public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* [bug][temp.patch] ide recalibrating bug
@ 2002-06-24  6:46 Dzmitry Chekmarou
  0 siblings, 0 replies; only message in thread
From: Dzmitry Chekmarou @ 2002-06-24  6:46 UTC (permalink / raw)
  To: martin; +Cc: linux-kernel

hi

the problem:
  kernel fails when ide recalibrating (dma turned on)
  when schedule called and in_interrupt() returns non-zero - BUG() executed
  do_recalibrate() is executed with in_interrupt() == 1

this happens after 2.5.20 to 2.5.21 patch

procedures call stack:
  drivers/ide/ide.c:
    do_recalibrate()
  drivers/ide/ide-taskfile.c:
    ide_raw_taskfile()
    ide_do_drive_cmd()
  kernel/sched.c:
    wait_for_completion()
    schedule()

solutions:
  do not use wait_for_completion when in_interrupt() non-zero (use old wait-sche
  return to old version

my hardware(this is not hw problem, i think):
  cpu: athlon 1000
  mb: sis735
  ram: 256ddr
  hdd: ibm
                    
compiler(and not compiler problem):
  gcc 3.1

tested kernels:
  2.5.17-20(good)
  2.5.21-24(broken)

sorry for my bad english

wbr zmiter

===here is patch(applies to 2.5.24) to return old code for do_recalibrate (temporary fix)===

diff -Nru a/drivers/ide/ide.c b/drivers/ide/ide.c
--- a/drivers/ide/ide.c Sun Jun 23 17:08:19 2002
+++ b/drivers/ide/ide.c Sun Jun 23 23:11:51 2002
@@ -550,6 +550,39 @@
  * We are still on the old request path here so issuing the recalibrate command
  * directly should just work.
  */
+
+/*
+ * Here is deleted ide_set_handler and recal_intr for do_recalibrate
+ */
+
+void ide_set_handler(struct ata_device *drive, ata_handler_t handler,
+                     unsigned long timeout, ata_expiry_t expiry)
+{
+       unsigned long flags;
+       struct ata_channel *ch = drive->channel;
+
+       spin_lock_irqsave(ch->lock, flags);
+
+       if (ch->handler != NULL) {
+               printk("%s: ide_set_handler: handler not null; old=%p, new=%p, f
+                       drive->name, ch->handler, handler, __builtin_return_addr
+       }
+       ch->handler = handler;
+       ch->expiry = expiry;
+       ch->timer.expires = jiffies + timeout;
+       add_timer(&ch->timer);
+
+       spin_unlock_irqrestore(ch->lock, flags);
+}
+
+ide_startstop_t recal_intr(struct ata_device *drive, struct request *rq)
+{
+       if (!ata_status(drive, READY_STAT, BAD_STAT))
+               return ata_error(drive, rq, __FUNCTION__);
+
+       return ide_stopped;
+}
+
 static int do_recalibrate(struct ata_device *drive)
 {

@@ -560,10 +593,28 @@
                struct ata_taskfile args;

                printk(KERN_INFO "%s: recalibrating...\n", drive->name);
+
                memset(&args, 0, sizeof(args));
                args.taskfile.sector_count = drive->sect;
                args.cmd = WIN_RESTORE;
-               ide_raw_taskfile(drive, &args);
+               args.command_type = IDE_DRIVE_TASK_NO_DATA;
+
+               ata_irq_enable(drive, 1);
+               ata_mask(drive);
+
+               if ((drive->id->command_set_2 & 0x0400) && (drive->id->cfs_enable_2 & 0x0400) && (drive->addressing == 1))
+                       ata_out_regfile(drive, &args.hobfile);
+
+               ata_out_regfile(drive, &args.taskfile);
+
+               {
+                       u8 HIHI = (drive->addressing) ? 0xE0 : 0xEF;
+                       OUT_BYTE((args.taskfile.device_head & HIHI) | drive->sel
ect.all, IDE_SELECT_REG);
+               }
+
+               ide_set_handler(drive, recal_intr, WAIT_CMD, NULL);
+               OUT_BYTE(args1->cmd, IDE_COMMAND_REG);
+
                printk(KERN_INFO "%s: done!\n", drive->name);
        }


^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2002-06-24  6:45 UTC | newest]

Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2002-06-24  6:46 [bug][temp.patch] ide recalibrating bug Dzmitry Chekmarou

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox