The kernel was allowing any component of the pt_regs to be updated either
by signal handlers writing to the stack, or by processes writing via
PTRACE_POKEUSR or PTRACE_SETREGS, which meant they could set their PL
up from 0 to 1 and get access to kernel code and data (or, in practice,
cause a kernel panic). We now always reset the ex1 field, allowing the
user to set their ICS bit only.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
This change is modelled on similar fixes for other architectures.
The pt_regs "faultnum" member is set to the trap (fault) number that
caused us to enter the kernel, and is INT_SWINT_1 for the syscall software
interrupt. We already supported a pseudo value, INT_SWINT_1_SIGRETURN,
that we used for the rt_sigreturn syscall; it avoided the case where
one signal was handled, then we "tail-called" to another handler.
This change avoids the similar case where we start to call one handler,
then are preempted into another handler when we start trying to run
the first handler. We clear ->faultnum after calling handle_signal(),
and to be paranoid also in the case where there was no signal to deliver.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
For the "initfree" boot argument it's not that big a deal, but
to avoid warnings in the code, we check for a valid value before
allowing the specified argument to override the kernel default.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
This completes the tile migration to the new naming scheme for
the architecture-specific irq management code.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
This change makes KM_TYPE_NR independent of the actual deprecated
list of km_type values, which are no longer used in tile code anywhere.
For now we leave it set to 8, allowing that many nested mappings,
and thus reserving 32MB of address space.
A few remaining places using KM_* values were cleaned up as well.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
According to the linux/clk.h definition we should be handing back an
errno value or a valid rate. This fixes up the case where 0 can be
returned for invalid frequencies or cases where rounding has no
selectable candidate.
Reported-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
Change the regsister/unregister routines to generate better
messages, and control arrival of new frames when USB device is
unplugged.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Serious bug in original code, if app reads 10 bytes but 20 byte msg
received memory would get overwritten.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Use standard network statistics variables and routines.
Transmit counters are per queue, and skb mapping is already in
skb and does not need to be recomputed. Move SearchVcId to only
place it is used.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Since classification and queue control are separate, allow
classification even if device is down now; this avoids races
on startup/shutdown.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Use namespace tool from kernel scripts to identify dead code and
functions that should be static.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Crash is triggered by commit e6484930d7 ("net: allocate tx queues in
register_netdevice"), which moved tx netqueue creation into register_netdev.
So now calling netif_stop_queue() before register_netdev causes an oops.
Move netif_stop_queue() after net device registration to fix crash.
Signed-off-by: Dmitry Artamonow <mad_soft@inbox.ru>
Signed-off-by: Denis Kirjanov <dkirjanov@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eliminate global variable in transmit path
The Leader can be on the stack, and get rid of unnecessary timeval.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Remove dead fields, change fields that only have true/false to boolean;
and rearrange to save space.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Touching the queue state before register_netdev is not
allowed, and besides the queue state before ->open()
is "don't care"
Reported-by: Josh Boyer <jwboyer@gmail.com>
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Many of the config bit are presently duplicated between the platforms,
which will gradually cleaned up through centralization. For the moment we
expose some new INTC features through drivers/sh/Kconfig that the ARM
platforms presently don't enable, so make it generally available.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
'sparse' spotted that the parameters to kzalloc in l2tp_dfs_seq_open
were swapped.
Tested on current git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
at 1792f17b72 build, boots and I can see that directory,
but there again I could see /sys/kernel/debug/l2tp with it swapped; I don't have
any l2tp in use.
Signed-off-by: Dr. David Alan Gilbert <linux@treblig.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Sparse caught several places where ioctl interface was incorrectly using user
memory. Fix all the ioctl cases for casting and __user annotation.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
This preserves the semantics of the original driver (unclassified packets
are dropped), but does it in a clean way; and fixes crash when packet
is sent to offline device.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Current kernels have multi-queue support which can be used by
this device. This has the advantage that a single type of traffic
will not block other types.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
