[git patches] ISDN cleanups (modularization prep)

[git patches] ISDN cleanups (modularization prep)

[Jeff Garzik]

ISDN is the last major subsystem that uses pci_find_* rather than the
"new" PCI hotplug API.  I have [untested (no h/w)] patches that convert
these drivers over to the new PCI API in #isdn-pci.

The changes below prepare us for those changes.  Both changes submitted
here are no-behavior-change patches, simply internal cleanups that
prepare us for #isdn-pci.  If any user is impacted at all, that's a bug.

These changes have been in -mm for spell, though they are not presently
there.  Since this project was started well over a year ago (with LKML,
kkeil, and isdn4linux always CC'd), the only objections raised were
not directly related:

- "isn't ISDN dead?"  Apparently, no it's not :)

- "don't waste your time, drop these drivers and get mISDN into the
  kernel instead" -- based on discussion it does sound like mISDN should
  get into the kernel.  Until these drivers are dropped, however, my
  patches remain appropriate.

And in case people are curious, #isdn-pci isn't ready for upstream yet
for two main reasons: (1) two drivers remain unconverted, and (2) none
of the changes have been tested, even once.

Please pull from 'isdn-cleanups' branch of
master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/misc-2.6.git isdn-cleanups

to receive the following updates:

 drivers/isdn/hisax/asuscom.c   |    5 +----
 drivers/isdn/hisax/avm_a1.c    |   27 +++++++++------------------
 drivers/isdn/hisax/bkm_a4t.c   |   19 ++++++++++---------
 drivers/isdn/hisax/bkm_a8.c    |   31 +++++++++++++------------------
 drivers/isdn/hisax/config.c    |   14 +++++++++-----
 drivers/isdn/hisax/gazel.c     |    5 ++---
 drivers/isdn/hisax/hfc_pci.c   |    2 --
 drivers/isdn/hisax/hfc_sx.c    |    2 --
 drivers/isdn/hisax/hisax_cfg.h |    2 ++
 drivers/isdn/hisax/isurf.c     |   13 ++++---------
 drivers/isdn/hisax/ix1_micro.c |    9 ++++-----
 drivers/isdn/hisax/mic.c       |    5 +----
 drivers/isdn/hisax/netjet.h    |    2 --
 drivers/isdn/hisax/niccy.c     |   19 ++++++++-----------
 drivers/isdn/hisax/nj_s.c      |    3 +--
 drivers/isdn/hisax/nj_u.c      |    4 ++--
 drivers/isdn/hisax/s0box.c     |   13 ++++---------
 drivers/isdn/hisax/saphir.c    |    8 +++-----
 drivers/isdn/hisax/sportster.c |   11 +++++------
 drivers/isdn/hisax/teleint.c   |    5 +----
 drivers/isdn/hisax/telespci.c  |    5 ++---
 drivers/isdn/hisax/w6692.c     |    2 --
 22 files changed, 81 insertions(+), 125 deletions(-)

Jeff Garzik (2):
      [ISDN] HiSax: modularization prep
      [ISDN] Hisax: eliminate many unnecessary references to CardType[]

diff --git a/drivers/isdn/hisax/asuscom.c b/drivers/isdn/hisax/asuscom.c
index 61e69e9..b96f318 100644
--- a/drivers/isdn/hisax/asuscom.c
+++ b/drivers/isdn/hisax/asuscom.c
@@ -20,8 +20,6 @@
 #include "hscx.h"
 #include "isdnl1.h"
 
-extern const char *CardType[];
-
 static const char *Asuscom_revision = "$Revision: 1.14.2.4 $";
 
 #define byteout(addr,val) outb(val,addr)
@@ -376,8 +374,7 @@ setup_asuscom(struct IsdnCard *card)
 	cs->irq = card->para[0];
 	if (!request_region(cs->hw.asus.cfg_reg, bytecnt, "asuscom isdn")) {
 		printk(KERN_WARNING
-		       "HiSax: %s config port %x-%x already in use\n",
-		       CardType[card->typ],
+		       "HiSax: ISDNLink config port %x-%x already in use\n",
 		       cs->hw.asus.cfg_reg,
 		       cs->hw.asus.cfg_reg + bytecnt);
 		return (0);
diff --git a/drivers/isdn/hisax/avm_a1.c b/drivers/isdn/hisax/avm_a1.c
index d9028e9..eb6b432 100644
--- a/drivers/isdn/hisax/avm_a1.c
+++ b/drivers/isdn/hisax/avm_a1.c
@@ -16,7 +16,6 @@
 #include "hscx.h"
 #include "isdnl1.h"
 
-extern const char *CardType[];
 static const char *avm_revision = "$Revision: 2.15.2.4 $";
 
 #define	 AVM_A1_STAT_ISAC	0x01
@@ -200,16 +199,14 @@ setup_avm_a1(struct IsdnCard *card)
 	cs->irq = card->para[0];
 	if (!request_region(cs->hw.avm.cfg_reg, 8, "avm cfg")) {
 		printk(KERN_WARNING
-		       "HiSax: %s config port %x-%x already in use\n",
-		       CardType[card->typ],
+		       "HiSax: AVM A1 config port %x-%x already in use\n",
 		       cs->hw.avm.cfg_reg,
 		       cs->hw.avm.cfg_reg + 8);
 		return (0);
 	}
 	if (!request_region(cs->hw.avm.isac + 32, 32, "HiSax isac")) {
 		printk(KERN_WARNING
-		       "HiSax: %s isac ports %x-%x already in use\n",
-		       CardType[cs->typ],
+		       "HiSax: AVM A1 isac ports %x-%x already in use\n",
 		       cs->hw.avm.isac + 32,
 		       cs->hw.avm.isac + 64);
 		release_ioregs(cs, 0);
@@ -217,16 +214,14 @@ setup_avm_a1(struct IsdnCard *card)
 	}
 	if (!request_region(cs->hw.avm.isacfifo, 1, "HiSax isac fifo")) {
 		printk(KERN_WARNING
-		       "HiSax: %s isac fifo port %x already in use\n",
-		       CardType[cs->typ],
+		       "HiSax: AVM A1 isac fifo port %x already in use\n",
 		       cs->hw.avm.isacfifo);
 		release_ioregs(cs, 1);
 		return (0);
 	}
 	if (!request_region(cs->hw.avm.hscx[0] + 32, 32, "HiSax hscx A")) {
 		printk(KERN_WARNING
-		       "HiSax: %s hscx A ports %x-%x already in use\n",
-		       CardType[cs->typ],
+		       "HiSax: AVM A1 hscx A ports %x-%x already in use\n",
 		       cs->hw.avm.hscx[0] + 32,
 		       cs->hw.avm.hscx[0] + 64);
 		release_ioregs(cs, 3);
@@ -234,16 +229,14 @@ setup_avm_a1(struct IsdnCard *card)
 	}
 	if (!request_region(cs->hw.avm.hscxfifo[0], 1, "HiSax hscx A fifo")) {
 		printk(KERN_WARNING
-		       "HiSax: %s hscx A fifo port %x already in use\n",
-		       CardType[cs->typ],
+		       "HiSax: AVM A1 hscx A fifo port %x already in use\n",
 		       cs->hw.avm.hscxfifo[0]);
 		release_ioregs(cs, 7);
 		return (0);
 	}
 	if (!request_region(cs->hw.avm.hscx[1] + 32, 32, "HiSax hscx B")) {
 		printk(KERN_WARNING
-		       "HiSax: %s hscx B ports %x-%x already in use\n",
-		       CardType[cs->typ],
+		       "HiSax: AVM A1 hscx B ports %x-%x already in use\n",
 		       cs->hw.avm.hscx[1] + 32,
 		       cs->hw.avm.hscx[1] + 64);
 		release_ioregs(cs, 0xf);
@@ -251,8 +244,7 @@ setup_avm_a1(struct IsdnCard *card)
 	}
 	if (!request_region(cs->hw.avm.hscxfifo[1], 1, "HiSax hscx B fifo")) {
 		printk(KERN_WARNING
-		       "HiSax: %s hscx B fifo port %x already in use\n",
-		       CardType[cs->typ],
+		       "HiSax: AVM A1 hscx B fifo port %x already in use\n",
 		       cs->hw.avm.hscxfifo[1]);
 		release_ioregs(cs, 0x1f);
 		return (0);
@@ -284,9 +276,8 @@ setup_avm_a1(struct IsdnCard *card)
 	printk(KERN_INFO "AVM A1: Byte at %x is %x\n",
 	       cs->hw.avm.cfg_reg, val);
 
-	printk(KERN_INFO
-	       "HiSax: %s config irq:%d cfg:0x%X\n",
-	       CardType[cs->typ], cs->irq,
+	printk(KERN_INFO "HiSax: AVM A1 config irq:%d cfg:0x%X\n",
+	       cs->irq,
 	       cs->hw.avm.cfg_reg);
 	printk(KERN_INFO
 	       "HiSax: isac:0x%X/0x%X\n",
diff --git a/drivers/isdn/hisax/bkm_a4t.c b/drivers/isdn/hisax/bkm_a4t.c
index 3d1bdc8..9ca2ee5 100644
--- a/drivers/isdn/hisax/bkm_a4t.c
+++ b/drivers/isdn/hisax/bkm_a4t.c
@@ -20,8 +20,6 @@
 #include <linux/pci.h>
 #include "bkm_ax.h"
 
-extern const char *CardType[];
-
 static const char *bkm_a4t_revision = "$Revision: 1.22.2.4 $";
 
 
@@ -284,15 +282,16 @@ static int __devinit a4t_cs_init(struct IsdnCard *card,
 	I20_REGISTER_FILE *pI20_Regs;
 
 	if (!cs->irq) {		/* IRQ range check ?? */
-		printk(KERN_WARNING "HiSax: %s: No IRQ\n", CardType[card->typ]);
+		printk(KERN_WARNING "HiSax: Telekom A4T: No IRQ\n");
 		return (0);
 	}
 	cs->hw.ax.base = (long) ioremap(pci_memaddr, 4096);
 	/* Check suspecious address */
 	pI20_Regs = (I20_REGISTER_FILE *) (cs->hw.ax.base);
 	if ((pI20_Regs->i20IntStatus & 0x8EFFFFFF) != 0) {
-		printk(KERN_WARNING "HiSax: %s address %lx-%lx suspecious\n",
-		       CardType[card->typ], cs->hw.ax.base, cs->hw.ax.base + 4096);
+		printk(KERN_WARNING "HiSax: Telekom A4T address "
+		       "%lx-%lx suspicious\n",
+		       cs->hw.ax.base, cs->hw.ax.base + 4096);
 		iounmap((void *) cs->hw.ax.base);
 		cs->hw.ax.base = 0;
 		return (0);
@@ -302,8 +301,9 @@ static int __devinit a4t_cs_init(struct IsdnCard *card,
 	cs->hw.ax.isac_ale = GCS_1;
 	cs->hw.ax.jade_ale = GCS_3;
 
-	printk(KERN_INFO "HiSax: %s: Card configured at 0x%lX IRQ %d\n",
-	       CardType[card->typ], cs->hw.ax.base, cs->irq);
+	printk(KERN_INFO "HiSax: Telekom A4T: Card configured at "
+	       "0x%lX IRQ %d\n",
+	       cs->hw.ax.base, cs->irq);
 
 	setup_isac(cs);
 	cs->readisac = &ReadISAC;
@@ -349,11 +349,12 @@ setup_bkm_a4t(struct IsdnCard *card)
 			break;
 	}
 	if (!found) {
-		printk(KERN_WARNING "HiSax: %s: Card not found\n", CardType[card->typ]);
+		printk(KERN_WARNING "HiSax: Telekom A4T: Card not found\n");
 		return (0);
 	}
 	if (!pci_memaddr) {
-		printk(KERN_WARNING "HiSax: %s: No Memory base address\n", CardType[card->typ]);
+		printk(KERN_WARNING "HiSax: Telekom A4T: "
+		       "No Memory base address\n");
 		return (0);
 	}
 
diff --git a/drivers/isdn/hisax/bkm_a8.c b/drivers/isdn/hisax/bkm_a8.c
index 99ef3b4..e1ff471 100644
--- a/drivers/isdn/hisax/bkm_a8.c
+++ b/drivers/isdn/hisax/bkm_a8.c
@@ -22,8 +22,6 @@
 
 #define	ATTEMPT_PCI_REMAPPING	/* Required for PLX rev 1 */
 
-extern const char *CardType[];
-
 static const char sct_quadro_revision[] = "$Revision: 1.22.2.4 $";
 
 static const char *sct_quadro_subtypes[] =
@@ -181,8 +179,7 @@ bkm_interrupt_ipac(int intno, void *dev_id)
 		goto Start_IPAC;
 	}
 	if (!icnt)
-		printk(KERN_WARNING "HiSax: %s (%s) IRQ LOOP\n",
-		       CardType[cs->typ],
+		printk(KERN_WARNING "HiSax: Scitel Quadro (%s) IRQ LOOP\n",
 		       sct_quadro_subtypes[cs->subtyp]);
 	writereg(cs->hw.ax.base, cs->hw.ax.data_adr, IPAC_MASK, 0xFF);
 	writereg(cs->hw.ax.base, cs->hw.ax.data_adr, IPAC_MASK, 0xC0);
@@ -296,8 +293,8 @@ setup_sct_quadro(struct IsdnCard *card)
 	if (card->para[0] >= SCT_1 && card->para[0] <= SCT_4)
 		cs->subtyp = card->para[0];
 	else {
-		printk(KERN_WARNING "HiSax: %s: Invalid subcontroller in configuration, default to 1\n",
-			CardType[card->typ]);
+		printk(KERN_WARNING "HiSax: Scitel Quadro: Invalid "
+		       "subcontroller in configuration, default to 1\n");
 		return (0);
 	}
 	if ((cs->subtyp != SCT_1) && ((sub_sys_id != PCI_DEVICE_ID_BERKOM_SCITEL_QUADRO) ||
@@ -322,16 +319,16 @@ setup_sct_quadro(struct IsdnCard *card)
 			}
 		}
 		if (!found) {
-			printk(KERN_WARNING "HiSax: %s (%s): Card not found\n",
-				CardType[card->typ],
+			printk(KERN_WARNING "HiSax: Scitel Quadro (%s): "
+				"Card not found\n",
 				sct_quadro_subtypes[cs->subtyp]);
 			return (0);
 		}
 #ifdef ATTEMPT_PCI_REMAPPING
 /* HACK: PLX revision 1 bug: PLX address bit 7 must not be set */
 		if ((pci_ioaddr1 & 0x80) && (dev_a8->revision == 1)) {
-			printk(KERN_WARNING "HiSax: %s (%s): PLX rev 1, remapping required!\n",
-				CardType[card->typ],
+			printk(KERN_WARNING "HiSax: Scitel Quadro (%s): "
+				"PLX rev 1, remapping required!\n",
 				sct_quadro_subtypes[cs->subtyp]);
 			/* Restart PCI negotiation */
 			pci_write_config_dword(dev_a8, PCI_BASE_ADDRESS_1, (u_int) - 1);
@@ -344,8 +341,7 @@ setup_sct_quadro(struct IsdnCard *card)
 #endif /* End HACK */
 	}
 	if (!pci_irq) {		/* IRQ range check ?? */
-		printk(KERN_WARNING "HiSax: %s (%s): No IRQ\n",
-		       CardType[card->typ],
+		printk(KERN_WARNING "HiSax: Scitel Quadro (%s): No IRQ\n",
 		       sct_quadro_subtypes[cs->subtyp]);
 		return (0);
 	}
@@ -355,8 +351,8 @@ setup_sct_quadro(struct IsdnCard *card)
 	pci_read_config_dword(dev_a8, PCI_BASE_ADDRESS_4, &pci_ioaddr4);
 	pci_read_config_dword(dev_a8, PCI_BASE_ADDRESS_5, &pci_ioaddr5);
 	if (!pci_ioaddr1 || !pci_ioaddr2 || !pci_ioaddr3 || !pci_ioaddr4 || !pci_ioaddr5) {
-		printk(KERN_WARNING "HiSax: %s (%s): No IO base address(es)\n",
-		       CardType[card->typ],
+		printk(KERN_WARNING "HiSax: Scitel Quadro (%s): "
+		       "No IO base address(es)\n",
 		       sct_quadro_subtypes[cs->subtyp]);
 		return (0);
 	}
@@ -411,8 +407,8 @@ setup_sct_quadro(struct IsdnCard *card)
 	/* For isac and hscx data path */
 	cs->hw.ax.data_adr = cs->hw.ax.base + 4;
 
-	printk(KERN_INFO "HiSax: %s (%s) configured at 0x%.4lX, 0x%.4lX, 0x%.4lX and IRQ %d\n",
-	       CardType[card->typ],
+	printk(KERN_INFO "HiSax: Scitel Quadro (%s) configured at "
+	       "0x%.4lX, 0x%.4lX, 0x%.4lX and IRQ %d\n",
 	       sct_quadro_subtypes[cs->subtyp],
 	       cs->hw.ax.plx_adr,
 	       cs->hw.ax.base,
@@ -432,8 +428,7 @@ setup_sct_quadro(struct IsdnCard *card)
 	cs->cardmsg = &BKM_card_msg;
 	cs->irq_func = &bkm_interrupt_ipac;
 
-	printk(KERN_INFO "HiSax: %s (%s): IPAC Version %d\n",
-		CardType[card->typ],
+	printk(KERN_INFO "HiSax: Scitel Quadro (%s): IPAC Version %d\n",
 		sct_quadro_subtypes[cs->subtyp],
 		readreg(cs->hw.ax.base, cs->hw.ax.data_adr, IPAC_ID));
 	return (1);
diff --git a/drivers/isdn/hisax/config.c b/drivers/isdn/hisax/config.c
index a0ee43c..84d75a3 100644
--- a/drivers/isdn/hisax/config.c
+++ b/drivers/isdn/hisax/config.c
@@ -1169,7 +1169,9 @@ outf_cs:
 /* Used from an exported function but calls __devinit functions.
  * Tell modpost not to warn (__ref)
  */
-static int __ref checkcard(int cardnr, char *id, int *busy_flag, struct module *lockowner)
+static int __ref checkcard(int cardnr, char *id, int *busy_flag,
+			   struct module *lockowner,
+			   hisax_setup_func_t card_setup)
 {
 	int ret;
 	struct IsdnCard *card = cards + cardnr;
@@ -1187,7 +1189,7 @@ static int __ref checkcard(int cardnr, char *id, int *busy_flag, struct module *
 	       (card->protocol == ISDN_PTYPE_NI1) ? "NI1" :
 	       "NONE", cs->iif.id, cs->myid);
 
-	ret = hisax_cs_setup_card(card);
+	ret = card_setup(card);
 	if (!ret) {
 		ll_unload(cs);
 		goto outf_cs;
@@ -1241,7 +1243,8 @@ static int HiSax_inithardware(int *busy_flag)
 			else
 				sprintf(ids, "%s%d", id, i);
 		}
-		if (checkcard(i, ids, busy_flag, THIS_MODULE)) {
+		if (checkcard(i, ids, busy_flag, THIS_MODULE,
+			      hisax_cs_setup_card)) {
 			foundcards++;
 			i++;
 		} else {
@@ -1549,7 +1552,8 @@ int hisax_init_pcmcia(void *pcm_iob, int *busy_flag, struct IsdnCard *card)
 		sprintf(ids, "HiSax%d", nrcards);
 	else
 		sprintf(ids, "HiSax");
-	if (!checkcard(nrcards, ids, busy_flag, THIS_MODULE))
+	if (!checkcard(nrcards, ids, busy_flag, THIS_MODULE,
+		       hisax_cs_setup_card))
 		goto error;
 
 	ret = nrcards;
@@ -1595,7 +1599,7 @@ int hisax_register(struct hisax_d_if *hisax_d_if, struct hisax_b_if *b_if[],
 	cards[i].protocol = protocol;
 	sprintf(id, "%s%d", name, i);
 	nrcards++;
-	retval = checkcard(i, id, NULL, hisax_d_if->owner);
+	retval = checkcard(i, id, NULL, hisax_d_if->owner, hisax_cs_setup_card);
 	if (retval == 0) { // yuck
 		cards[i].typ = 0;
 		nrcards--;
diff --git a/drivers/isdn/hisax/gazel.c b/drivers/isdn/hisax/gazel.c
index f66620a..0ea3b46 100644
--- a/drivers/isdn/hisax/gazel.c
+++ b/drivers/isdn/hisax/gazel.c
@@ -19,7 +19,6 @@
 #include "ipac.h"
 #include <linux/pci.h>
 
-extern const char *CardType[];
 static const char *gazel_revision = "$Revision: 2.19.2.4 $";
 
 #define R647      1
@@ -479,8 +478,8 @@ reserve_regions(struct IsdnCard *card, struct IsdnCardState *cs)
 	return 0;
 
       error:
-	printk(KERN_WARNING "Gazel: %s io ports 0x%x-0x%x already in use\n",
-	       CardType[cs->typ], adr, adr + len);
+	printk(KERN_WARNING "Gazel: io ports 0x%x-0x%x already in use\n",
+	       adr, adr + len);
 	return 1;
 }
 
diff --git a/drivers/isdn/hisax/hfc_pci.c b/drivers/isdn/hisax/hfc_pci.c
index fba8b62..f126566 100644
--- a/drivers/isdn/hisax/hfc_pci.c
+++ b/drivers/isdn/hisax/hfc_pci.c
@@ -22,8 +22,6 @@
 #include <linux/pci.h>
 #include <linux/interrupt.h>
 
-extern const char *CardType[];
-
 static const char *hfcpci_revision = "$Revision: 1.48.2.4 $";
 
 /* table entry in the PCI devices list */
diff --git a/drivers/isdn/hisax/hfc_sx.c b/drivers/isdn/hisax/hfc_sx.c
index 05482d2..f4a2138 100644
--- a/drivers/isdn/hisax/hfc_sx.c
+++ b/drivers/isdn/hisax/hfc_sx.c
@@ -18,8 +18,6 @@
 #include <linux/interrupt.h>
 #include <linux/isapnp.h>
 
-extern const char *CardType[];
-
 static const char *hfcsx_revision = "$Revision: 1.12.2.5 $";
 
 /***************************************/
diff --git a/drivers/isdn/hisax/hisax_cfg.h b/drivers/isdn/hisax/hisax_cfg.h
index ca3fe62..17a2fea 100644
--- a/drivers/isdn/hisax/hisax_cfg.h
+++ b/drivers/isdn/hisax/hisax_cfg.h
@@ -60,5 +60,7 @@ struct IsdnCard {
 	IsdnCardState_t	*cs;
 };
 
+typedef int (*hisax_setup_func_t)(struct IsdnCard *card);
+
 extern void	HiSax_closecard(int);
 extern int	hisax_init_pcmcia(void *, int *, IsdnCard_t *);
diff --git a/drivers/isdn/hisax/isurf.c b/drivers/isdn/hisax/isurf.c
index 55de069..ca41617 100644
--- a/drivers/isdn/hisax/isurf.c
+++ b/drivers/isdn/hisax/isurf.c
@@ -17,8 +17,6 @@
 #include "isdnl1.h"
 #include <linux/isapnp.h>
 
-extern const char *CardType[];
-
 static const char *ISurf_revision = "$Revision: 1.12.2.4 $";
 
 #define byteout(addr,val) outb(val,addr)
@@ -251,22 +249,19 @@ setup_isurf(struct IsdnCard *card)
 			return(0);
 		}
 #else
-		printk(KERN_WARNING "HiSax: %s port/mem not set\n",
-			CardType[card->typ]);
+		printk(KERN_WARNING "HiSax: Siemens I-Surf port/mem not set\n");
 		return (0);
 #endif
 	}
 	if (!request_region(cs->hw.isurf.reset, 1, "isurf isdn")) {
 		printk(KERN_WARNING
-			"HiSax: %s config port %x already in use\n",
-			CardType[card->typ],
+			"HiSax: Siemens I-Surf config port %x already in use\n",
 			cs->hw.isurf.reset);
 			return (0);
 	}
 	if (!request_region(cs->hw.isurf.phymem, ISURF_IOMEM_SIZE, "isurf iomem")) {
-		printk(KERN_WARNING
-			"HiSax: %s memory region %lx-%lx already in use\n",
-			CardType[card->typ],
+		printk(KERN_WARNING "HiSax: Siemens I-Surf memory region "
+			"%lx-%lx already in use\n",
 			cs->hw.isurf.phymem,
 			cs->hw.isurf.phymem + ISURF_IOMEM_SIZE);
 		release_region(cs->hw.isurf.reset, 1);
diff --git a/drivers/isdn/hisax/ix1_micro.c b/drivers/isdn/hisax/ix1_micro.c
index 252d79d..2d18d4f 100644
--- a/drivers/isdn/hisax/ix1_micro.c
+++ b/drivers/isdn/hisax/ix1_micro.c
@@ -24,7 +24,6 @@
 #include "hscx.h"
 #include "isdnl1.h"
 
-extern const char *CardType[];
 static const char *ix1_revision = "$Revision: 2.12.2.4 $";
 
 #define byteout(addr,val) outb(val,addr)
@@ -288,15 +287,15 @@ setup_ix1micro(struct IsdnCard *card)
 	if (cs->hw.ix1.cfg_reg) {
 		if (!request_region(cs->hw.ix1.cfg_reg, 4, "ix1micro cfg")) {
 			printk(KERN_WARNING
-			  "HiSax: %s config port %x-%x already in use\n",
-			       CardType[card->typ],
+			  "HiSax: ITK ix1-micro Rev.2 config port "
+			  "%x-%x already in use\n",
 			       cs->hw.ix1.cfg_reg,
 			       cs->hw.ix1.cfg_reg + 4);
 			return (0);
 		}
 	}
-	printk(KERN_INFO "HiSax: %s config irq:%d io:0x%X\n",
-		CardType[cs->typ], cs->irq, cs->hw.ix1.cfg_reg);
+	printk(KERN_INFO "HiSax: ITK ix1-micro Rev.2 config irq:%d io:0x%X\n",
+		cs->irq, cs->hw.ix1.cfg_reg);
 	setup_isac(cs);
 	cs->readisac = &ReadISAC;
 	cs->writeisac = &WriteISAC;
diff --git a/drivers/isdn/hisax/mic.c b/drivers/isdn/hisax/mic.c
index a81d175..2539430 100644
--- a/drivers/isdn/hisax/mic.c
+++ b/drivers/isdn/hisax/mic.c
@@ -16,8 +16,6 @@
 #include "hscx.h"
 #include "isdnl1.h"
 
-extern const char *CardType[];
-
 static const char *mic_revision = "$Revision: 1.12.2.4 $";
 
 #define byteout(addr,val) outb(val,addr)
@@ -210,8 +208,7 @@ setup_mic(struct IsdnCard *card)
 
 	if (!request_region(cs->hw.mic.cfg_reg, bytecnt, "mic isdn")) {
 		printk(KERN_WARNING
-		       "HiSax: %s config port %x-%x already in use\n",
-		       CardType[card->typ],
+		       "HiSax: ith mic config port %x-%x already in use\n",
 		       cs->hw.mic.cfg_reg,
 		       cs->hw.mic.cfg_reg + bytecnt);
 		return (0);
diff --git a/drivers/isdn/hisax/netjet.h b/drivers/isdn/hisax/netjet.h
index 4d89d3e..68e504d 100644
--- a/drivers/isdn/hisax/netjet.h
+++ b/drivers/isdn/hisax/netjet.h
@@ -12,8 +12,6 @@
  *
  */
 
-extern const char *CardType[];
-
 #define byteout(addr,val) outb(val,addr)
 #define bytein(addr) inb(addr)
 
diff --git a/drivers/isdn/hisax/niccy.c b/drivers/isdn/hisax/niccy.c
index bd99211..421b8e6 100644
--- a/drivers/isdn/hisax/niccy.c
+++ b/drivers/isdn/hisax/niccy.c
@@ -21,7 +21,6 @@
 #include <linux/pci.h>
 #include <linux/isapnp.h>
 
-extern const char *CardType[];
 static const char *niccy_revision = "$Revision: 1.21.2.4 $";
 
 #define byteout(addr,val) outb(val,addr)
@@ -284,14 +283,14 @@ int __devinit setup_niccy(struct IsdnCard *card)
 		cs->subtyp = NICCY_PNP;
 		cs->irq = card->para[0];
 		if (!request_region(cs->hw.niccy.isac, 2, "niccy data")) {
-			printk(KERN_WARNING "HiSax: %s data port %x-%x "
-				"already in use\n", CardType[card->typ],
+			printk(KERN_WARNING "HiSax: NICCY data port %x-%x "
+				"already in use\n",
 				cs->hw.niccy.isac, cs->hw.niccy.isac + 1);
 			return 0;
 		}
 		if (!request_region(cs->hw.niccy.isac_ale, 2, "niccy addr")) {
-			printk(KERN_WARNING "HiSax: %s address port %x-%x "
-				"already in use\n", CardType[card->typ],
+			printk(KERN_WARNING "HiSax: NICCY address port %x-%x "
+				"already in use\n",
 				cs->hw.niccy.isac_ale,
 				cs->hw.niccy.isac_ale + 1);
 			release_region(cs->hw.niccy.isac, 2);
@@ -339,15 +338,13 @@ int __devinit setup_niccy(struct IsdnCard *card)
 		cs->hw.niccy.hscx_ale = pci_ioaddr + HSCX_PCI_ADDR;
 		if (!request_region(cs->hw.niccy.isac, 4, "niccy")) {
 			printk(KERN_WARNING
-			       "HiSax: %s data port %x-%x already in use\n",
-			       CardType[card->typ],
+			       "HiSax: NICCY data port %x-%x already in use\n",
 			       cs->hw.niccy.isac, cs->hw.niccy.isac + 4);
 			return 0;
 		}
 		if (!request_region(cs->hw.niccy.cfg_reg, 0x40, "niccy pci")) {
 			printk(KERN_WARNING
-			       "HiSax: %s pci port %x-%x already in use\n",
-			       CardType[card->typ],
+			       "HiSax: NICCY pci port %x-%x already in use\n",
 			       cs->hw.niccy.cfg_reg,
 			       cs->hw.niccy.cfg_reg + 0x40);
 			release_region(cs->hw.niccy.isac, 4);
@@ -359,8 +356,8 @@ int __devinit setup_niccy(struct IsdnCard *card)
 		return 0;
 #endif				/* CONFIG_PCI_LEGACY */
 	}
-	printk(KERN_INFO "HiSax: %s %s config irq:%d data:0x%X ale:0x%X\n",
-		CardType[cs->typ], (cs->subtyp == 1) ? "PnP" : "PCI",
+	printk(KERN_INFO "HiSax: NICCY %s config irq:%d data:0x%X ale:0x%X\n",
+		(cs->subtyp == 1) ? "PnP" : "PCI",
 		cs->irq, cs->hw.niccy.isac, cs->hw.niccy.isac_ale);
 	setup_isac(cs);
 	cs->readisac = &ReadISAC;
diff --git a/drivers/isdn/hisax/nj_s.c b/drivers/isdn/hisax/nj_s.c
index a895dfe..8d36ccc 100644
--- a/drivers/isdn/hisax/nj_s.c
+++ b/drivers/isdn/hisax/nj_s.c
@@ -235,8 +235,7 @@ static int __devinit njs_cs_init_rest(struct IsdnCard *card,
 		cs->subtyp ? "TJ320" : "TJ300", cs->hw.njet.base, cs->irq);
 	if (!request_region(cs->hw.njet.base, bytecnt, "netjet-s isdn")) {
 		printk(KERN_WARNING
-		       "HiSax: %s config port %#lx-%#lx already in use\n",
-		       CardType[card->typ],
+		       "HiSax: NETjet-S config port %#lx-%#lx already in use\n",
 		       cs->hw.njet.base,
 		       cs->hw.njet.base + bytecnt);
 		return (0);
diff --git a/drivers/isdn/hisax/nj_u.c b/drivers/isdn/hisax/nj_u.c
index f017d38..d306c94 100644
--- a/drivers/isdn/hisax/nj_u.c
+++ b/drivers/isdn/hisax/nj_u.c
@@ -197,8 +197,8 @@ static int __devinit nju_cs_init_rest(struct IsdnCard *card,
 		cs->hw.njet.base, cs->irq);
 	if (!request_region(cs->hw.njet.base, bytecnt, "netspider-u isdn")) {
 		printk(KERN_WARNING
-		       "HiSax: %s config port %#lx-%#lx already in use\n",
-		       CardType[card->typ],
+		       "HiSax: NETspider-U config port %#lx-%#lx "
+		       "already in use\n",
 		       cs->hw.njet.base,
 		       cs->hw.njet.base + bytecnt);
 		return (0);
diff --git a/drivers/isdn/hisax/s0box.c b/drivers/isdn/hisax/s0box.c
index 150ef68..16d00b5 100644
--- a/drivers/isdn/hisax/s0box.c
+++ b/drivers/isdn/hisax/s0box.c
@@ -16,7 +16,6 @@
 #include "hscx.h"
 #include "isdnl1.h"
 
-extern const char *CardType[];
 static const char *s0box_revision = "$Revision: 2.6.2.4 $";
 
 static inline void
@@ -231,19 +230,15 @@ setup_s0box(struct IsdnCard *card)
 	cs->hw.teles3.hscxfifo[1] = cs->hw.teles3.hscx[1] + 0x3e;
 	cs->irq = card->para[0];
 	if (!request_region(cs->hw.teles3.cfg_reg,8, "S0Box parallel I/O")) {
-		printk(KERN_WARNING
-		       "HiSax: %s ports %x-%x already in use\n",
-		       CardType[cs->typ],
+		printk(KERN_WARNING "HiSax: S0Box ports %x-%x already in use\n",
                        cs->hw.teles3.cfg_reg,
                        cs->hw.teles3.cfg_reg + 7);
 		return 0;
 	}
-	printk(KERN_INFO
-		"HiSax: %s config irq:%d isac:0x%x  cfg:0x%x\n",
-		CardType[cs->typ], cs->irq,
+	printk(KERN_INFO "HiSax: S0Box config irq:%d isac:0x%x  cfg:0x%x\n",
+		cs->irq,
 		cs->hw.teles3.isac, cs->hw.teles3.cfg_reg);
-	printk(KERN_INFO
-		"HiSax: hscx A:0x%x  hscx B:0x%x\n",
+	printk(KERN_INFO "HiSax: hscx A:0x%x  hscx B:0x%x\n",
 		cs->hw.teles3.hscx[0], cs->hw.teles3.hscx[1]);
 	setup_isac(cs);
 	cs->readisac = &ReadISAC;
diff --git a/drivers/isdn/hisax/saphir.c b/drivers/isdn/hisax/saphir.c
index c99b166..b34a81d 100644
--- a/drivers/isdn/hisax/saphir.c
+++ b/drivers/isdn/hisax/saphir.c
@@ -18,7 +18,6 @@
 #include "hscx.h"
 #include "isdnl1.h"
 
-extern const char *CardType[];
 static char *saphir_rev = "$Revision: 1.10.2.4 $";
 
 #define byteout(addr,val) outb(val,addr)
@@ -260,15 +259,14 @@ setup_saphir(struct IsdnCard *card)
 	cs->irq = card->para[0];
 	if (!request_region(cs->hw.saphir.cfg_reg, 6, "saphir")) {
 		printk(KERN_WARNING
-			"HiSax: %s config port %x-%x already in use\n",
-			CardType[card->typ],
+			"HiSax: HST Saphir config port %x-%x already in use\n",
 			cs->hw.saphir.cfg_reg,
 			cs->hw.saphir.cfg_reg + 5);
 		return (0);
 	}
 
-	printk(KERN_INFO "HiSax: %s config irq:%d io:0x%X\n",
-		CardType[cs->typ], cs->irq, cs->hw.saphir.cfg_reg);
+	printk(KERN_INFO "HiSax: HST Saphir config irq:%d io:0x%X\n",
+	       cs->irq, cs->hw.saphir.cfg_reg);
 
 	setup_isac(cs);
 	cs->hw.saphir.timer.function = (void *) SaphirWatchDog;
diff --git a/drivers/isdn/hisax/sportster.c b/drivers/isdn/hisax/sportster.c
index 0220950..0a53759 100644
--- a/drivers/isdn/hisax/sportster.c
+++ b/drivers/isdn/hisax/sportster.c
@@ -18,7 +18,6 @@
 #include "hscx.h"
 #include "isdnl1.h"
 
-extern const char *CardType[];
 static const char *sportster_revision = "$Revision: 1.16.2.4 $";
 
 #define byteout(addr,val) outb(val,addr)
@@ -192,9 +191,9 @@ get_io_range(struct IsdnCardState *cs)
 	for (i=0;i<64;i++) {
 		adr = cs->hw.spt.cfg_reg + i *1024;
 		if (!request_region(adr, 8, "sportster")) {
-			printk(KERN_WARNING
-				"HiSax: %s config port %x-%x already in use\n",
-				CardType[cs->typ], adr, adr + 8);
+			printk(KERN_WARNING "HiSax: USR Sportster config port "
+				"%x-%x already in use\n",
+				adr, adr + 8);
 			break;
 		} 
 	}
@@ -247,8 +246,8 @@ setup_sportster(struct IsdnCard *card)
 			printk(KERN_WARNING "Sportster: wrong IRQ\n");
 			return(0);
 	}
-	printk(KERN_INFO "HiSax: %s config irq:%d cfg:0x%X\n",
-		CardType[cs->typ], cs->irq, cs->hw.spt.cfg_reg);
+	printk(KERN_INFO "HiSax: USR Sportster config irq:%d cfg:0x%X\n",
+		cs->irq, cs->hw.spt.cfg_reg);
 	setup_isac(cs);
 	cs->readisac = &ReadISAC;
 	cs->writeisac = &WriteISAC;
diff --git a/drivers/isdn/hisax/teleint.c b/drivers/isdn/hisax/teleint.c
index 0909662..b0ce4ae 100644
--- a/drivers/isdn/hisax/teleint.c
+++ b/drivers/isdn/hisax/teleint.c
@@ -16,8 +16,6 @@
 #include "hfc_2bs0.h"
 #include "isdnl1.h"
 
-extern const char *CardType[];
-
 static const char *TeleInt_revision = "$Revision: 1.16.2.5 $";
 
 #define byteout(addr,val) outb(val,addr)
@@ -286,8 +284,7 @@ setup_TeleInt(struct IsdnCard *card)
 	init_timer(&cs->hw.hfc.timer);
 	if (!request_region(cs->hw.hfc.addr, 2, "TeleInt isdn")) {
 		printk(KERN_WARNING
-		       "HiSax: %s config port %x-%x already in use\n",
-		       CardType[card->typ],
+		       "HiSax: TeleInt config port %x-%x already in use\n",
 		       cs->hw.hfc.addr,
 		       cs->hw.hfc.addr + 2);
 		return (0);
diff --git a/drivers/isdn/hisax/telespci.c b/drivers/isdn/hisax/telespci.c
index 4393003..28b08de 100644
--- a/drivers/isdn/hisax/telespci.c
+++ b/drivers/isdn/hisax/telespci.c
@@ -19,7 +19,6 @@
 #include "isdnl1.h"
 #include <linux/pci.h>
 
-extern const char *CardType[];
 static const char *telespci_revision = "$Revision: 2.23.2.3 $";
 
 #define ZORAN_PO_RQ_PEN	0x02000000
@@ -329,8 +328,8 @@ setup_telespci(struct IsdnCard *card)
 	/* writel(0x00800000, cs->hw.teles0.membase + 0x200); */
 
 	printk(KERN_INFO
-	       "HiSax: %s config irq:%d mem:%p\n",
-	       CardType[cs->typ], cs->irq,
+	       "HiSax: Teles PCI config irq:%d mem:%p\n",
+	       cs->irq,
 	       cs->hw.teles0.membase);
 
 	setup_isac(cs);
diff --git a/drivers/isdn/hisax/w6692.c b/drivers/isdn/hisax/w6692.c
index 39129b9..bb1c8dd 100644
--- a/drivers/isdn/hisax/w6692.c
+++ b/drivers/isdn/hisax/w6692.c
@@ -38,8 +38,6 @@ static const PCI_ENTRY id_list[] =
 #define W6692_DYNALINK 1
 #define W6692_USR      2
 
-extern const char *CardType[];
-
 static const char *w6692_revision = "$Revision: 1.18.2.4 $";
 
 #define DBUSY_TIMER_VALUE 80

ISDN testing (was: [git patches] ISDN cleanups (modularization prep))

[Frans Pop]

(Dropping private CCs.)

Jeff Garzik wrote:
> - "isn't ISDN dead?"  Apparently, no it's not :)
> 
> - "don't waste your time, drop these drivers and get mISDN into the
>   kernel instead" -- based on discussion it does sound like mISDN should
>   get into the kernel.  Until these drivers are dropped, however, my
>   patches remain appropriate.
> 
> And in case people are curious, #isdn-pci isn't ready for upstream yet
> for two main reasons: (1) two drivers remain unconverted, and (2) none
> of the changes have been tested, even once.

I have three different ISDN cards. However, since I've had ADSL since about 
3 years (on top of ISDN), I don't actually use it anymore.

I could be persuaded to do some testing though, or to donate some or all of 
my cards for that purpose.

For testing I'd at least need somewhere to connect to, preferably in NL.
I have no idea if I could still connect to my previous provider using ISDN. 
The account was transferred to a different provider about 2 years ago, but 
the mail accounts with the old provider are still active.

The cards I have are:
- Teles 16.3 ISA -- worked way back when with Debian Woody
- ITK ix1-micro 2.1 ISA
- Eicon Diva 2.01 PCI

I also have a test box I can put them in :-)

Looks like I can still at least successfully load the hisax driver with 
2.6.18 for the Teles card. For the last two I'm not sure if I ever had them 
working with Linux (or even if I tried).

Cheers,
FJP

Re: ISDN testing

[Jeff Garzik]

Frans Pop wrote:
> (Dropping private CCs.)
> 
> Jeff Garzik wrote:
>> - "isn't ISDN dead?"  Apparently, no it's not :)
>>
>> - "don't waste your time, drop these drivers and get mISDN into the
>>   kernel instead" -- based on discussion it does sound like mISDN should
>>   get into the kernel.  Until these drivers are dropped, however, my
>>   patches remain appropriate.
>>
>> And in case people are curious, #isdn-pci isn't ready for upstream yet
>> for two main reasons: (1) two drivers remain unconverted, and (2) none
>> of the changes have been tested, even once.
> 
> I have three different ISDN cards. However, since I've had ADSL since about 
> 3 years (on top of ISDN), I don't actually use it anymore.
> 
> I could be persuaded to do some testing though, or to donate some or all of 
> my cards for that purpose.
> 
> For testing I'd at least need somewhere to connect to, preferably in NL.
> I have no idea if I could still connect to my previous provider using ISDN. 
> The account was transferred to a different provider about 2 years ago, but 
> the mail accounts with the old provider are still active.
> 
> The cards I have are:
> - Teles 16.3 ISA -- worked way back when with Debian Woody
> - ITK ix1-micro 2.1 ISA
> - Eicon Diva 2.01 PCI
> 
> I also have a test box I can put them in :-)
> 
> Looks like I can still at least successfully load the hisax driver with 
> 2.6.18 for the Teles card. For the last two I'm not sure if I ever had them 
> working with Linux (or even if I tried).

Honestly, a module-load test on real hardware can give us some useful 
information, even if it's not connected to anything.

My changes are primarily in the probe-the-hardware area, so successfully 
making it past that new code would indeed be a useful data point.

Check out the kernel found in 'isdn-pci' branch of
git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/misc-2.6.git

Any help is certainly appreciated!

	Jeff

Re: ISDN testing

[Frans Pop]

On Monday 21 April 2008, Jeff Garzik wrote:
> Frans Pop wrote:
> > The cards I have are:
> > - Teles 16.3 ISA -- worked way back when with Debian Woody
> > - ITK ix1-micro 2.1 ISA
> > - Eicon Diva 2.01 PCI
>
> Honestly, a module-load test on real hardware can give us some useful
> information, even if it's not connected to anything.
>
> My changes are primarily in the probe-the-hardware area, so successfully
> making it past that new code would indeed be a useful data point.
>
> Check out the kernel found in 'isdn-pci' branch of
> git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/misc-2.6.git

Turns out all 3 cards are supported by the same basic driver: hisax, so
I can only test for the one driver. The Eicon Diva uses an additional
module though: hisaxdiva.

I've put the last 2 cards in one machine and (after some fiddling with
jumpers and interrupts and IO ports :-) both are detected perfectly with
2.6.24 when hisax is modprobed.

Advantage of using the two cards is that it covers both PCI and ISA.
The PCI card gets configured automagically, while the ISA card requires
module parameters for io/irq.

I've compiled a kernel based on your 'isdn-pci' branch, and it boots fine.

But when I tried to modprobe hisax, I got:
# modprobe hisax
FATAL: Module hisax not found.

Cause seems to be:
# depmod -a
WARNING: Module /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/hisaxdiva.ko 
ignored, due to loop
WARNING: Module /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/libhisax.ko 
ignored, due to loop
WARNING: Loop 
detected: /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/hisax.ko needs 
libhisax.ko which needs hisax.ko again!
WARNING: Module /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/hisax.ko 
ignored, due to loop

If needed I can also try with vanilla 2.6.25 to get a bit more current
baseline test than I now have with 2.6.24.

Cheers,
FJP

Re: ISDN testing

[Jeff Garzik]

Frans Pop wrote:
> On Monday 21 April 2008, Jeff Garzik wrote:
>> Frans Pop wrote:
>>> The cards I have are:
>>> - Teles 16.3 ISA -- worked way back when with Debian Woody
>>> - ITK ix1-micro 2.1 ISA
>>> - Eicon Diva 2.01 PCI
>> Honestly, a module-load test on real hardware can give us some useful
>> information, even if it's not connected to anything.
>>
>> My changes are primarily in the probe-the-hardware area, so successfully
>> making it past that new code would indeed be a useful data point.
>>
>> Check out the kernel found in 'isdn-pci' branch of
>> git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/misc-2.6.git
> 
> Turns out all 3 cards are supported by the same basic driver: hisax, so
> I can only test for the one driver. The Eicon Diva uses an additional
> module though: hisaxdiva.
> 
> I've put the last 2 cards in one machine and (after some fiddling with
> jumpers and interrupts and IO ports :-) both are detected perfectly with
> 2.6.24 when hisax is modprobed.
> 
> Advantage of using the two cards is that it covers both PCI and ISA.
> The PCI card gets configured automagically, while the ISA card requires
> module parameters for io/irq.
> 
> I've compiled a kernel based on your 'isdn-pci' branch, and it boots fine.
> 
> But when I tried to modprobe hisax, I got:
> # modprobe hisax
> FATAL: Module hisax not found.
> 
> Cause seems to be:
> # depmod -a
> WARNING: Module /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/hisaxdiva.ko 
> ignored, due to loop
> WARNING: Module /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/libhisax.ko 
> ignored, due to loop
> WARNING: Loop 
> detected: /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/hisax.ko needs 
> libhisax.ko which needs hisax.ko again!
> WARNING: Module /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/hisax.ko 
> ignored, due to loop

Yeah, that's what my patchset does -- splits the "hisax" driver into 
multiple smaller drivers.  Thus, "modprobe hisax" produces the results 
that you are seeing.

You'll need to determine which newly-created kernel module applies to 
your hardware.  Let me know if you need help with that.  The new drivers 
should be in /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/ 
provided that you enabled them in the kernel configuration.

	Jeff

Re: ISDN testing

[Andi Kleen]

Jeff Garzik <jeff@garzik.org> writes:
>
> Yeah, that's what my patchset does -- splits the "hisax" driver into
> multiple smaller drivers.  Thus, "modprobe hisax" produces the results
> that you are seeing.

That will break a lot of user's setup.

> You'll need to determine which newly-created kernel module applies to
> your hardware.  Let me know if you need help with that.  The new
> drivers should be in
> /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/ provided that you
> enabled them in the kernel configuration.

One way to avoid that issue would be to continue having a dummy
"hisax" module that just depends on all the other drivers with 
a symbol reference.
 
That could be optional with some compat CONFIG.

-Andi

Re: ISDN testing

[Jeff Garzik]

Andi Kleen wrote:
> Jeff Garzik <jeff@garzik.org> writes:
>> Yeah, that's what my patchset does -- splits the "hisax" driver into
>> multiple smaller drivers.  Thus, "modprobe hisax" produces the results
>> that you are seeing.
> 
> That will break a lot of user's setup.

At present, yes :/  Yet another why it's not upstream :)


>> You'll need to determine which newly-created kernel module applies to
>> your hardware.  Let me know if you need help with that.  The new
>> drivers should be in
>> /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/ provided that you
>> enabled them in the kernel configuration.
> 
> One way to avoid that issue would be to continue having a dummy
> "hisax" module that just depends on all the other drivers with 
> a symbol reference.

I'm definitely open to any solution like this.  Unfortunately (as Frans, 
I think, pointed out) the hisax module routinely depends on critical 
setup info being passed as module options, mainly for non-PCI hardware.

One solution might be to create a dummy hisax module that calls 
foo_module_init() and foo_module_exit(), and stitch them together in an 
unconventional way.

	Jeff

Re: [git patches] ISDN cleanups (modularization prep) (was: ISDN testing)

[Frans Pop]

(Let's change the subject back to what it was :-)

Jeff Garzik wrote:
> Frans Pop wrote:
>> But when I tried to modprobe hisax, I got:
>> # modprobe hisax
>> FATAL: Module hisax not found.
>> 
>> Cause seems to be:
>> # depmod -a
>> WARNING: Module
>> /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/hisaxdiva.ko ignored,
>> due to loop WARNING: Module
>> /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/libhisax.ko ignored,
>> due to loop WARNING: Loop
>> detected: /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/hisax.ko
>> needs libhisax.ko which needs hisax.ko again!
>> WARNING: Module
>> /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/hisax.ko ignored, due
>> to loop
> 
> Yeah, that's what my patchset does -- splits the "hisax" driver into
> multiple smaller drivers.  Thus, "modprobe hisax" produces the results
> that you are seeing.

Hmmm. Isn't that going to break things for existing users?

For example, I currently have this in /etc/modprobe.d/hisax to get the two 
cards to work correctly:
options hisax type=9,11 protocol=2,2 irq=10 io=0x398

This effectively translates to:
- ITK card:  type=9  protocol=2 irq=10 io=0x398
- Diva card: type=11 protocol=2

To what would I have to change this?

I suspect there will at least have to be a major update in Documentation to 
tell people how options should now be set.

> You'll need to determine which newly-created kernel module applies to
> your hardware.  Let me know if you need help with that.  The new drivers
> should be in /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/
> provided that you enabled them in the kernel configuration.

I think that I have the correct config (unset options omitted):
CONFIG_ISDN=m
CONFIG_ISDN_I4L=m
# ISDN feature submodules
# ISDN4Linux hardware drivers
# Passive cards
CONFIG_ISDN_DRV_HISAX=m
# D-channel protocol features
CONFIG_HISAX_EURO=y
CONFIG_HISAX_MAX_CARDS=8
# HiSax supported cards
CONFIG_HISAX_16_3=y
CONFIG_HISAX_IX1MICROR2=y
CONFIG_HISAX_DIEHLDIVA=m

Note that HISAX_16_3 and HISAX_IX1MICROR2 are y/n config options; making 
those modular is not allowed.

This results in the following modules under drivers/isdn:
/lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax:
hisaxdiva.ko  hisax.ko  libhisax.ko
/lib/modules/2.6.25-isdn/kernel/drivers/isdn/i4l:
isdn.ko

So AFAICT, I should still be loading hisax for the ITK card.

Also, I doubt that depmod is supposed to fail like the way I saw. Seems to 
me that the "loop" errors during depmod are the cause of the modules not 
being loadable.
Note that 'modprobe hisaxdiva' fails with the exact same message as for 
hisax and that depmod even explicitly says that the modules are "ignored"!

Cheers,
FJP

Re: [git patches] ISDN cleanups (modularization prep)

[Jeff Garzik]

Frans Pop wrote:
> (Let's change the subject back to what it was :-)
> 
> Jeff Garzik wrote:
>> Frans Pop wrote:
>>> But when I tried to modprobe hisax, I got:
>>> # modprobe hisax
>>> FATAL: Module hisax not found.
>>>
>>> Cause seems to be:
>>> # depmod -a
>>> WARNING: Module
>>> /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/hisaxdiva.ko ignored,
>>> due to loop WARNING: Module
>>> /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/libhisax.ko ignored,
>>> due to loop WARNING: Loop
>>> detected: /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/hisax.ko
>>> needs libhisax.ko which needs hisax.ko again!
>>> WARNING: Module
>>> /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/hisax.ko ignored, due
>>> to loop
>> Yeah, that's what my patchset does -- splits the "hisax" driver into
>> multiple smaller drivers.  Thus, "modprobe hisax" produces the results
>> that you are seeing.
> 
> Hmmm. Isn't that going to break things for existing users?

At present, yes.  Ultimately I want to find a solution that doesn't 
break existing users, but I am still at the early "trying to see if it 
works at all" stage.


> For example, I currently have this in /etc/modprobe.d/hisax to get the two 
> cards to work correctly:
> options hisax type=9,11 protocol=2,2 irq=10 io=0x398
> 
> This effectively translates to:
> - ITK card:  type=9  protocol=2 irq=10 io=0x398
> - Diva card: type=11 protocol=2
> 
> To what would I have to change this?

Two changes:

1) 'type' is now automatically deduced based on your choice of driver, 
so it is not needed anymore.

2) The module options are passed to the specific module, rather than the 
hisax module.

'protocol', 'irq' and 'io' module options should remain completely 
compatible with existing setups.


>> You'll need to determine which newly-created kernel module applies to
>> your hardware.  Let me know if you need help with that.  The new drivers
>> should be in /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/
>> provided that you enabled them in the kernel configuration.
> 
> I think that I have the correct config (unset options omitted):
> CONFIG_ISDN=m
> CONFIG_ISDN_I4L=m
> # ISDN feature submodules
> # ISDN4Linux hardware drivers
> # Passive cards
> CONFIG_ISDN_DRV_HISAX=m
> # D-channel protocol features
> CONFIG_HISAX_EURO=y
> CONFIG_HISAX_MAX_CARDS=8
> # HiSax supported cards
> CONFIG_HISAX_16_3=y
> CONFIG_HISAX_IX1MICROR2=y
> CONFIG_HISAX_DIEHLDIVA=m
> 
> Note that HISAX_16_3 and HISAX_IX1MICROR2 are y/n config options; making 
> those modular is not allowed.
> 
> This results in the following modules under drivers/isdn:
> /lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax:
> hisaxdiva.ko  hisax.ko  libhisax.ko
> /lib/modules/2.6.25-isdn/kernel/drivers/isdn/i4l:
> isdn.ko
> 
> So AFAICT, I should still be loading hisax for the ITK card.
> 
> Also, I doubt that depmod is supposed to fail like the way I saw. Seems to 
> me that the "loop" errors during depmod are the cause of the modules not 
> being loadable.
> Note that 'modprobe hisaxdiva' fails with the exact same message as for 
> hisax and that depmod even explicitly says that the modules are "ignored"!

Yeah, it looks like the ITK card is still served by the main hisax 
module (not all were converted, just the ones related to PCI).

Maybe try manually modprobing libhisax and then hisaxdiva.

Also try disabling the ITK support in Kconfig and see if that at least 
permits you to load the hisaxdiva driver.


My apologies for all the caveats, this is quite raw code.  I appreciate 
your taking a look at this!

	Jeff

Re: [git patches] ISDN cleanups (modularization prep)

[Frans Pop]

On Friday 25 April 2008, Jeff Garzik wrote:
> Frans Pop wrote:
> > Also, I doubt that depmod is supposed to fail like the way I saw. Seems
> > to me that the "loop" errors during depmod are the cause of the modules
> > not being loadable.
> > Note that 'modprobe hisaxdiva' fails with the exact same message as for
> > hisax and that depmod even explicitly says that the modules are
> > "ignored"!
>
> Yeah, it looks like the ITK card is still served by the main hisax
> module (not all were converted, just the ones related to PCI).
>
> Maybe try manually modprobing libhisax and then hisaxdiva.

Even manually insmod'ing them does not work: you'll really need to resolve 
those interdependencies that cause the dependency loops first!

> Also try disabling the ITK support in Kconfig and see if that at least
> permits you to load the hisaxdiva driver.

Using a kernel with both Teles and ITK support unselected, depmod -a no 
longer complains.

First I left the old modprobe config:
options        hisax type=9,11 protocol=2,2 irq=10 io=0x398

When I modprobe hisaxdiva with that, I get:
# modprobe hisaxdiva
WARNING: Error inserting hisax 
(/lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/hisax.ko): No such 
device
WARNING: Error inserting libhisax 
(/lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/libhisax.ko): Unknown 
symbol in module, or unknown parameter (see dmesg)
FATAL: Error inserting hisaxdiva 
(/lib/modules/2.6.25-isdn/kernel/drivers/isdn/hisax/hisaxdiva.ko): Unknown 
symbol in module, or unknown parameter (see dmesg)

fjp@selinux:~$ dmesg
[...]
HiSax: Linux Driver for passive ISDN cards
HiSax: Version 3.5 (module)
HiSax: Layer1 Revision 2.46.2.5
HiSax: Layer2 Revision 2.30.2.4
HiSax: TeiMgr Revision 2.20.2.3
HiSax: Layer3 Revision 2.22.2.3
HiSax: LinkLayer Revision 2.59.2.4
HiSax: Total 2 cards defined
HiSax: Card 1 Protocol EDSS1 Id=HiSax (0)
HiSax: Support for ITK ix1-micro Rev.2 Card not selected
HiSax: Card ITK ix1-micro Rev.2 not installed !
HiSax: Card 1 Protocol EDSS1 Id=HiSax (0)
HiSax: Support for Eicon.Diehl Diva Card has moved to separate PCI driver 
module
HiSax: Card Eicon.Diehl Diva not installed !
libhisax: Unknown symbol LogFrame
libhisax: Unknown symbol Logl2Frame
libhisax: Unknown symbol DChannel_proc_xmt
libhisax: Unknown symbol QuickHex
libhisax: Unknown symbol DChannel_proc_rcv
libhisax: Unknown symbol l1_msg_b
libhisax: Unknown symbol setstack_l1_B
libhisax: Unknown symbol dlogframe
libhisax: Unknown symbol l1_msg
libhisax: Unknown symbol setstack_manager
libhisax: Unknown symbol debugl1
hisaxdiva: Unknown symbol LogFrame
hisaxdiva: Unknown symbol HiSax_closecard
hisaxdiva: Unknown symbol Logl2Frame
hisaxdiva: Unknown symbol HscxVersion
hisaxdiva: Unknown symbol QuickHex
hisaxdiva: Unknown symbol hisax_init_hotplug
hisaxdiva: Unknown symbol setup_isac
hisaxdiva: Unknown symbol l1_msg_b
hisaxdiva: Unknown symbol setstack_l1_B
hisaxdiva: Unknown symbol dlogframe
hisaxdiva: Unknown symbol isac_interrupt
hisaxdiva: Unknown symbol ISACVersion
hisaxdiva: Unknown symbol inithscxisac
hisaxdiva: Unknown symbol HiSax_getrev
hisaxdiva: Unknown symbol setstack_manager
hisaxdiva: Unknown symbol hisax_pci_remove_one
hisaxdiva: Unknown symbol debugl1

lsmod does not show any isdn/hisax modules loaded after that.


I then changed the modprobe config to just:
options         hisaxdiva protocol=2

When I 'modprobe hisaxdiva' with that, the result is a lot better:
ISDN subsystem Rev: 1.1.2.3/1.1.2.3/1.1.2.2/none/none/1.1.2.2 loaded
HiSax: Linux Driver for passive ISDN cards
HiSax: Version 3.5 (module)
HiSax: Layer1 Revision 2.46.2.5
HiSax: Layer2 Revision 2.30.2.4
HiSax: TeiMgr Revision 2.20.2.3
HiSax: Layer3 Revision 2.22.2.3
HiSax: LinkLayer Revision 2.59.2.4
libhisax: module license 'unspecified' taints kernel.
HiSax: Card 1 Protocol EDSS1 Id=HiSax (0)
ACPI: PCI Interrupt Link [LNKC] enabled at IRQ 11
ACPI: PCI Interrupt 0000:00:0f.0[A] -> Link [LNKC] -> GSI 11 (level, low) -> 
IRQ 11
HiSax: Eicon.Diehl Diva driver Rev. 1.33.2.6
Diva: IPAC PCI card configured at 0xc8cc2000 IRQ 11
Diva: IPAC PCI space at 0xc8cc0000
Diva: IPAC version ff
Eicon.Diehl Diva: IRQ 11 count 746
Eicon.Diehl Diva: IRQ 11 count 752
HiSax: DSS1 Rev. 2.32.2.3
HiSax: 2 channels added
HiSax: MAX_WAITING_CALLS added
HiSax: if_command 6 called with invalid driverId 0!
HiSax: if_command 6 called with invalid driverId 0!

Issues I can see in this dmesg output:
- the module licence warning for libhisax
- I've never seen the last two lines before

Otherwise the driver seems to have loaded correctly and lsmod now shows:
Module                  Size  Used by
hisaxdiva              27564  0
libhisax               27872  1 hisaxdiva
hisax                 113924  2 hisaxdiva,libhisax
isdn                   70096  1 hisax
crc_ccitt               2144  1 libhisax

Cheers,
FJP

Re: ISDN testing (was: [git patches] ISDN cleanups (modularization prep))

[Jan Engelhardt]

On Monday 2008-04-21 02:05, Frans Pop wrote:
>
>For testing I'd at least need somewhere to connect to, preferably in NL.
>I have no idea if I could still connect to my previous provider using ISDN. 
>The account was transferred to a different provider about 2 years ago, but 
>the mail accounts with the old provider are still active.

At least over here in DE, ISPs provide login-less dialins that
are automatically accredited to the number that called it.