Skip to content

L
linux
Switch branch/tag
  1. 15 Aug, 2005 8 commits
  3. 05 Aug, 2005 14 commits
  5. 15 Jul, 2005 11 commits
  7. 29 Jun, 2005 7 commits
    • Chris Wright's avatar
      Linux 2.6.12.2 · 88800936
      Chris Wright authored
      88800936
    • David S. Miller's avatar
      [PATCH] Fix two socket hashing bugs. · 1b1f8246
      David S. Miller authored 
      1) netlink_release() should only decrement the hash entry
   count if the socket was actually hashed.

   This was causing hash->entries to underflow, which
   resulting in all kinds of troubles.

   On 64-bit systems, this would cause the following
   conditional to erroneously trigger:

	err = -ENOMEM;
	if (BITS_PER_LONG > 32 && unlikely(hash->entries >= UINT_MAX))
		goto err;

2) netlink_autobind() needs to propagate the error return from
   netlink_insert().  Otherwise, callers will not see the error
   as they should and thus try to operate on a socket with a zero pid,
   which is very bad.

   However, it should not propagate -EBUSY.  If two threads race
   to autobind the socket, that is fine.  This is consistent with the
   autobind behavior in other protocols.

   So bug #1 above, combined with this one, resulted in hangs
   on netlink_sendmsg() calls to the rtnetlink socket.  We'd try
   to do the user sendmsg() with the socket's pid set to zero,
   later we do a socket lookup using that pid (via the value we
   stashed away in NETLINK_CB(skb).pid), but that won't give us the
   user socket, it will give us the rtnetlink socket.  So when we
   try to wake up the receive queue, we dive back into rtnetlink_rcv()
   which tries to recursively take the rtnetlink semaphore.

Thanks to Jakub Jelink for providing backtraces.  Also, thanks to
Herbert Xu for supplying debugging patches to help track this down,
and also finding a mistake in an earlier version of this fix.
Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
Signed-off-by: default avatarChris Wright <chrisw@osdl.org>
      1b1f8246
    • Linus Torvalds's avatar
      [PATCH] ACPI: Make sure we call acpi_register_gsi() even for default PCI interrupt assignment · 1ef0867a
      Linus Torvalds authored 
      ACPI: Make sure we call acpi_register_gsi() even for default PCI interrupt assignment

That's the part that keeps track of the ELCR register, and we want to
make sure that the PCI interrupts are properly marked level/low.
Signed-off-by: default avatarChris Wright <chrisw@osdl.org>
      1ef0867a
    • Linus Torvalds's avatar
      [PATCH] Add "memory" clobbers to the x86 inline asm of strncmp and friends · 2e8e1219
      Linus Torvalds authored 
      Add "memory" clobbers to the x86 inline asm of strncmp and friends

They don't actually clobber memory, but gcc doesn't even know they
_read_ memory, so can apparently re-order memory accesses around them.

Which obviously does the wrong thing if the memory access happens to
change the memory that the compare function is accessing..

Verified to fix a strange boot problem by Jens Axboe.
Signed-off-by: default avatarChris Wright <chrisw@osdl.org>
      2e8e1219
    • Mitch Williams's avatar
      [PATCH] e1000: fix spinlock bug · 26b1062c
      Mitch Williams authored 
      This patch fixes an obvious and nasty bug where we could exit the transmit
routine while holding tx_lock.
Signed-off-by: default avatarMitch Williams <mitch.a.williams@intel.com>
Signed-off-by: default avatarChris Wright <chrisw@osdl.org>
      26b1062c
    • Hugh Dickins's avatar
      [PATCH] fix remap_pte_range BUG · 986bd430
      Hugh Dickins authored 
      Out-of-tree user of remap_pfn_range hit kernel BUG at mm/memory.c:1112!
It passes an unrounded size to remap_pfn_range, which was okay before
2.6.12, but misses remap_pte_range's new end condition.  An audit of
all the other ptwalks confirms that this is the only one so exposed.
Signed-off-by: default avatarHugh Dickins <hugh@veritas.com>
Signed-off-by: default avatarChris Wright <chrisw@osdl.org>
      986bd430
    • Andrew Vasquez's avatar
      [PATCH] qla2xxx: Pull-down scsi-host-addition to follow board initialization. · 906aa4d2
      Andrew Vasquez authored 
      Return to previous held-logic of calling scsi_add_host() only
after the board has been completely initialized.  Also return
pci_*() error-codes during probe failure paths.

This also corrects an issue where only lun 0 is being scanned for
a given port.
Signed-off-by: default avatarAndrew Vasquez <andrew.vasquez@qlogic.com>
Signed-off-by: default avatarJames Bottomley <James.Bottomley@SteelEye.com>
Signed-off-by: default avatarChris Wright <chrisw@osdl.org>
      906aa4d2
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7