New FW has chunks that are larger than the size limit of the
FH's DMA. To make sure we don't crash it - actively limit the
max size of each chunk.
Signed-off-by: Liad Kaufman <liad.kaufman@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
When we reset the device, the CSR_INT gets cleared as well
as CSR_INT_MASK. Meaning that we shouldn't get any interrupt
but, due to a hardware bug, recent devices will keep sending
interrupts. This leads to an interrupt storm while stopping
the device.
The way to fix this is to ACK all the interrupts after the
device is reset so that the value of CSR_INT will stay
0xffffffff.
Fixes: 522713c81e ("iwlwifi: pcie: properly reset the device")
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Until this patch, dumping the monitor data could be done only
for PCIe external (DRAM) mode in 7000 HW family. This patch
allows to pull the monitor data also on other families, and
also to pull the monitor data if an internal buffer is used.
Signed-off-by: Liad Kaufman <liad.kaufman@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Sometimes there is a need to configure some registers for
setting some FW properties, such as the FW monitor mode
(internal/external). This patch supports setting this for
PCIe mode.
Signed-off-by: Liad Kaufman <liad.kaufman@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Adds all FH registers between FH_MEM_UPPER_BOUND and
FH_MEM_LOWER_BOUND (which should be readable to the driver)
to the dump data when it is collected.
Signed-off-by: Liad Kaufman <liad.kaufman@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Not doing so would allow other possible users of the device
to take ownership and prevent normal WiFi operation.
This fixes the second part of:
https://bugzilla.kernel.org/show_bug.cgi?id=87191
Reviewed-by: Moshe Harel <moshe.harel@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Toggle the LMPM_CHICK register when writing chunks into the FW's extended
SRAM. This tells the FW to put the chunk into a different memory space.
Signed-off-by: Arik Nemtsov <arikx.nemtsov@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
We were toggling the wrong bit when we reset the device,
fix that. Moreover, since the reset can take time, we need
to wait before we set the rfkill interrupt. Not doing so
can be racy since the driver is enabling the rfkill
interrupt while the device is resetting which will clear
all the registers including the CSR_INT_MASK.
This can basically lead to a situation where we don't
enable the rfkill interrupt. If that happens, the user will
not be able to re-enable the device when de-asserting
rfkill.
This scenario happened to the submitter of:
https://bugzilla.kernel.org/show_bug.cgi?id=87191
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
In some rare cases, the firmware can put the device to
sleep after the driver requested the access. This is
because the access request can take a short time to be
propagated to the firmware.
If that happens, the driver may think that it has access
since the firmware hasn't put the device to sleep yet, but
right after the driver's check, the firmware might put the
device to sleep.
Warn when this happens by allowing the firmware to finish
the "put the device sleep" flow so that the driver will
not get access to the device. This will make the issue
visible.
This still doesn't fix the race, but at least it makes
it more visible.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
The length counting previously done had an error in it, causing
the length down the data dumping function to be shorter than it
should be, causing the end of the data to get truncated off and
lost.
Cc: <stable@vger.kernel.org> [3.17+]
Fixes: 67c65f2cf7 ("iwlwifi: dump periphery registers to fw-error-dump")
Signed-off-by: Liad Kaufman <liad.kaufman@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
If the RFkill interrupt fires while we calibrate, it would
make the firmware fail and the driver wasn't able to recover.
Change the flow so that the driver will kill the firmware
in that case.
Since we have now two flows that are calling
trans_stop_device (the RFkill interrupt and the
op_mode_mvm_start function) - we need to better sync this.
Use the STATUS_DEVICE_ENABLED in the pcie transport in an
atomic way to achieve this.
This fixes: https://bugzilla.kernel.org/show_bug.cgi?id=86231
CC: <stable@vger.kernel.org> [3.10+]
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Reviewed-by: Luciano Coelho <luciano.coelho@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
iwlwifi features a debug mechanism that allows to dump
binary data which is helpful to debug the firmware.
Until now, this data was made available for the userspace
through debugfs. For this exact purpose, devcoredump was
created. Move to the new infrastructure.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
In the new format the "CSS section" has the same TLV type
as the "mem section". So we need to run the secured flow
for all the 8000 products.
Signed-off-by: Eran Harary <eran.harary@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
When the ARC is reset when we exit from Sx in case we had
WoWLAN running, we can't access the prph before we reset
the NIC.
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
iwl_poll_bit may return a strictly positive value when the
poll doesn't match on the first try.
This was caught when WoWLAN started failing upon resume
even if the poll_bit actually succeeded.
Also change a wrong print. If we reach the end of
iwl_pcie_prepare_card_hw, it means that we couldn't
get the devices.
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Reviewed-by: Luciano Coelho <luciano.coelho@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
The LTR is the handshake between the device and the root
complex about the latency allowed when the bus exits power
save. This configuration was missing and this led to high
latency in the link power up. The end user could experience
high latency in the network because of this.
Cc: <stable@vger.kernel.org> [3.10+]
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Align the trans->hw_rev variable format with previous series
format.
Signed-off-by: Liad Kaufman <liad.kaufman@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Rather than ANDing with a mask - use existing macros, which
are more readable.
Signed-off-by: Liad Kaufman <liad.kaufman@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
This configuration is not needed for dvm, and it actually
broke it.
Reported-by: Oliver Hartkopp <socketcan@hartkopp.net>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Our legal structure changed at some point (see wikipedia), but
we forgot to immediately switch over to the new copyright
notice.
For files that we have modified in the time since the change,
add the proper copyright notice now.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Add the Control Status Registers to the firmware error dump
infrastructure.
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Use the fw-error-dump infrastructure to dump the periphery
registers. Only certain ranges are readable, so dump only
these.
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
The mvm op_mode won't allocate the buffer for the transport
any more. The transport allocates its own buffer and mvm
is in charge of splicing the buffers in the debugfs hook.
This makes the repartition easier to handle.
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
The format of the CSR_HW_REV register has changed in 8000
HW family. To keep backwards compatibility, we store the
value of this register as usual in trans->hw_rev, only we
store it in the old format in this variable.
Signed-off-by: Liad Kaufman <liad.kaufman@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
This allows to use the firmware monitor. This capability
uses a lot of contiguous memory (up to 64MB), so make its
usage module parameter dependent.
The driver will try to allocate as much contiguous memory
as possible downgrading its requirements until the
allocation succeeds.
Dump this data into the fw-error dump file when an error
happens.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Some races with the hardware can happen when we take
ownership of the device. Don't give up after the first try.
Cc: <stable@vger.kernel.org>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
In case RFKILL is in KILL position, the NIC will issue an
interrupt straight away. This interrupt won't be sent
because it is masked in the hardware.
But if our interrupt service routine is called for another
reason (SHARED_IRQ), then we'll look at the interrupt cause
and service it. This can cause bad things if we are not
ready yet.
Explicitly clean the interrupt cause register to make sure
we won't service anything before we are ready to.
Cc: <stable@vger.kernel.org> [3.14]
Reported-and-tested-by: Alexander Monakov <amonakov@gmail.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
When a firmware error occurs, capture the last 32 commands
(which are still in memory) in the error dump debugfs file.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
This variable always tracks a constant value (256) so there's
no need to have it. Removing it simplifies code generation,
reducing the .text size (by about 240 bytes on x86-64.)
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
While the scan_cmd should really be allocated in init (and
we do fail init in case the allocation failed), it doesn't
mean we should lock up the machine if something really bad
happened.
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
This must not happen - otherwise we might keep flushing
forever.
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
This will be used later to flush / wait for queues that are
related to a specific vif.
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
This define is related to the firmware packaging and is
needed by more than one transport.
Signed-off-by: Eran Harary <eran.harary@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Use the new NAPI infrastructure added to mac80211 to get
GRO. We don't really implement NAPI since we don't have
a real poll function and we never schedule a NAPI poll.
Instead of this, we collect all the packets we got from a
single interrupt and then call napi_gro_flush().
This allows us to benefit from GRO. In half duplex medium
like WiFi, its main advantage is that it reduces the number
of TCP Acks, hence improving the TCP Rx performance.
Since we call the Rx path with a spinlock held, remove
the might_sleep mention from the op_mode's API.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Reviewed-by: Ido Yariv <ido@wizery.com>
[Squash different patches and rewrite the commit message]
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
When indicating RF-kill toggle to the higher layer, that
may in turn call back to the transport (for MVM at least)
to turn off the device quickly. Instead of that, allow it
to return whether or not the device should be turned off,
this gets rid of the call indirection and will help make
the API more consistent when we go back to non-threaded
interrupts again for PCIe.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
1. Enable LP XTAL to avoid HW bug where device may consume much
power if FW is not loaded after device reset. LP XTAL is
disabled by default after device HW reset. Configure device's
"persistence" mode to avoid resetting XTAL again when SHRD_HW_RST
occurs in S3.
2. Add methods to access SHR (shared block memory space) directly from PCI
bus w/o need to power up MAC HW.
Shared internal registers (e.g. SHR_APMG_GP1, SHR_APMG_XTAL_CFG)can be
accessed directly from PCI bus through SHR arbiter even when MAC HW is
powered down. This is possible due to indirect read/write via
HEEP_CTRL_WRD_PCIEX_CTRL (0xEC) and HEEP_CTRL_WRD_PCIEX_DATA (0xF4)
registers.
Use iwl_write32()/iwl_read32() family to access these registers. The MAC HW
need not be powered up so no "grab inc access" is required.
For example, to read from SHR_APMG_GP1 register (0x1DC),
first, write to the control register:
HEEP_CTRL_WRD_PCIEX_CTRL[15:0] = 0x1DC (offset of the SHR_APMG_GP1 register)
HEEP_CTRL_WRD_PCIEX_CTRL[29:28] = 2 (read access)
second, read from the data register HEEP_CTRL_WRD_PCIEX_DATA[31:0].
To write the register, first, write to the data register
HEEP_CTRL_WRD_PCIEX_DATA[31:0] and then:
HEEP_CTRL_WRD_PCIEX_CTRL[15:0] = 0x1DC (offset of the SHR_APMG_GP1 register)
HEEP_CTRL_WRD_PCIEX_CTRL[29:28] = 3 (write access)
Signed-off-by: Alexander Bondar <alexander.bondar@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Newer devices have two embedded CPUs, and the firwmare for
both of them is include in the .ucode file requested upon
enumeration.
An empty section with address=0xFFFFCCCC separates between
the sections intended for cpu1 and the sections intended
for cpu2.
Update the driver to parse the .ucode file with this format
and act accordingly.
Signed-off-by: Eran Harary <eran.harary@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
The code seems fine, as buf won't be assigned when an error
is returned, but checking for the error first is easier to
understand.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Also handle the bypass mode in which the second CPU doesn't
interfere.
Signed-off-by: Eran Harary <eran.harary@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
This configuration is invalid for this family.
Signed-off-by: Eran Harary <eran.harary@intel.com>
Reviewed-by: Dor Shaish <dor.shaish@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
This register is not present in 8000 family devices.
There is prph register instead.
Signed-off-by: Eran Harary <eran.harary@intel.com>
Reviewed-by: Dor Shaish <dor.shaish@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
APMG HW block was removed in this NIC, hence, no need to
configure it.
Signed-off-by: Eran Harary <eran.harary@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
