Re: busdma/scsi trm(4) related panic
Ruslan Ermilov wrote:
On Wed, Aug 06, 2003 at 07:42:06AM -0600, Scott Long wrote:
I know what the problem is and I'm working on a patch right now.
I have the same problem with amd(4) and this commit to busdma_machdep.c.
Cheers,
Attached is an untested patch for amd(4). Please let me know if it
fixes your problem.
Scott
Index: amd.c
===
RCS file: /home/ncvs/src/sys/dev/amd/amd.c,v
retrieving revision 1.22
diff -u -r1.22 amd.c
--- amd.c 31 Jul 2003 16:55:44 - 1.22
+++ amd.c 9 Aug 2003 06:44:28 -
@@ -2115,12 +2115,26 @@
{
u_int16_t count, i;
struct amd_srb *psrb;
+ int error;
count = amd-SRBCount;
for (i = 0; i count; i++) {
psrb = (struct amd_srb *)amd-SRB_array[i];
psrb-TagNumber = i;
+
+ /*
+* Create the dmamap. This is no longer optional!
+*
+* XXX Since there is no detach method in this driver,
+* this does not get freed!
+*/
+ if ((error = bus_dmamap_create(amd-buffer_dmat, 0,
+ psrb-dmamap)) != 0) {
+ device_printf(amd-dev, Error %d creating buffer
+ dmamap!\n, error);
+ return;
+ }
TAILQ_INSERT_TAIL(amd-free_srbs, psrb, links);
}
}
___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: busdma/scsi trm(4) related panic
I know what the problem is and I'm working on a patch right now. Scott Jon Kuster wrote: On Wed, 2003-08-06 at 05:00, Craig Rodrigues wrote: On Tue, Aug 05, 2003 at 11:08:33PM -0700, Jon Kuster wrote: If there's anything you'd like me to type at the db prompt, just ask and I'll send the output. Send the backtrace, with the 't' command. Here you go. Once again, this was hand transcribed, so sorry for any typos. -Jon db t _bus_dmamap_load_buffer(c3fd3c40,c044660,c4193e84,24,0) at _bus_dmamap_load_buffer+0x3ff bus_dmamap_load(c3fd3c40,0,c4193e84,24,c01c9760) at bus_dmamap_load+0x6c trm_action(c406c200,c418dc00,c0130cc4,c4193e00,c4193e00) at trm_action+0x27a xpt_run_dev_sendq(c406c1c0,c4193e18,1,7,4) at xpt_run_dev_sendq+0x192 xpt_release_devq_device(c4193e00,1,1,d68fbcec,c0243b5e) at xpt_release_devq_device+0xe2 xpt_release_devq_timeout(c4193e00,c1521974,10,f8c6020d,c0131dd0) at xpt_release_devq_timeout+0x21 softclock (0,0,0,0,c1521974) at softclock+0x24e ithread_loop(c1520280,d68fbd48,0,0,c1520280) at ithread_loop+0x1c8 fork_exit(c021d770,c1520280,d68fbd48 at fork_exit+0xb1 fork_trampoline() at fork_trampoline+0x8 --- trap 0x1, eip = 0, esp = 0xd68fbd7c, ebp = 0 --- ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to [EMAIL PROTECTED] ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to [EMAIL PROTECTED]
busdma/scsi trm(4) related panic
Hi, I've been getting what appears to be a busdma / scsi related panic for the past couple of days. This is based solely on what little info I get when it drops to the debugger - I haven't been able to get a core dump. It panics during the boot process immediately after it tries to probe my scsi devices (just a cdrom) attached to my Tekram DC-315U adapter. If I remove trm(4) from my kernel, I can boot without a panic, but if I load the trm module, and then issue `camcontrol rescan all` it will panic. Reverting /usr/src/sys/i386/i386/busdma_machdep.c to version 1.52 lets me boot and use my scsi devices without panics. Version 1.53 will panic too. This is with -current sources as of 5:57am MST (12:47UTC) on Aug. 5. FreeBSD jonnyv.kwsn.lan 5.1-CURRENT FreeBSD 5.1-CURRENT #3: Tue Aug 5 21:09:50 MST 2003 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/JONNYV i386 Here's what little info I have (apologies for typos, I had to transcribe this by hand): Fatal trap 12: page fault while in kernel mode fault virtual address = 0x0 fault code = supervisor write, page not present instruction pointer = 0x8:0xc35682f stack pointer = 0x10:0xd68fbb94 frame pointer = 0x10:0xd68fbbd8 code segment= base 0x0, limit 0xf, type 0x1b = DPL 0, pres 1, def32 1, gran 1 processor eflags= interrupt enabled, resume, IOPL = 0 current process = 12 (swi7: tty:sio clock) kernel: type 12 trap, code = 0 Stopped at _bus_dmamap_load_buffer+0x3ff: movl %ecx,0(%edx,%eax,8) $ nm -n /boot/kernel/kernel | grep c03568 c03568e0 T bus_dmamap_load If there's anything you'd like me to type at the db prompt, just ask and I'll send the output. Thanks, Jon ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: busdma/scsi trm(4) related panic
On Wed, 2003-08-06 at 06:42, Scott Long wrote: I know what the problem is and I'm working on a patch right now. Scott Excellent. I'll be happy to test it if needed. Jon ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: busdma/scsi trm(4) related panic
On Wed, 2003-08-06 at 05:00, Craig Rodrigues wrote: On Tue, Aug 05, 2003 at 11:08:33PM -0700, Jon Kuster wrote: If there's anything you'd like me to type at the db prompt, just ask and I'll send the output. Send the backtrace, with the 't' command. Here you go. Once again, this was hand transcribed, so sorry for any typos. -Jon db t _bus_dmamap_load_buffer(c3fd3c40,c044660,c4193e84,24,0) at _bus_dmamap_load_buffer+0x3ff bus_dmamap_load(c3fd3c40,0,c4193e84,24,c01c9760) at bus_dmamap_load+0x6c trm_action(c406c200,c418dc00,c0130cc4,c4193e00,c4193e00) at trm_action+0x27a xpt_run_dev_sendq(c406c1c0,c4193e18,1,7,4) at xpt_run_dev_sendq+0x192 xpt_release_devq_device(c4193e00,1,1,d68fbcec,c0243b5e) at xpt_release_devq_device+0xe2 xpt_release_devq_timeout(c4193e00,c1521974,10,f8c6020d,c0131dd0) at xpt_release_devq_timeout+0x21 softclock (0,0,0,0,c1521974) at softclock+0x24e ithread_loop(c1520280,d68fbd48,0,0,c1520280) at ithread_loop+0x1c8 fork_exit(c021d770,c1520280,d68fbd48 at fork_exit+0xb1 fork_trampoline() at fork_trampoline+0x8 --- trap 0x1, eip = 0, esp = 0xd68fbd7c, ebp = 0 --- ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: busdma/scsi trm(4) related panic
On Fri, 2003-08-08 at 23:46, Scott Long wrote: Attached is an untested patch. Please let me know if it solves the problem Scott That did the trick. Thanks! Jon ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: busdma/scsi trm(4) related panic
Jon Kuster wrote:
On Wed, 2003-08-06 at 06:42, Scott Long wrote:
I know what the problem is and I'm working on a patch right now.
Scott
Excellent. I'll be happy to test it if needed.
Jon
Attached is an untested patch. Please let me know if it solves the
problem
Scott
Index: trm.c
===
RCS file: /home/ncvs/src/sys/dev/trm/trm.c,v
retrieving revision 1.12
diff -u -r1.12 trm.c
--- trm.c 1 Jul 2003 15:52:02 - 1.12
+++ trm.c 9 Aug 2003 06:45:05 -
@@ -2999,6 +2999,7 @@
{
u_int16_ti;
PSRBpSRB;
+ int error;
for (i = 0; i TRM_MAX_SRB_CNT; i++) {
pSRB = (PSRB)pACB-pFreeSRB[i];
@@ -3040,6 +3041,17 @@
pSRB-pNextSRB = NULL;
}
pSRB-TagNumber = i;
+
+ /*
+* Create the dmamap. This is no longer optional!
+*
+* XXX This is not freed on unload! None of the other
+* allocations in this function are either!
+*/
+ if ((error = bus_dmamap_create(pACB-buffer_dmat, 0,
+ pSRB-dmamap)) != 0)
+ return (error);
+
}
return (0);
}
___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to [EMAIL PROTECTED]
