changelog

一个2.4.21版本的嵌入式linux内核

Affected files ...

... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_osm.h#96 edit
... //depot/linux_mod_devel/scsi.aic79xx/rpm/aic79xx.spec#20 edit
... //depot/linux_mod_devel/scsi.aic79xx/rpm/install.sh#23 edit

Change 1786 by gibbs@overdrive on 2002/12/19 16:10:05

	Regenerate Linux firmware.

Affected files ...

... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_reg.h#53 edit
... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_reg_print.c#34 edit
... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_seq.h#54 edit

Change 1785 by gibbs@aslan on 2002/12/19 16:08:47

	aic79xx.c:
		Add a statistics timer that decides when to enable
		or disable interrupt coalessing based on load.
	
		Add ahd_flush_qoutfifo() which will run the qoutfifo
		as well as complete any commands sitting on the 
		sequencer's COMPLETE_SCB lists.  Use this routine in
		several places that did similar things in an add-hoc,
		but incomplete, fashion.  A call to this routine was
		also added to ahd_abort_scbs() to close a race.
	
		Add a routine, ahd_reset_cmds_pending() which is used
		to update the CMDS_PENDING sequencer variable whenever
		error recovery compeltes SCBs without notifying the
		sequencer.  Since ahd_reset_cmds_pending is called
		during ahd_unpause() only if we've aborted SCBs, its
		call to ahd_flush_qoutfifo should not cause recursion
		through ahd_run_qoutfifo().  A panic has been added to
		ensure that this recursion does not occur.
	
		In ahd_search_qinfifo, update the CMDS_PENDING sequencer
		variable directly.  ahd_search_qinififo can be called
		in situations where using ahd_reset_cmds_pending() might
		cause recursion.  Since we can safely determine the
		exact number to reduce CMDS_PENDING by in this scenario
		without running the qoutfifo, the manual update is more
		than adequate.
	
		Clean up diagnostics.
	
	aic79xx.h:
		Update per-softc variables for the stats "daemon".
		The defaults for interrupt coalessing have been updated
		slightly, but these need additional tuning.  A new
		tunable paramter "minimum commands still outstanding
		required to attempt coalessing" has been added.
	
		Add a debug option for interrupt coalessing activities.
	
		Add two new softc flags:
		o AHD_UPDATE_PEND_CMDS
			Run ahd_reset_cmds_pending() on the next unpause.
	
		o AHD_RUNNING_QOUTFIFO
			Used to catch recursion through ahd_run_qoutfifo().
	
	aic79xx.reg:
		Break INTMASK in SEQITNCTL out into INTMASK1 and INTMASK2.
		In at least the REV A, these are writable bits.  We make
		use of that for a swtimer workaround in the sequencer.
	
		Add sequencer variables to recovery the min-commands
		still outstanding for coalessing and the current number
		of commands outstanding.
	
		Since HS_MAILBOX autoclears, provide a sequencer variable
		to store its contents.
	
	aic79xx.seq:
		In idle_loop_cchan, update LOCAL_HS_MAILBOX everytime
		we are notified of an HS_MAILBOX update via the
		HS_MAILBOX_ACT bit in QOFF_CTLSTA.
	
		Enhance our coalessing algorithm.  If we have more
		SCBs to complete to the host (sitting in COMPLETE_SCB
		lists), always try to coaless them up to our coalessing
		limit.  If coalessing is enabled, but we have fewer
		commands oustantind than the hosts limit, complete the
		command immediately.
	
		Since we cannot disable the swtimer's countdown, simply
		mask its interrupt once we no longer care about it firing.
	
		Add code to track the number of commands outstanding.
		Commands are outstanding from the time they are placed
		into the execution queue until the DMA to post completion
		is setup.
	
		Add a workaround for intvec_2 interrupts on the H2A4.
		In H2A4, the mode pointer is not saved for intvec2, but
		is restored on iret.  This can lead to the restoration
		of a bogus mode ptr.  Manually clear the intmask bits and
		do a normal return to compensate.
	
	aic79xx_inline.h:
		Call ahd_reset_cmds_pending() in ahd_unpause if required.
	
		Update cmdcmplt interrupt statistics in our interrupt
		handler.

Affected files ...

... //depot/aic7xxx/aic7xxx/aic79xx.c#146 edit
... //depot/aic7xxx/aic7xxx/aic79xx.h#78 edit
... //depot/aic7xxx/aic7xxx/aic79xx.reg#59 edit
... //depot/aic7xxx/aic7xxx/aic79xx.seq#77 edit
... //depot/aic7xxx/aic7xxx/aic79xx_inline.h#40 edit

Change 1784 by gibbs@bitkeeper-linux-2.5 on 2002/12/18 10:33:25

	aic79xx_osm.c:
		Remove a stray ';' that was short cicuiting an if statement.
	
	aic79xx_osm.c:
	aic7xxx_osm.c:
	aic79xx_osm.h:
	aic7xxx_osm.h:
		Only allow slave_destroy() to destroy a real device
		if that device has had slave_configure() called on it.
		This prevents "silly deletions" caused by bogons in
		the scsi_scan code from deleting DV state for devices
		that are still attached and will have a slave_alloc()
		called for them as soon as the slave_destroy() returns.

Affected files ...

... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_osm.c#100 edit
... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_osm.h#95 edit
... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic7xxx_osm.c#163 edit
... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic7xxx_osm.h#110 edit

Change 1783 by gibbs@bitkeeper-linux-2.5 on 2002/12/18 10:29:46

	aic7xxx.c:
		Don't clobber ppr_options when forcing a renegotiation.
		The current ppr_options may be referenced while queuing
		new commands.

Affected files ...

... //depot/aic7xxx/aic7xxx/aic7xxx.c#104 edit

Change 1782 by gibbs@bitkeeper-linux-2.5 on 2002/12/17 19:27:05

	aic79xx_osm.c:
		Add a prototype for ahd_linux_dv_fallback().
	
	aic7xxx_osm.c:
		Complete the merge of the new biosparam API from
		Christoph Hellwig.
	
	aic79xx_osm.c:
	aic7xxx_osm.c:
		Move the setting of max_sectors to the host template
		to quiet a warning in 2.5.X.
	
		Add slave_alloc calls and "bootverbose" diagnostics
		for help in sorting out exactly how this slave_*
		API functions.
	
		Move call to ahd_linux_device_queue_depth to within
		the test to see if dev is NULL.
	
		Add code to actually destroy our data structures to
		the slave_destroy routine instead of waiting for a
		command that may not come to do so.
	
		Initialize the name field of the host structure.

Affected files ...

... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_osm.c#99 edit
... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic7xxx_osm.c#162 edit

Change 1781 by gibbs@bitkeeper-linux-2.5 on 2002/12/17 19:20:16

	aic79xx.c:
		Remove "Now packetized" diagnostic now that this
		information is incorperated into the actual negotiation
		messages that are displayed.
	
		When forcing renegotiation, don't clober the current
		ppr_options.  Much of the driver uses this information
		to determine if we are currently packetized or not.
	
		Remove some stray spaces at column 1 and ahd_set_tags.
	
		When complaining about getting a host message loop
		request with no pending messages, print out the 
		SCB_CONTROL register down on the card.
	
		Modify the ahd_sent_msg() routine to handle a search
		for an outgoing identify message.  Use this to detect
		a msg reject on an identify message which typically
		indicates that the target thought we were packetized.
		Force a renegotiation in this case.
	
		In ahd_search_qinfifo(), wait more effectively for SCB
		DMA activities to cease.  We also clear out the state
		that tells the sequencer that a DMA was in progress for
		SCB fetch operations since we are about to change the
		qinfifo.
	
		In ahd_pause_and_flush_work(), actually itterate through
		the two Complete SCB lists.  Ooops.
	
		In ahd_qinfifo_count(), fix the qinfifo empty case.
	
		In ahd_dump_card_state(), print out CCSCBCTL in the
		correct mode.

Affected files ...

... //depot/aic7xxx/aic7xxx/aic79xx.c#145 edit

Change 1780 by gibbs@bitkeeper-linux-2.5 on 2002/12/17 19:11:25

	aic79xx.seq:
		Add a missing ret to the last instruction in
		load_overrun_buf.

Affected files ...

... //depot/aic7xxx/aic7xxx/aic79xx.seq#76 edit

Change 1779 by gibbs@bitkeeper-linux-2.5 on 2002/12/17 19:10:42

	aic79xx.reg:
		Correct address for the DFDBCTL register.

Affected files ...

... //depot/aic7xxx/aic7xxx/aic79xx.reg#58 edit

Change 1774 by scottl@scottl-template on 2002/12/16 03:08:08

	       Update driver version to 1.3.0.ALPHA4

Affected files ...

... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_osm.h#94 edit
... //depot/linux_mod_devel/scsi.aic79xx/rpm/aic79xx.spec#19 edit
... //depot/linux_mod_devel/scsi.aic79xx/rpm/install.sh#22 edit

Change 1772 by scottl@scottl-belfalas-template on 2002/12/16 02:43:36

	ahd_linux_fallback can't aquire the ahd lock since it might be
	called from the completion handler.  Create a wrapper called
	ahd_linux_dv_fallback that aquires and releases the ahd lock so
	that is can be called from the DV code and still satisfy it's locking
	needs.
	Move the error stat code into ahd_linux_queue_cmd_complete so it
	can catch more error conditions.

Affected files ...

... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_osm.c#98 edit

Change 1771 by scottl@scottl-belfalas-template on 2002/12/16 02:39:22

	aic79xx_proc.c:
		Add a newline when printing the renegotiation pending message.

Affected files ...

... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_proc.c#11 edit

Change 1770 by scottl@scottl-belfalas-template on 2002/12/15 23:38:06

	aic79xx_osm.c:
		Don't grab the ahd lock when updating error stats since
		it's already held.
		When deciding to do a fallback because of an error, don't
		alias the existing devinfo struct with a new one.  Why
		the compiler didn't catch this is beyond me.

Affected files ...

... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_osm.c#97 edit

Change 1769 by scottl@scottl-hobbiton-0b47 on 2002/12/15 22:24:16

	Regen firmware

Affected files ...

... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_reg.h#52 edit
... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_reg_print.c#33 edit
... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_seq.h#53 edit

Change 1768 by scottl@scottl-hobbiton-0b47 on 2002/12/15 22:24:02

	aic79xx_osm.c:
	aic79xx_osm.h:
		Switch to a bucket-based method of counting the number off
		good commands in between bad ones.  Use only a single bucket
		right now for simplicity, though this can grow in the future.
		This also removes the stats timer.

Affected files ...

... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_osm.c#96 edit
... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_osm.h#93 edit

Change 1767 by scottl@scottl-hobbiton-0b47 on 2002/12/15 21:23:38

	aic79xx_osm.h:
		Move the errors_detected field to the target structure, and
		add new_errors_detected field for doing stats decay.
	
	aic79xx_osm.c:
		If a transmission error is detected, increment the
		new_errors_detected counter.  If it's passed the threshold,
		call for a speed fallback.
		Update the errors_detected field with new_errors_detected and
		reset new_errors_detected on every stats timeout call.
	
	aic79xx_proc.c:
		Display the per-target errors detected field.

Affected files ...

... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_osm.c#95 edit
... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_osm.h#92 edit
... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_proc.c#10 edit

Change 1766 by gibbs@aslan on 2002/12/15 19:32:13

	Add interrupt coalessing hooks.
	
	aic79xx.c:
		Keep a copy of the hs_mailbox in our softc so that
		we can perform read/modify/write operations on the
		hs_mailbox without having to pause the sequencer to
		read the last written value.  Use the ENINT_COALESS
		flag in the hs_mailbox to toggle interrupt coalessing.
	
		Add entrypoints for enabling interrupt coalessing and
		setting both a timeout (how long to wait for commands
		to be coalessed) and a maximum commands to coaless value. 
	
		In ahd_pause_and_flushwork() only return one selections
		are safely disabled.  We now also flush the two sequencer
		lists of completed commands.
	
	aic79xx.h:
		Add coalessing and HS_MAILBOX fields.
	
	aic79xx.reg:
		Correct register addresses related to the software timer.
	
		Add constants paramaterizing the software timer.
	
		Add scratch ram locations for storing interrupt coalessing
		tunables.
	
	aic79xx.seq:
		Use the software timer and a commands completed count to
		implement interrupt coalessing.  The command complete is
		deferred until either the maximum command threshold or a
		timer since the first command completed since the last
		completion expires.
	
		Move the test for the cfg4istat interrupt up an instruction
		to hopefully close a race between the next outgoing selection
		and our disabling of selections.

Affected files ...

... //depot/aic7xxx/aic7xxx/aic79xx.c#144 edit
... //depot/aic7xxx/aic7xxx/aic79xx.h#77 edit
... //depot/aic7xxx/aic7xxx/aic79xx.reg#57 edit
... //depot/aic7xxx/aic7xxx/aic79xx.seq#75 edit

Change 1765 by gibbs@aslan on 2002/12/15 17:48:29

	aicasm_gram.y:
		Allow constants to be complex expressions so long
		as those expressions can be fully evaluated during
		assembly.

Affected files ...

... //depot/aic7xxx/aic7xxx/aicasm/aicasm_gram.y#28 edit

Change 1764 by scottl@scottl-hobbiton-0b47 on 2002/12/13 18:09:34

	aic79xx_osm.c:
	aic79xx_osm.h:
		Start of runtime fallback and interrupt caolescing stat
		gathering.  Run ahd_linux_stats_update() every
		AHD_LINUX_STATS_INTERVAL period.  Consense the packetized
		and non-packetized autosense code and start some logic to
		figure out what to do with the sense information.

Affected files ...

... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_osm.c#94 edit
... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_osm.h#91 edit

Change 1762 by gibbs@bitkeeper-linux-2.5 on 2002/12/13 13:08:43

	aic79xx_osm.c:
	aic7xxx_osm.c:
		Remove *slave_alloc declarations as we don't need or
		export this entry point.

Affected files ...

... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic79xx_osm.c#93 edit
... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aic7xxx_osm.c#161 edit

Change 1761 by gibbs@bitkeeper-linux-2.4 on 2002/12/12 14:31:11

	aiclib.h:
		Add aic_sector_div macro.

Affected files ...

... //depot/aic7xxx/linux/drivers/scsi/aic7xxx/aiclib.h#4 edit