* Re: bkl: Pushdowns for sound/oss ?
2010-05-21 15:34 ` Maciej W. Rozycki
@ 2010-05-21 15:46 ` Ralf Baechle
2010-05-22 21:26 ` Maciej W. Rozycki
0 siblings, 1 reply; 6+ messages in thread
From: Ralf Baechle @ 2010-05-21 15:46 UTC (permalink / raw)
To: Maciej W. Rozycki
Cc: John Kacur, linux-kernel, Frederic Weisbecker, Arnd Bergmann,
Thomas Gleixner, Ingo Molnar, linux-mips, Paul Mundt, linux-sh,
Geert Uytterhoeven, Roman Zippel, linux-m68k
On Fri, May 21, 2010 at 04:34:54PM +0100, Maciej W. Rozycki wrote:
> > > particular in:
> > > sound/oss/swarm_cs4297a.c
> >
> > This one is specific to the Swarm, a MIPS-based platform indeed; I'll cc
> > Maciej Rozycki who most likely is the only person on the planet knowing the
> > technical details. I don't even recall touching that file so my (C)
> > header in there is a surprise :)
>
> You probably added a missing header inclusion or suchlike. ;)
That may well be true - I haven't touched much sound code since my time
at Waldorf Electronics but that was a slightly different calibre than
sound in Linux.
> That's a Crystal Sound CS4297A AC'97 codec wired to a synchronous serial
> interface of the SWARM board. It used to work with 2.4 after some tweaks
> I did back then (it broke in the little-endian mode or something like
> that), but I can't say anything about 2.6. I think the driver should be
> dropped and the serial port in the sound mode (there's a demux to switch
> the interface's external connection between the codec and a DE-9
> connector; the serial port supports asynchronous mode as well) properly
> abstracted as a "sound card".
>
> There's a separate CS4297A driver already in our tree, so it should be
> used in place of the codec bits from this driver (which I believe were
> simply copied over at some point). The rest is glue logic to set up
> serial line parameters correctly for the codec and switch the demux to the
> codec (no proper resource management is done for that though; the
> selection used to be made at the kernel build time). This glue logic is
> all that's needed to be carried over to the new "sound card" driver.
>
> I have plans to do so in some indefinite future, probably when I retire
> and my grandchildren have grown up; anyone please feel free to take it
> first. ;)
I've patched up what there to build without warnings but that doesn't
make any less ugly or wrong. I'm trying to motivate a few folks to
offload the work from you; until that has happened it might be nice if this
driver would stick around as some form of documentation.
As for the grandchildren - is there something I haven't heared of yet? :)
Ralf
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
diff --git a/sound/oss/swarm_cs4297a.c b/sound/oss/swarm_cs4297a.c
index 3136c88..17ae26b 100644
--- a/sound/oss/swarm_cs4297a.c
+++ b/sound/oss/swarm_cs4297a.c
@@ -78,6 +78,7 @@
#include <linux/mutex.h>
#include <linux/kernel.h>
+#include <asm/addrspace.h>
#include <asm/byteorder.h>
#include <asm/dma.h>
#include <asm/io.h>
@@ -160,10 +161,10 @@ module_param(cs_debugmask, int, 0);
#define CS_TYPE_ADC 0
#define CS_TYPE_DAC 1
-#define SER_BASE (A_SER_BASE_1 + KSEG1)
-#define SS_CSR(t) (SER_BASE+t)
-#define SS_TXTBL(t) (SER_BASE+R_SER_TX_TABLE_BASE+(t*8))
-#define SS_RXTBL(t) (SER_BASE+R_SER_RX_TABLE_BASE+(t*8))
+#define SER_BASE CKSEG1ADDR(A_SER_BASE_1)
+#define SS_CSR(t) ((void *) SER_BASE + (t))
+#define SS_TXTBL(t) ((void *) (SER_BASE + R_SER_TX_TABLE_BASE + (t * 8)))
+#define SS_RXTBL(t) ((void *) (SER_BASE + R_SER_RX_TABLE_BASE + (t * 8)))
#define FRAME_BYTES 32
#define FRAME_SAMPLE_BYTES 4
@@ -1179,7 +1180,7 @@ static int mixer_ioctl(struct cs4297a_state *s, unsigned int cmd,
#if CSDEBUG
cs_printioctl(cmd);
#endif
-#if CSDEBUG_INTERFACE
+#ifdef CSDEBUG_INTERFACE
if ((cmd = SOUND_MIXER_CS_GETDBGMASK) ||
(cmd = SOUND_MIXER_CS_SETDBGMASK) ||
@@ -1859,7 +1860,7 @@ static ssize_t cs4297a_write(struct file *file, const char *buffer,
"cs4297a: copy in %d to swptr %x\n", cnt, swptr));
swptr = (swptr + (cnt/FRAME_SAMPLE_BYTES)) % d->ringsz;
- __raw_writeq(cnt/FRAME_SAMPLE_BYTES, SS_CSR(R_SER_DMA_DSCR_COUNT_TX));
+ __raw_writeq(cnt/FRAME_SAMPLE_BYTES, (void *) SS_CSR(R_SER_DMA_DSCR_COUNT_TX));
spin_lock_irqsave(&s->lock, flags);
d->swptr = swptr;
d->count += cnt;
@@ -2502,7 +2503,7 @@ static const struct file_operations cs4297a_audio_fops = {
.release = cs4297a_release,
};
-static void cs4297a_interrupt(int irq, void *dev_id)
+static irqreturn_t cs4297a_interrupt(int irq, void *dev_id)
{
struct cs4297a_state *s = (struct cs4297a_state *) dev_id;
u32 status;
@@ -2517,14 +2518,14 @@ static void cs4297a_interrupt(int irq, void *dev_id)
if (!(status & (M_SYNCSER_RX_EOP_COUNT | M_SYNCSER_RX_OVERRUN | M_SYNCSER_RX_SYNC_ERR))) {
status = __raw_readq(SS_CSR(R_SER_STATUS));
printk(KERN_ERR "cs4297a: unexpected interrupt (status %08x)\n", status);
- return;
+ return IRQ_HANDLED;
}
#endif
if (status & M_SYNCSER_RX_SYNC_ERR) {
status = __raw_readq(SS_CSR(R_SER_STATUS));
printk(KERN_ERR "cs4297a: rx sync error (status %08x)\n", status);
- return;
+ return IRQ_HANDLED;
}
if (status & M_SYNCSER_RX_OVERRUN) {
@@ -2534,9 +2535,9 @@ static void cs4297a_interrupt(int irq, void *dev_id)
/* Fix things up: get the receive descriptor pool
clean and give them back to the hardware */
- while (__raw_readq(SS_CSR(R_SER_DMA_DSCR_COUNT_RX)))
+ while (__raw_readq((void *)SS_CSR(R_SER_DMA_DSCR_COUNT_RX)))
;
- newptr = (unsigned) (((__raw_readq(SS_CSR(R_SER_DMA_CUR_DSCR_ADDR_RX)) & M_DMA_CURDSCR_ADDR) -
+ newptr = (unsigned) (((__raw_readq((void *)SS_CSR(R_SER_DMA_CUR_DSCR_ADDR_RX)) & M_DMA_CURDSCR_ADDR) -
s->dma_adc.descrtab_phys) / sizeof(serdma_descr_t));
for (i=0; i<DMA_DESCR; i++) {
s->dma_adc.descrtab[i].descr_a &= ~M_DMA_SERRX_SOP;
@@ -2544,7 +2545,7 @@ static void cs4297a_interrupt(int irq, void *dev_id)
s->dma_adc.swptr = s->dma_adc.hwptr = newptr;
s->dma_adc.count = 0;
s->dma_adc.sb_swptr = s->dma_adc.sb_hwptr = s->dma_adc.sample_buf;
- __raw_writeq(DMA_DESCR, SS_CSR(R_SER_DMA_DSCR_COUNT_RX));
+ __raw_writeq(DMA_DESCR, (void *)SS_CSR(R_SER_DMA_DSCR_COUNT_RX));
}
spin_lock(&s->lock);
@@ -2553,6 +2554,8 @@ static void cs4297a_interrupt(int irq, void *dev_id)
CS_DBGOUT(CS_INTERRUPT, 6, printk(KERN_INFO
"cs4297a: cs4297a_interrupt()-\n"));
+
+ return IRQ_HANDLED;
}
#if 0
@@ -2588,14 +2591,15 @@ static int __init cs4297a_init(void)
"cs4297a: cs4297a_init_module()+ \n"));
#ifndef CONFIG_BCM_CS4297A_CSWARM
- mdio_val = __raw_readq(KSEG1 + A_MAC_REGISTER(2, R_MAC_MDIO)) &
+ mdio_val = __raw_readq((void *)CKSEG1ADDR(A_MAC_REGISTER(2, R_MAC_MDIO))) &
(M_MAC_MDIO_DIR|M_MAC_MDIO_OUT);
/* Check syscfg for synchronous serial on port 1 */
- cfg = __raw_readq(KSEG1 + A_SCD_SYSTEM_CFG);
+ cfg = __raw_readq((void *)CKSEG1ADDR(A_SCD_SYSTEM_CFG));
if (!(cfg & M_SYS_SER1_ENABLE)) {
- __raw_writeq(cfg | M_SYS_SER1_ENABLE, KSEG1+A_SCD_SYSTEM_CFG);
- cfg = __raw_readq(KSEG1 + A_SCD_SYSTEM_CFG);
+ __raw_writeq(cfg | M_SYS_SER1_ENABLE,
+ (void *)CKSEG1ADDR(A_SCD_SYSTEM_CFG));
+ cfg = __raw_readq((void *)CKSEG1ADDR(A_SCD_SYSTEM_CFG));
if (!(cfg & M_SYS_SER1_ENABLE)) {
printk(KERN_INFO "cs4297a: serial port 1 not configured for synchronous operation\n");
return -1;
@@ -2605,12 +2609,14 @@ static int __init cs4297a_init(void)
/* Force the codec (on SWARM) to reset by clearing
GENO, preserving MDIO (no effect on CSWARM) */
- __raw_writeq(mdio_val, KSEG1+A_MAC_REGISTER(2, R_MAC_MDIO));
+ __raw_writeq(mdio_val,
+ (void *)CKSEG1ADDR(A_MAC_REGISTER(2, R_MAC_MDIO)));
udelay(10);
}
/* Now set GENO */
- __raw_writeq(mdio_val | M_MAC_GENC, KSEG1+A_MAC_REGISTER(2, R_MAC_MDIO));
+ __raw_writeq(mdio_val | M_MAC_GENC,
+ (void *)CKSEG1ADDR(A_MAC_REGISTER(2, R_MAC_MDIO)));
/* Give the codec some time to finish resetting (start the bit clock) */
udelay(100);
#endif
@@ -2665,8 +2671,6 @@ static int __init cs4297a_init(void)
} while (!rval && (pwr != 0xf));
if (!rval) {
- char *sb1250_duart_present;
-
fs = get_fs();
set_fs(KERNEL_DS);
#if 0
@@ -2688,9 +2692,20 @@ static int __init cs4297a_init(void)
cs4297a_read_ac97(s, AC97_VENDOR_ID1, &id);
+#if 0
+ /*
+ * Since the rewrite of the SB1250 UART driver for the
+ * drivers/serial subsystem there is no more
+ * sb1250_duart_present symbol in the kernel so this resource
+ * sharing mechanism which arguably has always been fishy
+ * does no longer work.
+ */
+ char *sb1250_duart_present_p;
+
sb1250_duart_present = symbol_get(sb1250_duart_present);
if (sb1250_duart_present)
sb1250_duart_present[1] = 0;
+#endif
printk(KERN_INFO "cs4297a: initialized (vendor id = %x)\n", id);
@@ -2732,6 +2747,7 @@ static void __exit cs4297a_cleanup(void)
MODULE_AUTHOR("Kip Walker, Broadcom Corp.");
MODULE_DESCRIPTION("Cirrus Logic CS4297a Driver for Broadcom SWARM board");
+MODULE_LICENSE("GPL");
// ---------------------------------------------------------------------
^ permalink raw reply related [flat|nested] 6+ messages in thread