From the Windows INF file, we know the firmware ranges for all RME
cards. For PCIe, a single revision ID per device (RayDAT, MADI, AIO,
AES) is used. Contrary, the older PCI versions use ranges, that is,
one revision ID per firmware version.
Instead of listing all possible revisions individually, match the range.
This commit enables all MADI and AES PCI versions ever shipped.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
HDSP_VERSION_BIT has to be ORed with HDSP_S_LOAD. This fixes the detection
of at least some RME RPM boxes.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
SNDRV_HDSPM_IOCTL_GET_STATUS is supposed to query the current card
status, so we have to return what we receive on the MADI wire (RX), not
what we transmit (TX) to others. The latter is a config item to be
queried via SNDRV_HDSPM_IOCTL_GET_CONFIG.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Smatch has a new check for Rosenberg type information leaks where
structs are copied to the user with uninitialized stack data in them.
The status struct has a hole in it, and on some paths not all the
members were initialized.
struct hdspm_status {
unsigned char card_type; /* 0 1 */
/* XXX 3 bytes hole, try to pack */
enum hdspm_syncsource autosync_source; /* 4 4 */
long long unsigned int card_clock; /* 8 8 */
The hdspm_version struct had holes in it as well.
struct hdspm_version {
unsigned char card_type; /* 0 1 */
char cardname[20]; /* 1 20 */
/* XXX 3 bytes hole, try to pack */
unsigned int serial; /* 24 4 */
short unsigned int firmware_rev; /* 28 2 */
/* XXX 2 bytes hole, try to pack */
int addons; /* 32 4 */
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Refactoring the code using snd_pcm_hw_constraint_pow2() helper function.
Tested-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
AES32 supports the non-standard 128kHZ, and this is enabled only when
SNDRV_PCM_RATE_KNOT is set in hw.rates field.
Tested-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Newer RME cards like RayDAT and AIO support 32 samples per period. This
value is encoded as {1,1,1} in the HDSP_LatencyMask bits in the control
register.
Since {1,1,1} is also the representation for 8192 samples/period on
older RME cards, we have to special case 32 samples and 32768 bytes
according to the actual card.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Currently, hdspm_decode_latency is called several times, violating the
DRY principle. Given that we need to distinguish between old and new
cards when decoding the latency bits in the control register, introduce
hdspm_get_latency() to provide the required functionality.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
On newer RME cards like RayDAT and AIO, the 8192 samples per period size
are no longer supported. Instead, setting all three bits of
HDSP_LatencyMask to one ({1,1,1}) now corresponds to 32 samples per
period.
To make this more obvious to future developers, let's reorder the array
according to their bit representation, starting at 64 ({0,0,0}) up to
4096 ({1,1,0}) and finally 32 ({1,1,1}).
Note that this patch doesn't change semantics.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
On newer RME cards like RayDAT and AIO, the lower bound is 32 samples
per period in contrast to 64 samples as seen on older cards.
We hence lower period_bytes_min to 32 * 4. Four bytes per sample.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Older RME cards like MADI and AES support period sizes of 8192 samples.
The original hdspm driver already featured this value, apparently, it
was lost during the rewrite.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Put the exception checks for io_type switch() for possible mistakes in
future. Also this shuts up annoying compile warnings.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Apparently, there are multiple old firmware revisions in the wild for
the PCI RME MADI cards. Just add them to the list of supported devices
and treat them like their modern counterparts.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In slave mode, the card can only detect the base frequency (32..48kHz)
on the MADI link (exception: 96k frames), so the real external sample
rate is this base frequency multiplied by 1, 2 or 4 depending on the
speed mode.
This patch enables 64..192kHz sample rates in clock slave mode, which
failed before due to an alleged sample rate mismatch between the MADI
link (e.g., 48kHz) and the application in DS/QS mode (e.g., 96kHz,
192kHz).
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
When running in slave mode (no clock master), there is no way to
determine the real wirespeed on the MADI link (single/double/quad
speed). Like physical gear, simply provide the user with a tristate
switch to select the appropriate format.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The char can be unsigned on some architectures. Since the code checks
the negative values, they should be declared as signed char explicitly.
sound/pci/rme9652/hdspm.c:5449: warning: comparison is always false due to limited range of data type
sound/pci/rme9652/hdspm.c:5462: warning: comparison is always false due to limited range of data type
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The PCI version of the RME HDSP MADI card uses 0xcf as revision ID. Just
add this to the list of supported cards.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
When using Word Clock on RME MADI cards, AutoSync mode was alternating
betweeen MADI and WC due to a typo: AutoSync is indicated in the second
status register (status2), not the first one (status).
While the proc output was always correct, the reported WC frequency to
ALSA was unstable as mentioned in
http://mailman.alsa-project.org/pipermail/alsa-devel/2008-March/006723.html
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
For the MIDI part, we need to acquire (and release) the hmidi->lock,
access to the global hdspm structure is serialized through
hmidi->hdspm->lock instead.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The name argument of request_irq() appears in /proc/interrupts, and
it's quite ugly when the name entry contains a space or special letters.
In general, it's simpler and more readable when the module name appears
there, so let's replace all entries with KBUILD_MODNAME.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The convention for pci_driver.name entry in kernel drivers seem to be
the module name or equivalent ones. But, so far, almost all PCI sound
drivers use more verbose name like "ABC Xyz (12)", and these are fairly
confusing when appearing as a file name.
This patch converts the all pci_driver.name entries in sound/pci/* to
use KBUILD_MODNAME for more unified appearance.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The current AES32 firmware revision ID is 234, however, a user confirmed
that everything works fine with the previous revision, too.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Depending on the model and the presence of a TCO module, the number of
midi ports varies. Some have 1 port (MADIface), some have 2 (default),
with TCO, there are 3.
Don't hardcode the number of midi ports to initialize.
This patch also fixes a boot lockup on MADIface.
[Coding-style fixes by tiwai]
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In contrast to the PCIe version (RME AES), the PCI version (RME AES32)
has a different firmware revision.
This patch adds the missing PCI revision.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Original patch by Dan Rosenberg <drosenberg@vsecurity.com> under commit
e68d3b316a. I'm copying his text here:
The SNDRV_HDSPM_IOCTL_GET_CONFIG_INFO ioctl in hdspm.c allow unprivileged
users to read uninitialized kernel stack memory, because several fields
of the hdspm_config struct declared on the stack are not altered
or zeroed before being copied back to the user. This patch takes care
of it.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Fredrik Lingvall <fredrik.lingvall@gmail.com> has discovered wrong
frequency and sync detection on AES32. According to him, the provided
patch fixes these issues.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In contrast to the RME MADI card, coax/optical selection on the MADIface
is done via a physical switch located at the breakout box. Obviously,
the driver cannot switch ports in software.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Caused by two typos, no output channel mappings were assigned for
MADI/MADIface at double/quad speed.
The channel mapping is indeed identical to the single speed mapping, the
cards will simply use the first N channels.
Signed-off-by: Florian Faber <faber@faberman.de>
Signed-off-by: Fredrik Lingvall <fredrik.lingvall@gmail.com>
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Without calling an appropriate rule, AES/AES32 cards would announce a
theoretical channel count of 64 (HDSPM_MAX_CHANNELS), leading to the
already known bug:
[37422.640481] ------------[ cut here ]------------
[37422.640487] WARNING: at sound/pci/rme9652/hdspm.c:5449
snd_hdspm_ioctl+0x18f/0x202 [snd_hdspm]()
[37422.640489] Hardware name: PRIMERGY RX100 S6
[37422.640490] BUG? (info->channel >= hdspm->max_channels_in)
[37422.640492] Modules linked in: snd_hdspm snd_seq_midi ipmi_watchdog
ipmi_poweroff ipmi_si ipmi_devintf ipmi_msghandler i2c_i801 e1000e
snd_rawmidi power_meter [last unloaded: snd_hdspm]
[37422.640501] Pid: 22231, comm: jackd Tainted: G D W
2.6.36-gentoo-r5 #5
[37422.640502] Call Trace:
[37422.640508] [<ffffffff8103db3a>] warn_slowpath_common+0x80/0x98
[37422.640511] [<ffffffff8103dbe6>] warn_slowpath_fmt+0x41/0x43
[37422.640514] [<ffffffff81034306>] ? get_parent_ip+0x11/0x42
[37422.640518] [<ffffffffa0055763>] snd_hdspm_ioctl+0x18f/0x202
[snd_hdspm]
[37422.640522] [<ffffffff813fd626>] snd_pcm_channel_info+0x73/0x7c
[37422.640525] [<ffffffff814001e9>] snd_pcm_common_ioctl1+0x326/0xb01
[37422.640527] [<ffffffff81034306>] ? get_parent_ip+0x11/0x42
[37422.640531] [<ffffffff8105be6c>] ? __srcu_read_unlock+0x3b/0x59
[37422.640533] [<ffffffff81400bce>] snd_pcm_capture_ioctl1+0x20a/0x227
[37422.640537] [<ffffffff811e599c>] ? file_has_perm+0x90/0x9e
[37422.640540] [<ffffffff81400c15>] snd_pcm_capture_ioctl+0x2a/0x2e
[37422.640543] [<ffffffff810f2c69>] do_vfs_ioctl+0x404/0x453
[37422.640546] [<ffffffff810f2d09>] sys_ioctl+0x51/0x74
[37422.640549] [<ffffffff81002aab>] system_call_fastpath+0x16/0x1b
[37422.640552] ---[ end trace 0cd919cd68118082 ]---
We already have all the right values in place, we simply have to inform
the upper layers about this restriction.
Note that snd_hdspm_hw_rule_rate_out_channels and
snd_hdspm_hw_rule_rate_in_channels must not be called on AES32, because
the channel count is always 16, no matter of the samplerate in use.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Only RayDAT and AIO provide sane buffer pointers that can be used with
HDSPM_BufferPositionMask, on all other cards, this would result in a
wrong HW pointer leading to xruns and these messages:
[260808.916788] BUG: pcmC0D0p:0, pos = 2976, buffer size = 1024, period size = 512
[260808.961124] BUG: pcmC0D0c:0, pos = 4944, buffer size = 1024, period size = 512
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
As requested by Takashi and Jaroslav, these arrays should not be in the
header file.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Incorporate changes by Florian Faber into hdspm.c. Code taken from
http://wiki.linuxproaudio.org/index.php/Driver:hdspe
Heavily reworked to mostly comply with the coding standard (whitespace
fixes, line width, C++ style comments)
The code was tested and confirmed to be working on RME RayDAT.
Signed-off-by: Adrian Knoth <adi@drcomp.erfurt.thur.de>
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Add support for the RME HDSP RPM IO box. Changes have been made in the identification of the IO box and the neccessary controls have been added.
Signed-off-by: Florian Faber <faberman@linuxproaudio.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The SNDRV_HDSP_IOCTL_GET_CONFIG_INFO and
SNDRV_HDSP_IOCTL_GET_CONFIG_INFO ioctls in hdspm.c and hdsp.c allow
unprivileged users to read uninitialized kernel stack memory, because
several fields of the hdsp{m}_config_info structs declared on the stack
are not altered or zeroed before being copied back to the user. This
patch takes care of it.
Signed-off-by: Dan Rosenberg <dan.j.rosenberg@gmail.com>
Cc: <stable@kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
