Hi Manu, Joel, I have exactly the same problem with my TwinHan VP2031 as Joel (Fedora Core 4, kernel 2.6.14). Tuning and zapping now work, I find all channels (thanks for the czap -r switch info!), but signal is garbled (loads of green squares, very occasionally a second or so with some recognizable images in between the green). I use dvb-kernel modules compiled from fresh cvs source code. Have been struggling with this for weeks now, and am happy to find a few shoulders to cry on ;-) Just kidding, this email list is great and you all are very helpful. I mailed Joel directly yesterday, and he forwarded me the TarBall that Manu Abrahams sended him today (cgi-bin-hgwebdir-cgi-v4l-algo-test-675755d11eb2.tar.bz2). I am also living in the Netherlands, and even use the same cable provider as Joel (Casema). I compiled the tarbal, which works fine, but when first removing and then modprobing the modules I get loads of error messages (see below, also dmesg). Joel just emailed me he got the same (lack of) results. Do we compile against the wrong kernel? Any ideas? I'd love to start using this card, and hate having to capitulate and install Windoze... Thanks for looking into this! Cheers, Bas --- error messages when inserting modules: [root@mediapc cgi-bin-hgwebdir-cgi-v4l-algo-test-675755d11eb2]# rmmod bttv ERROR: Module bttv is in use by bt878 [root@mediapc cgi-bin-hgwebdir-cgi-v4l-algo-test-675755d11eb2]# rmmod bt878 ERROR: Module bt878 is in use by dst [root@mediapc cgi-bin-hgwebdir-cgi-v4l-algo-test-675755d11eb2]# rmmod dst ERROR: Module dst is in use by dst_ca [root@mediapc cgi-bin-hgwebdir-cgi-v4l-algo-test-675755d11eb2]# rmmod dst_ca [root@mediapc cgi-bin-hgwebdir-cgi-v4l-algo-test-675755d11eb2]# rmmod dst [root@mediapc cgi-bin-hgwebdir-cgi-v4l-algo-test-675755d11eb2]# rmmod bt878 [root@mediapc cgi-bin-hgwebdir-cgi-v4l-algo-test-675755d11eb2]# rmmod bttv [root@mediapc cgi-bin-hgwebdir-cgi-v4l-algo-test-675755d11eb2]# modprobe bttv FATAL: Error inserting bttv (/lib/modules/2.6.14-1.1656_FC4.netdev.7/kernel/drivers/media/video/bttv.ko): Unknown symbol in module, or unknown parameter (see dmesg) FATAL: Error running install command for bttv [root@mediapc cgi-bin-hgwebdir-cgi-v4l-algo-test-675755d11eb2]# modprobe bt878 FATAL: Error inserting bttv (/lib/modules/2.6.14-1.1656_FC4.netdev.7/kernel/drivers/media/video/bttv.ko): Unknown symbol in module, or unknown parameter (see dmesg) WARNING: Error running install command for bttv FATAL: Error inserting bt878 (/lib/modules/2.6.14-1.1656_FC4.netdev.7/kernel/drivers/media/dvb/bt8xx/bt878.ko): Unknown symbol in module, or unknown parameter (see dmesg) [root@mediapc cgi-bin-hgwebdir-cgi-v4l-algo-test-675755d11eb2]# modprobe dst FATAL: Error inserting bttv (/lib/modules/2.6.14-1.1656_FC4.netdev.7/kernel/drivers/media/video/bttv.ko): Unknown symbol in module, or unknown parameter (see dmesg) WARNING: Error running install command for bttv WARNING: Error inserting bt878 (/lib/modules/2.6.14-1.1656_FC4.netdev.7/kernel/drivers/media/dvb/bt8xx/bt878.ko): Unknown symbol in module, or unknown parameter (see dmesg) FATAL: Error inserting dst (/lib/modules/2.6.14-1.1656_FC4.netdev.7/kernel/drivers/media/dvb/bt8xx/dst.ko): Unknown symbol in module, or unknown parameter (see dmesg) [root@mediapc cgi-bin-hgwebdir-cgi-v4l-algo-test-675755d11eb2]# modprobe bt878 FATAL: Error inserting bttv (/lib/modules/2.6.14-1.1656_FC4.netdev.7/kernel/drivers/media/video/bttv.ko): Unknown symbol in module, or unknown parameter (see dmesg) WARNING: Error running install command for bttv FATAL: Error inserting bt878 (/lib/modules/2.6.14-1.1656_FC4.netdev.7/kernel/drivers/media/dvb/bt8xx/bt878.ko): Unknown symbol in module, or unknown parameter (see dmesg) [root@mediapc cgi-bin-hgwebdir-cgi-v4l-algo-test-675755d11eb2]# rmmod dvb_bt8xx ERROR: Module dvb_bt8xx does not exist in /proc/modules [root@mediapc cgi-bin-hgwebdir-cgi-v4l-algo-test-675755d11eb2]# modprobe dvb_bt8xx FATAL: Error inserting bttv (/lib/modules/2.6.14-1.1656_FC4.netdev.7/kernel/drivers/media/video/bttv.ko): Unknown symbol in module, or unknown parameter (see dmesg) WARNING: Error running install command for bttv WARNING: Error inserting bt878 (/lib/modules/2.6.14-1.1656_FC4.netdev.7/kernel/drivers/media/dvb/bt8xx/bt878.ko): Unknown symbol in module, or unknown parameter (see dmesg) WARNING: Error inserting dst (/lib/modules/2.6.14-1.1656_FC4.netdev.7/kernel/drivers/media/dvb/bt8xx/dst.ko): Unknown symbol in module, or unknown parameter (see dmesg) WARNING: Error inserting dst_ca (/lib/modules/2.6.14-1.1656_FC4.netdev.7/kernel/drivers/media/dvb/bt8xx/dst_ca.ko): Unknown symbol in module, or unknown parameter (see dmesg) FATAL: Error inserting dvb_bt8xx (/lib/modules/2.6.14-1.1656_FC4.netdev.7/kernel/drivers/media/dvb/bt8xx/dvb-bt8xx.ko): Unknown symbol in module, or unknown parameter (see dmesg) [root@mediapc cgi-bin-hgwebdir-cgi-v4l-algo-test-675755d11eb2]# The relevant parts of my dmesg (before modules loaded fine, card was detected): bttv: disagrees about version of symbol btcx_riscmem_alloc bttv: Unknown symbol btcx_riscmem_alloc bttv: disagrees about version of symbol btcx_riscmem_free bttv: Unknown symbol btcx_riscmem_free bttv: disagrees about version of symbol tveeprom_hauppauge_analog bttv: Unknown symbol tveeprom_hauppauge_analog bttv: disagrees about version of symbol btcx_riscmem_alloc bttv: Unknown symbol btcx_riscmem_alloc bttv: disagrees about version of symbol btcx_riscmem_free bttv: Unknown symbol btcx_riscmem_free bttv: disagrees about version of symbol tveeprom_hauppauge_analog bttv: Unknown symbol tveeprom_hauppauge_analog bt878: Unknown symbol bttv_read_gpio bt878: Unknown symbol bttv_write_gpio bt878: Unknown symbol bttv_gpio_enable bttv: disagrees about version of symbol btcx_riscmem_alloc bttv: Unknown symbol btcx_riscmem_alloc bttv: disagrees about version of symbol btcx_riscmem_free bttv: Unknown symbol btcx_riscmem_free bttv: disagrees about version of symbol tveeprom_hauppauge_analog bttv: Unknown symbol tveeprom_hauppauge_analog bt878: Unknown symbol bttv_read_gpio bt878: Unknown symbol bttv_write_gpio bt878: Unknown symbol bttv_gpio_enable dst: Unknown symbol bt878_device_control bttv: disagrees about version of symbol btcx_riscmem_alloc bttv: Unknown symbol btcx_riscmem_alloc bttv: disagrees about version of symbol btcx_riscmem_free bttv: Unknown symbol btcx_riscmem_free bttv: disagrees about version of symbol tveeprom_hauppauge_analog bttv: Unknown symbol tveeprom_hauppauge_analog bt878: Unknown symbol bttv_read_gpio bt878: Unknown symbol bttv_write_gpio bt878: Unknown symbol bttv_gpio_enable bttv: disagrees about version of symbol btcx_riscmem_alloc bttv: Unknown symbol btcx_riscmem_alloc bttv: disagrees about version of symbol btcx_riscmem_free bttv: Unknown symbol btcx_riscmem_free bttv: disagrees about version of symbol tveeprom_hauppauge_analog bttv: Unknown symbol tveeprom_hauppauge_analog bt878: Unknown symbol bttv_read_gpio bt878: Unknown symbol bttv_write_gpio bt878: Unknown symbol bttv_gpio_enable dst: Unknown symbol bt878_device_control dst_ca: Unknown symbol write_dst dst_ca: Unknown symbol dst_pio_disable dst_ca: Unknown symbol dst_comm_init dst_ca: Unknown symbol read_dst dst_ca: Unknown symbol dst_check_sum dst_ca: Unknown symbol dst_error_bailout dst_ca: Unknown symbol dst_wait_dst_ready dst_ca: Unknown symbol rdc_reset_state dst_ca: Unknown symbol dst_error_recovery dvb_bt8xx: disagrees about version of symbol dvb_dmxdev_init dvb_bt8xx: Unknown symbol dvb_dmxdev_init dvb_bt8xx: disagrees about version of symbol nxt6000_attach dvb_bt8xx: Unknown symbol nxt6000_attach dvb_bt8xx: disagrees about version of symbol mt352_attach dvb_bt8xx: Unknown symbol mt352_attach dvb_bt8xx: Unknown symbol bt878_num dvb_bt8xx: Unknown symbol bttv_sub_unregister dvb_bt8xx: Unknown symbol dst_attach dvb_bt8xx: Unknown symbol bttv_write_gpio dvb_bt8xx: disagrees about version of symbol dvb_dmx_swfilter_204 dvb_bt8xx: Unknown symbol dvb_dmx_swfilter_204 dvb_bt8xx: Unknown symbol bttv_sub_register dvb_bt8xx: disagrees about version of symbol dvb_dmx_release dvb_bt8xx: Unknown symbol dvb_dmx_release dvb_bt8xx: disagrees about version of symbol sp887x_attach dvb_bt8xx: Unknown symbol sp887x_attach dvb_bt8xx: Unknown symbol bttv_get_pcidev dvb_bt8xx: disagrees about version of symbol dvb_net_init dvb_bt8xx: Unknown symbol dvb_net_init dvb_bt8xx: disagrees about version of symbol dvb_dmx_swfilter dvb_bt8xx: Unknown symbol dvb_dmx_swfilter dvb_bt8xx: disagrees about version of symbol dvb_dmxdev_release dvb_bt8xx: Unknown symbol dvb_dmxdev_release dvb_bt8xx: Unknown symbol dst_ca_attach dvb_bt8xx: Unknown symbol bt878_start dvb_bt8xx: disagrees about version of symbol dvb_net_release dvb_bt8xx: Unknown symbol dvb_net_release dvb_bt8xx: Unknown symbol bttv_gpio_enable dvb_bt8xx: disagrees about version of symbol cx24110_pll_write dvb_bt8xx: Unknown symbol cx24110_pll_write dvb_bt8xx: disagrees about version of symbol or51211_attach dvb_bt8xx: Unknown symbol or51211_attach dvb_bt8xx: disagrees about version of symbol dvb_unregister_frontend dvb_bt8xx: Unknown symbol dvb_unregister_frontend dvb_bt8xx: Unknown symbol bt878_stop dvb_bt8xx: disagrees about version of symbol dvb_register_frontend dvb_bt8xx: Unknown symbol dvb_register_frontend dvb_bt8xx: disagrees about version of symbol mt352_write dvb_bt8xx: Unknown symbol mt352_write dvb_bt8xx: disagrees about version of symbol dvb_dmx_init dvb_bt8xx: Unknown symbol dvb_dmx_init dvb_bt8xx: Unknown symbol bt878 dvb_bt8xx: disagrees about version of symbol cx24110_attach dvb_bt8xx: Unknown symbol cx24110_attach Message: 4 Date: Mon, 09 Jan 2006 14:04:25 +0400 From: Manu Abraham <abraham.manu@xxxxxxxxx> Subject: Re: [linux-dvb] Another Twinhan VP20310 problem To: "J. van Amerongen" <joel@xxxxxxxxx> Cc: linux-dvb@xxxxxxxxxxx Message-ID: <43C23529.7020700@xxxxxxxxx> Content-Type: text/plain; charset=ISO-8859-1; format=flowed J. van Amerongen wrote: >I have already going through: >http://www.linuxtv.org/pipermail/linux-dvb/2005-September/004871.html > >The card can tune and scan. But no picture. >I have try: > > > > >There is one thing I didn't try, because I can't find: > > >>try changing the dst_tlist[] (for DCT-CI ) >>.type_flags = DST_TYPE_HAS_NEWTUNE | DST_TYPE_HAS_FW_2 >>only and see what happens. >> >> You would require this one. This is probably fixed in the v4l-algo test tree.. But unfortunately my link is down (on dialup right now) and hence will not be able to access the repo. (It's holidays here and the ISP does not have tech. people to fix it) I will send you a tarball of the tree, in case you can test it out. Manu ------------------------------ Message: 5 Date: Mon, 09 Jan 2006 14:09:16 +0400 From: Manu Abraham <abraham.manu@xxxxxxxxx> Subject: Re: [linux-dvb] [PATCH 3/6] DVB-PinnSat: Remove op_sync_orin and irq_err_ignore To: Edgar Toernig <froese@xxxxxx> Cc: linux-dvb@xxxxxxxxxxx Message-ID: <43C2364C.1060201@xxxxxxxxx> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Edgar Toernig wrote: >The op_sync_orin and irq_err_ignore code is not really card >specific. Some cards set op_sync_orin but any card would >benefit from setting it. And irq_err_ignore just shuts up >some error printks in the irq handler. > >This patch activates resyncing for all cards thus making >op_sync_orin superfluous. > >irq_err_ignore is removed. The error printks are only >activated then debugging is turned on. > > > To make this work you need not move the card specific application of these parameters globally to all cards. The reason being that individual modules loose the capability of changing this option as all cards will be affected in the very same manner. As i said earlier, i am against this patch. The same feature can be achieved, without touching all modules .. You can see how the very same can be achieved without stirring up so much. Manu Signed-off-by: Manu Abraham <manu@xxxxxxxxxxx> # HG changeset patch # User root@xxxxxxxxxxxxxxxx # Date Sat Dec 24 17:01:34 2005 # Node ID 1b4ffaea140aaff910e5d2b26451c57d2a88caa5 # parent: f515634631e7165ebcc6c995be92661f70587862 Do a RISC_SYNC also --- a/linux/drivers/media/dvb/bt8xx/dvb-bt8xx.c Sat Dec 24 16:54:12 2005 +++ b/linux/drivers/media/dvb/bt8xx/dvb-bt8xx.c Sat Dec 24 17:01:34 2005 @@ -801,13 +801,13 @@ card->gpio_mode = 0x0400c060; /* should be: BT878_A_GAIN=0,BT878_A_PWRDN,BT878_DA_DPM,BT878_DA_SBR, BT878_DA_IOM=1,BT878_DA_APP to enable serial highspeed mode. */ - card->op_sync_orin = 0; + card->op_sync_orin = BT878_RISC_SYNC_MASK; card->irq_err_ignore = BT878_AFBUS | BT878_AFDSR; break; case BTTV_BOARD_DVICO_DVBT_LITE: card->gpio_mode = 0x0400C060; - card->op_sync_orin = 0; + card->op_sync_orin = BT878_RISC_SYNC_MASK; card->irq_err_ignore = BT878_AFBUS | BT878_AFDSR; /* 26, 15, 14, 6, 5 * A_PWRDN DA_DPM DA_SBR DA_IOM_DA @@ -823,7 +823,7 @@ case BTTV_BOARD_NEBULA_DIGITV: case BTTV_BOARD_AVDVBT_761: card->gpio_mode = (1 << 26) | (1 << 14) | (1 << 5); - card->op_sync_orin = 0; + card->op_sync_orin = BT878_RISC_SYNC_MASK; card->irq_err_ignore = BT878_AFBUS | BT878_AFDSR; /* A_PWRDN DA_SBR DA_APP (high speed serial) */ break; @@ -855,7 +855,7 @@ case BTTV_BOARD_PC_HDTV: card->gpio_mode = 0x0100EC7B; - card->op_sync_orin = 0; + card->op_sync_orin = BT878_RISC_SYNC_MASK; card->irq_err_ignore = BT878_AFBUS | BT878_AFDSR; break; # HG changeset patch # User root@xxxxxxxxxxxxxxxx # Date Sat Dec 24 16:54:12 2005 # Node ID f515634631e7165ebcc6c995be92661f70587862 # parent: 70e9c12d0066a13eb2902f5a5cf0610e92a7457f Ignore FBUS and FDSR errors --- a/linux/drivers/media/dvb/bt8xx/dvb-bt8xx.c Wed Dec 21 13:24:21 2005 +++ b/linux/drivers/media/dvb/bt8xx/dvb-bt8xx.c Sat Dec 24 16:54:12 2005 @@ -802,13 +802,13 @@ /* should be: BT878_A_GAIN=0,BT878_A_PWRDN,BT878_DA_DPM,BT878_DA_SBR, BT878_DA_IOM=1,BT878_DA_APP to enable serial highspeed mode. */ card->op_sync_orin = 0; - card->irq_err_ignore = 0; + card->irq_err_ignore = BT878_AFBUS | BT878_AFDSR; break; case BTTV_BOARD_DVICO_DVBT_LITE: card->gpio_mode = 0x0400C060; card->op_sync_orin = 0; - card->irq_err_ignore = 0; + card->irq_err_ignore = BT878_AFBUS | BT878_AFDSR; /* 26, 15, 14, 6, 5 * A_PWRDN DA_DPM DA_SBR DA_IOM_DA * DA_APP(parallel) */ @@ -824,14 +824,14 @@ case BTTV_BOARD_AVDVBT_761: card->gpio_mode = (1 << 26) | (1 << 14) | (1 << 5); card->op_sync_orin = 0; - card->irq_err_ignore = 0; + card->irq_err_ignore = BT878_AFBUS | BT878_AFDSR; /* A_PWRDN DA_SBR DA_APP (high speed serial) */ break; case BTTV_BOARD_AVDVBT_771: //case 0x07711461: card->gpio_mode = 0x0400402B; card->op_sync_orin = BT878_RISC_SYNC_MASK; - card->irq_err_ignore = 0; + card->irq_err_ignore = BT878_AFBUS | BT878_AFDSR; /* A_PWRDN DA_SBR DA_APP[0] PKTP=10 RISC_ENABLE FIFO_ENABLE*/ break; @@ -856,7 +856,7 @@ case BTTV_BOARD_PC_HDTV: card->gpio_mode = 0x0100EC7B; card->op_sync_orin = 0; - card->irq_err_ignore = 0; + card->irq_err_ignore = BT878_AFBUS | BT878_AFDSR; break; default: ------------------------------ Message: 6 Date: Mon, 09 Jan 2006 14:10:13 +0400 From: Manu Abraham <abraham.manu@xxxxxxxxx> Subject: Re: [linux-dvb] [PATCH 6/6] DVB-PinnSat: Misc cleanup and robustness tweaks To: Edgar Toernig <froese@xxxxxx> Cc: linux-dvb@xxxxxxxxxxx Message-ID: <43C23685.7060407@xxxxxxxxx> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Edgar Toernig wrote: >General cleanups and some tweaks to make the code more >robust, especially against fifo overrun errors. > >I still don't know what causes the audio fifo to overflow >on some chipsets - the video fifo seems to never show >this problem (yeah, it's two time as big but the data >rate is much more then twice of that comming via DVB). > > The overflow does happen on certain chipset's (chipset quirks ) or due to PCI bus saturation. Where i have seen this issue is when using multiple cards in the order of cards > 3. But even then the PCI bus should be able to handle this bandwidth. since a transponder is only approx 27.5 Mbps. Even with three such cards, data rate does not even go near the actual PCI b/w. The reason why it overflows is data is not read from the FIFO as it should be and the PCI bus expects larger chunks. Another way of handling this is to reduce the latency, such that even smaller chunks can be handled. But even then for example having a motherboard with all peripherals on the same bus can create problems, as data bursts from other peripherals can create a substantial delay, if latency is set too high. >Maybe some PCI-bridges honour the (ro) MAX_LAT register >nowadays - it's much too high for DVB data rates. > >--- > > bt878.c | 105 ++++++++++++++++++++++++++++++-------------------------- > bt878.h | 14 ++----- > 2 files changed, 61 insertions(+), 58 deletions(-) > >--- 0.6/drivers/media/dvb/bt8xx/bt878.h Sat, 07 Jan 2006 02:03:06 +0100 froese (kernel-dvb/h/15_bt878.h 1.2 644) >+++ 0.7/drivers/media/dvb/bt8xx/bt878.h Sun, 08 Jan 2006 01:45:15 +0100 froese (kernel-dvb/h/15_bt878.h 1.3 644) >@@ -28,7 +28,7 @@ > #include "bt848.h" > #include "bttv.h" > >-#define BT878_VERSION_CODE 0x000000 >+#define BT878_VERSION_CODE 0x000001 > > #define BT878_AINT_STAT 0x100 > #define BT878_ARISCS (0xf<<28) >@@ -75,15 +75,6 @@ > > #define BT878_ARISC_PC 0x120 > >-/* BT878 FUNCTION 0 REGISTERS */ >-#define BT878_GPIO_DMA_CTL 0x10c >- >-/* Interrupt register */ >-#define BT878_INT_STAT 0x100 >-#define BT878_INT_MASK 0x104 >-#define BT878_I2CRACK (1<<25) >-#define BT878_I2CDONE (1<<8) >- > #define BT878_MAX 4 > > extern int bt878_num; >@@ -117,6 +108,9 @@ > dma_addr_t risc_dma; > u32 risc_pos; > >+ unsigned int errors; >+ unsigned long error_expire; >+ > struct tasklet_struct tasklet; > int shutdown; > }; >--- 0.6/drivers/media/dvb/bt8xx/bt878.c Sat, 07 Jan 2006 02:03:06 +0100 froese (kernel-dvb/h/16_bt878.c 1.3 644) >+++ 0.7/drivers/media/dvb/bt8xx/bt878.c Sun, 08 Jan 2006 01:45:15 +0100 froese (kernel-dvb/h/16_bt878.c 1.4 644) >@@ -143,9 +143,9 @@ > #define RISC_SYNC_VRO 0x0C > > #define RISC_FLUSH() bt->risc_pos = 0 >-#define RISC_INSTR(instr) bt->risc_cpu[bt->risc_pos++] = cpu_to_le32(instr) >+#define RISC_INSTR(instr) bt->risc_cpu[bt->risc_pos++] = cpu_to_le32((instr)) > >-static int bt878_make_risc(struct bt878 *bt) >+static int bt878_calc_line_size(struct bt878 *bt) > > calc_line_size would be misleading, since it is in fact creating the RISC program itself. IIRC, there is another function calculating the line size too. So you have merged those two functions .. ? Any specific reason why those two functions should be merged into one function ? > { > bt->block_bytes = bt->buf_size >> 4; > bt->block_count = 1 << 4; >@@ -177,16 +177,15 @@ > dprintk("bt878: risc len lines %u, bytes per line %u\n", > bt->line_count, bt->line_bytes); > for (line = 0; line < bt->line_count; line++) { >- // At the beginning of every block we issue an IRQ with previous (finished) block number set >+ // At the beginning of every block we issue an IRQ with >+ // previous (finished) block number set > if (!(buf_pos % bt->block_bytes)) > RISC_INSTR(RISC_WRITE | RISC_WR_SOL | RISC_WR_EOL | > RISC_IRQ | >- RISC_STATUS(((buf_pos / >- bt->block_bytes) + >- (bt->block_count - >- 1)) % >- bt->block_count) | bt-> >- line_bytes); >+ RISC_STATUS(((buf_pos / bt->block_bytes) + >+ bt->block_count - 1) % >+ bt->block_count) | >+ bt->line_bytes); > else > RISC_INSTR(RISC_WRITE | RISC_WR_SOL | RISC_WR_EOL | > bt->line_bytes); >@@ -212,23 +211,26 @@ > u32 int_mask; > > dprintk("bt878 debug: bt878_start (ctl=%8.8x)\n", controlreg); >- /* complete the writing of the risc dma program now we have >- * the card specifics >- */ >- bt878_risc_program(bt); >+ > controlreg &= ~0x1f; >- controlreg |= 0x1b; >+ btwrite(controlreg, BT878_AGPIO_DMA_CTL); > > > > btwrite(bt->risc_dma, BT878_ARISC_START); >+ bt->last_block = bt->block_count - 1; >+ >+ bt->errors = 0; > > int_mask = BT878_ARISCI; > if (bt878_debug) > int_mask |= BT878_ASCERR | BT878_AOCERR | > BT878_APABORT | BT878_ARIPERR | BT878_APPERR | > BT878_AFDSR | BT878_AFTRGT | BT878_AFBUS; >- >+ btwrite(int_mask, BT878_AINT_STAT); > btwrite(int_mask, BT878_AINT_MASK); >- btwrite(controlreg, BT878_AGPIO_DMA_CTL); >+ >+ /* start dma (with fifo threshold at minimum) */ >+ /* just in case with a read-modify-write to flush prior writes */ >+ btor(0x13, BT878_AGPIO_DMA_CTL); > } > > void bt878_stop(struct bt878 *bt) >@@ -261,7 +263,7 @@ > > static irqreturn_t bt878_irq(int irq, void *dev_id, struct pt_regs *regs) > { >- u32 stat, astat, mask; >+ u32 stat, astat; > int count; > struct bt878 *bt; > >@@ -270,13 +272,16 @@ > count = 0; > while (1) { > stat = btread(BT878_AINT_STAT); >- mask = btread(BT878_AINT_MASK); >- if (!(astat = (stat & mask))) >- return IRQ_NONE; /* this interrupt is not for me */ >-/* dprintk("bt878(%d) debug: irq count %d, stat 0x%8.8x, mask 0x%8.8x\n",bt->nr,count,stat,mask); */ >- btwrite(astat, BT878_AINT_STAT); /* try to clear interupt condition */ >+ astat = stat & btread(BT878_AINT_MASK); >+ if (!astat) >+ break; /* not for us */ > >+ btwrite(astat, BT878_AINT_STAT); /* ack interrupt */ > >+ if (astat & BT878_ARISCI) { >+ bt->finished_block = (stat & BT878_ARISCS) >> 28; >+ tasklet_schedule(&bt->tasklet); >+ } > if (astat & (BT878_ASCERR | BT878_AOCERR)) { > if (bt878_verbose) { > printk(KERN_ERR >@@ -309,21 +314,33 @@ > btread(BT878_ARISC_PC)); > } > } >- if (astat & BT878_ARISCI) { >- bt->finished_block = (stat & BT878_ARISCS) >> 28; >- tasklet_schedule(&bt->tasklet); >- break; >+ if (astat & ~BT878_ARISCI) { >+ if (time_after(jiffies, bt->error_expire)) >+ bt->errors = 0; >+ bt->error_expire = jiffies + 5*HZ; >+ bt->errors++; > > IMHO, You shouldn't be sleeping inside an interrupt handler. Use the tasklet itself, if you need to sleep, for a predefined period, sleep within the tasklet. But i wouldn't advise sleeping there either, since this will cause other *working* modules which depend on this also to sleep. Such changes would require all 878 cards to be tested quite thoroughly. Right now i can't test this part ( for the dst, since it is a bit broken at present ) >+ if (bt->errors == 10 || bt->errors == 15) { >+ printk(KERN_ERR "bt878(%d): too many errors, " >+ "resetting dma\n", bt->nr); >+ /* reset dma and set fifo-trigger to minimum */ >+ btand(~0x1f, BT878_AGPIO_DMA_CTL); >+ btor(0x13, BT878_AGPIO_DMA_CTL); >+ } >+ if (bt->errors == 20) { >+ printk(KERN_ERR "bt878(%d): too many errors, " >+ "shutting up\n", bt->nr); >+ btwrite(BT878_ARISCI, BT878_AINT_MASK); >+ } > > I would say that just, errors > error_count, do_operations would be sufficient .. But anyway what's the idea behind reducing the fifo size if in case of errors ? normally when we have communication errors , we generally increase the buffer size, not reduce it... > } > count++; >- if (count > 20) { >+ if (count > 30) { > btwrite(0, BT878_AINT_MASK); >- printk(KERN_ERR >- "bt878(%d): IRQ lockup, cleared int mask\n", >- bt->nr); >+ printk(KERN_ERR "bt878(%d): IRQ lockup," >+ " device disabled\n", bt->nr); > break; > } > } >- return IRQ_HANDLED; >+ return IRQ_RETVAL(count); > } > > int >@@ -424,42 +441,35 @@ > bt->bt878_mem = ioremap(bt->bt878_adr, 0x1000); > #endif > >- /* clear interrupt mask */ >- btwrite(0, BT848_INT_MASK); >+ /* disable interrupts and dma */ >+ btwrite(0, BT878_AINT_MASK); >+ btwrite(0, BT878_AGPIO_DMA_CTL); > > result = request_irq(bt->irq, bt878_irq, > SA_SHIRQ | SA_INTERRUPT, "bt878", > (void *) bt); >- if (result == -EINVAL) { >- printk(KERN_ERR "bt878(%d): Bad irq number or handler\n", >- bt878_num); >- goto fail1; >- } > if (result == -EBUSY) { > printk(KERN_ERR > "bt878(%d): IRQ %d busy, change your PnP config in BIOS\n", > bt878_num, bt->irq); > goto fail1; > } >- if (result < 0) >+ if (result < 0) { >+ printk(KERN_ERR "bt878(%d): Bad irq number or handler (%d)\n", >+ bt878_num, result); > goto fail1; >+ } > > pci_set_master(dev); > pci_set_drvdata(dev, bt); > >-/* if(init_bt878(btv) < 0) { >- bt878_remove(dev); >- return -EIO; >- } >-*/ >- > if ((result = bt878_mem_alloc(bt))) { > printk(KERN_ERR "bt878: failed to allocate memory!\n"); > goto fail2; > } > >- bt878_make_risc(bt); >- btwrite(0, BT878_AINT_MASK); >+ bt878_calc_line_size(bt); >+ bt878_risc_program(bt); > bt878_num++; > > return 0; >@@ -491,7 +501,6 @@ > > /* disable PCI bus-mastering */ > pci_read_config_byte(bt->dev, PCI_COMMAND, &command); >- /* Should this be &=~ ?? */ > command &= ~PCI_COMMAND_MASTER; > pci_write_config_byte(bt->dev, PCI_COMMAND, command); > > > > Manu ------------------------------ _______________________________________________ linux-dvb@xxxxxxxxxxx http://www.linuxtv.org/cgi-bin/mailman/listinfo/linux-dvb End of linux-dvb Digest, Vol 12, Issue 31 ***************************************** -- 10 GB Mailbox, 100 FreeSMS/Monat http://www.gmx.net/de/go/topmail +++ GMX - die erste Adresse für Mail, Message, More +++
