppc-dev with explanation can be found at:
  http://ozlabs.org/pipermail/linuxppc-dev/2004-August/017458.html
  
  Signed-off-by: Eugene Surovegin <ebs@ebshome.net>
  
  ===== arch/ppc/kernel/time.c 1.16 vs edited =====

ChangeSet@1.1506, 2005-04-27 12:28:47-03:00, mikpe@user.it.uu.se
  [PATCH] rwsem-spinlock linkage error
  
  The
  
  Andrew Morton:
    o rwsem: Make rwsems use interrupt disabling spinlocks
  
  change in 2.4.31-pre1 has a typo: one occurrence of spin_unlock() was
  changed to spin_unlock_restore() instead of spin_unlock_irqrestore()
  as was obviously the intention. Since spin_unlock_restore() doesn't
  exist, this results in linkage errors on x86_64 and other archs using
  CONFIG_RWSEM_GENERIC_SPINLOCK.
  
  Trival fix below.
  
  Signed-off-by: Mikael Pettersson <mikpe@csd.uu.se>

ChangeSet@1.1505, 2005-04-22 07:07:58-03:00, marcelo@logos.cnet
  Change VERSION to 2.4.31-pre1
  TAG: v2.4.31-pre1