Rocketport driver fails if more than four cards are inserted
Inserting a fifth Rocketport card into a machines gives me the backtrace included below.
Looking at the source, it looks like there's two problems: the sController array only has four entries, even though CTL_SIZE is eight, and some pc104 initialisation is run even though we're not running on a pc104 platform. (Apparently the maximum number of controllers on pc104 is four.)
The patch (against 2.6.13.3) included below fixes both these problems. The other way to fix the latter problem would be to extend the pc104 (i.e. ISA) case to allow eight cards, but that seems somewhat less likely...
Unable to handle kernel NULL pointer dereference at virtual address 00000201 printing eip: dc863796 *pde = 00000000 Oops: 0000 [#1] SMP Modules linked in: rocket CPU: 0 EIP: 0060:[<dc863796>] Not tainted VLI EFLAGS: 00010246 (2.6.13.3)
EIP is at init_r_port+0x126/0x334 [rocket]
eax: 00000001 ebx: 00000000 ecx: 00000004 edx: 00000000
esi: 00000000 edi: 00000080 ebp: dc869c40 esp: db8abee8
ds: 007b es: 007b ss: 0068
Process modprobe (pid: 886, threadinfo=db8aa000 task=db813a40)
Stack: 00000000 00000000 00000001 dc869c7c dbbe0800 00000080 dc8029c2 00000004
00000000 00000000 dbe17800 dc866ae0 00000004 00000000 00000000 dbe17800
00000004 00000000 00000000 dbe178e8 00000010 00000008 00000000 00000000
Call Trace:
[<dc8029c2>] register_PCI+0x9c2/0xacc [rocket]
[<dc802aea>] init_PCI+0x1e/0x44 [rocket]
[<dc80316f>] init_module+0x293/0x2e4 [rocket]
[<c01311bd>] sys_init_module+0xd1/0x1cc
[<c0102ac9>] syscall_call+0x7/0xb
Code: 01 00 00 00 83 c0 0c 89 87 0c 01 00 00 89 87 10 01 00 00 8b 57 08 80 e6 cf 89 57 08 8b 4c 24 1c 8b 7c 24 14 8b 04 8d 60 99 86 dc <8b> 04 b8 3d a6 01 00 00 74 10 76 28 3d e5 01 00 00 74 17 eb 1f rocket.c renamed from rocket.c.~1~ +21 -11
@@ -206,6 +206,14 @@ {-1, -1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, {0, 0, 0, 0}, {0, 0, 0, 0}, {-1, -1, -1, -1}, {0, 0, 0, 0}}, {-1, -1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, {0, 0, 0, 0}, + {0, 0, 0, 0}, {-1, -1, -1, -1}, {0, 0, 0, 0}}, + {-1, -1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, {0, 0, 0, 0}, + {0, 0, 0, 0}, {-1, -1, -1, -1}, {0, 0, 0, 0}}, + {-1, -1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, {0, 0, 0, 0}, + {0, 0, 0, 0}, {-1, -1, -1, -1}, {0, 0, 0, 0}}, + {-1, -1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, {0, 0, 0, 0}, + {0, 0, 0, 0}, {-1, -1, -1, -1}, {0, 0, 0, 0}}, + {-1, -1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, {0, 0, 0, 0}, {0, 0, 0, 0}, {-1, -1, -1, -1}, {0, 0, 0, 0}} };
@@ -675,17 +683,19 @@init_waitqueue_head (&info->open_wait);init_waitqueue_head (&info->close_wait); info->flags &= ~ROCKET_MODE_MASK ; -switch (pc104[board][line]) { -case 422 : - info->flags |=ROCKET_MODE_RS422 ; -break ; -case 485 : - info->flags |=ROCKET_MODE_RS485 ; -break ; -case 232 : -default : - info->flags |=ROCKET_MODE_RS232 ; -break ; +if (ctlp->boardType ==ROCKET_TYPE_PC104 ) { +switch (pc104[board][line]) { +case 422 : + info->flags |=ROCKET_MODE_RS422 ; +break ; +case 485 : + info->flags |=ROCKET_MODE_RS485 ; +break ; +case 232 : +default : + info->flags |=ROCKET_MODE_RS232 ; +break ; + } } info->intmask =RXF_TRIG |TXFIFO_MT |SRC_INT |DELTA_CD |DELTA_CTS |DELTA_DSR ;
Signature · 2 lines
(domestic pets only, the antidote for overdose, milk.) [email protected] * Lars Magne Ingebrigtsen
Quoted 9 lines
Inserting a fifth Rocketport card into a machines gives me the backtrace included below.
Looking at the source, it looks like there's two problems: the sController array only has four entries, even though
CTL_SIZEis eight, and some pc104 initialisation is run even though we're not running on a pc104 platform. (Apparently the maximum number of controllers on pc104 is four.)
Quoted 4 lines
The patch (against 2.6.13.3) included below fixes both these problems. The other way to fix the latter problem would be to extend the pc104 (i.e. ISA) case to allow eight cards, but that seems somewhat less likely...
[snip]
They are in process of rewriting the driver (they wrote this to me on Sep 16):
<cite>
We have a driver in beta for 2.6 kernels. Eventually we do plan to update the main kernel with our changes, but after the driver is more stable and we have more time to work on it.
Here is a link to the beta if you are interested ftp://ftp.comtrol.com/contribs/Drivers/RPort/Linux/3.02%20Beta%202.4-2.6/rocketport-linux-3.02.tgz </cite>
The link is the same, so there is no new version, so I don't know the state of the rewriting now, we'll see...
regards,
--
Jiri Slaby www.fi.muni.cz/~xslaby
~\-/~ [email protected] ~\-/~
B67499670407CE62ACC8 22A032CC55C339D47A7E