Talk:Porting Theora to 353 cameras - Revision history

JeremViewsurf at 14:07, 22 May 2008

2008-05-22T14:07:30Z

← Older revision		Revision as of 14:07, 22 May 2008
Line 240:		Line 240:
	Working on the driver made me realize that the compressor never reached the end of a frame... The compressor just stopped processing data after a while and a bunch of signal switched to an undetermined state. I was working on this issue and realized that I got a problem with clock phases and data was written to wrong locations in the sdram. Now it looks like it's fixed, I don't have undetermined signals anymore and the done_compress signal goes high after a frame is processed. Fixed files have been committed.		Working on the driver made me realize that the compressor never reached the end of a frame... The compressor just stopped processing data after a while and a bunch of signal switched to an undetermined state. I was working on this issue and realized that I got a problem with clock phases and data was written to wrong locations in the sdram. Now it looks like it's fixed, I don't have undetermined signals anymore and the done_compress signal goes high after a frame is processed. Fixed files have been committed.

	I am now dealing with interrupts and I can't make the IRQ signal go high when done_compress is high like the 333 design. In the 333 design there's a simple interrupt manager that let the irq go through according to a certain mask value. However I don't understand how the 353 interrupt vector block works... At that time I just plugged the interrupt sources (done_compress, reset_compress etc) in the irq_in input but obviously I think there are ~~mor~~ to do to make that work. --[[User:JeremViewsurf\|JeremViewsurf]] 03:32, 22 May 2008 (CDT)		I am now dealing with interrupts and I can't make the IRQ signal go high when done_compress is high like the 333 design. In the 333 design there's a simple interrupt manager that let the irq go through according to a certain mask value. However I don't understand how the 353 interrupt vector block works... At that time I just plugged the interrupt sources (done_compress, reset_compress etc) in the irq_in input but obviously I think there are more to do to make that work. --[[User:JeremViewsurf\|JeremViewsurf]] 03:32, 22 May 2008 (CDT)

			By the way, I tried to use the old 333 interrupts manager block with the new design and it seems to work with the simulation (IRQ goes high for some time when the frame is done processing). However I'm stuck with how to modify the device driver with this design, (R_IRQ_MASK0_SET used but undeclared). --[[User:JeremViewsurf\|JeremViewsurf]] 09:07, 22 May 2008 (CDT)

JeremViewsurf at 08:32, 22 May 2008

2008-05-22T08:32:03Z

← Older revision		Revision as of 08:32, 22 May 2008
Line 231:		Line 231:

	Andrey is right, that was just a stupid space problem... By the way, I'll get the latest version of the source code as I'm working on the driver (well just trying to understand how is it structured right now...). Thanks for the quick answer. --[[User:JeremViewsurf\|JeremViewsurf]] 02:17, 22 April 2008 (CDT)		Andrey is right, that was just a stupid space problem... By the way, I'll get the latest version of the source code as I'm working on the driver (well just trying to understand how is it structured right now...). Thanks for the quick answer. --[[User:JeremViewsurf\|JeremViewsurf]] 02:17, 22 April 2008 (CDT)



			----



			Working on the driver made me realize that the compressor never reached the end of a frame... The compressor just stopped processing data after a while and a bunch of signal switched to an undetermined state. I was working on this issue and realized that I got a problem with clock phases and data was written to wrong locations in the sdram. Now it looks like it's fixed, I don't have undetermined signals anymore and the done_compress signal goes high after a frame is processed. Fixed files have been committed.

			I am now dealing with interrupts and I can't make the IRQ signal go high when done_compress is high like the 333 design. In the 333 design there's a simple interrupt manager that let the irq go through according to a certain mask value. However I don't understand how the 353 interrupt vector block works... At that time I just plugged the interrupt sources (done_compress, reset_compress etc) in the irq_in input but obviously I think there are mor to do to make that work. --[[User:JeremViewsurf\|JeremViewsurf]] 03:32, 22 May 2008 (CDT)

JeremViewsurf at 07:17, 22 April 2008

2008-04-22T07:17:50Z

← Older revision		Revision as of 07:17, 22 April 2008
Line 229:		Line 229:
	1053: #define X313_IRQSTATE (port_csp0_addr[0x11],port_csp0_addr[0x11])		1053: #define X313_IRQSTATE (port_csp0_addr[0x11],port_csp0_addr[0x11])
	there is a space between "X313_IRQSTATE" and "(" separating macro name with 0 parameters and the macro body. Where could that space character disappear in your example?--[[User:Andrey.filippov\|Andrey.filippov]] 14:22, 21 April 2008 (CDT)		there is a space between "X313_IRQSTATE" and "(" separating macro name with 0 parameters and the macro body. Where could that space character disappear in your example?--[[User:Andrey.filippov\|Andrey.filippov]] 14:22, 21 April 2008 (CDT)

			Andrey is right, that was just a stupid space problem... By the way, I'll get the latest version of the source code as I'm working on the driver (well just trying to understand how is it structured right now...). Thanks for the quick answer. --[[User:JeremViewsurf\|JeremViewsurf]] 02:17, 22 April 2008 (CDT)

Andrey.filippov at 19:22, 21 April 2008

2008-04-21T19:22:52Z

← Older revision		Revision as of 19:22, 21 April 2008
Line 226:		Line 226:

	Strange. Could you pls use LANG=CC ? So others will be able to read error messages too :) What GNU/Linux distribution do you use ? what version of gcc-cris? Also it would be better to sync to last CVS, many bugs was resolved since .11. --[[User:Polto\|Alexandre.Poltorak]] 12:45, 21 April 2008 (CDT)		Strange. Could you pls use LANG=CC ? So others will be able to read error messages too :) What GNU/Linux distribution do you use ? what version of gcc-cris? Also it would be better to sync to last CVS, many bugs was resolved since .11. --[[User:Polto\|Alexandre.Poltorak]] 12:45, 21 April 2008 (CDT)
			In the http://elphel.cvs.sourceforge.net/elphel/elphel353-7.1/os/linux-2.6-tag--devboard-R2_10-4/arch/cris/arch-v32/drivers/elphel/x353.h?view=markup
			1053: #define X313_IRQSTATE (port_csp0_addr[0x11],port_csp0_addr[0x11])
			there is a space between "X313_IRQSTATE" and "(" separating macro name with 0 parameters and the macro body. Where could that space character disappear in your example?--[[User:Andrey.filippov\|Andrey.filippov]] 14:22, 21 April 2008 (CDT)

Polto at 17:45, 21 April 2008

2008-04-21T17:45:04Z

← Older revision		Revision as of 17:45, 21 April 2008
Line 224:		Line 224:

	At the first occurrence of #include "x3x3.h" some 353 macros such as #define X313_IRQSTATE(port_csp0_addr[0x11],port_csp0_addr[0x11]) with the "[" character stops the compilation. --[[User:JeremViewsurf\|JeremViewsurf]] 10:48, 21 April 2008 (CDT)		At the first occurrence of #include "x3x3.h" some 353 macros such as #define X313_IRQSTATE(port_csp0_addr[0x11],port_csp0_addr[0x11]) with the "[" character stops the compilation. --[[User:JeremViewsurf\|JeremViewsurf]] 10:48, 21 April 2008 (CDT)

			Strange. Could you pls use LANG=CC ? So others will be able to read error messages too :) What GNU/Linux distribution do you use ? what version of gcc-cris? Also it would be better to sync to last CVS, many bugs was resolved since .11. --[[User:Polto\|Alexandre.Poltorak]] 12:45, 21 April 2008 (CDT)

JeremViewsurf at 15:48, 21 April 2008

2008-04-21T15:48:43Z

← Older revision		Revision as of 15:48, 21 April 2008
Line 208:		Line 208:

	I carried out the synthesis of the design with ISE, I am able to get the .bit file with some warnings (old constraints file from 353 : some timings not met, some clocks warnings and obviously unused signals) but no error. I started to have a look at the driver and made a first version of the definition file x3x3.h (see CVS). Some macros at the end may be obsolete as they command the 353 memory controller. --[[User:JeremViewsurf\|JeremViewsurf]] 08:32, 17 April 2008 (CDT)		I carried out the synthesis of the design with ISE, I am able to get the .bit file with some warnings (old constraints file from 353 : some timings not met, some clocks warnings and obviously unused signals) but no error. I started to have a look at the driver and made a first version of the definition file x3x3.h (see CVS). Some macros at the end may be obsolete as they command the 353 memory controller. --[[User:JeremViewsurf\|JeremViewsurf]] 08:32, 17 April 2008 (CDT)

			Problem while compiling with definition file x3x3.h :
			Making install in packages/os/linux-2.6 for crisv32-axis-linux-gnu
			make[2]: entrant dans le répertoire « /home/larrey/Elphel/353t_clean/elphel353-7.1.7.11/elphel353t/packages/os/linux-2.6-R1_4_2 »
			subdirs=
			MAKE /home/larrey/Elphel/353t_clean/elphel353-7.1.7.11/elphel353t/os/linux-2.6 zImage
			make[3]: entrant dans le répertoire « /home/larrey/Elphel/353t_clean/elphel353-7.1.7.11/elphel353t/os/linux-2.6-tag--devboard-R2_10-4 »
			CHK include/linux/version.h
			CHK include/linux/utsrelease.h
			CHK include/linux/compile.h
			CC arch/cris/arch-v32/drivers/elphel/fpgajtag.o
			In file included from arch/cris/arch-v32/drivers/elphel/fpgajtag.c:95:
			arch/cris/arch-v32/drivers/elphel/x3x3.h:963:45: "[" may not appear in macro parameter list
			make[5]: *** [arch/cris/arch-v32/drivers/elphel/fpgajtag.o] Erreur 1

			At the first occurrence of #include "x3x3.h" some 353 macros such as #define X313_IRQSTATE(port_csp0_addr[0x11],port_csp0_addr[0x11]) with the "[" character stops the compilation. --[[User:JeremViewsurf\|JeremViewsurf]] 10:48, 21 April 2008 (CDT)

JeremViewsurf at 14:08, 17 April 2008

2008-04-17T14:08:51Z

← Older revision		Revision as of 14:08, 17 April 2008
Line 1:		Line 1:
	~~--[[User:JeremViewsurf\|JeremViewsurf]] 08:32, 17 April 2008 (CDT)~~Ok, I started a new page on the Theora project in 353 cameras. I started with a global approach and a list of modules that need to be changed or replaced on the 353 model. Feel free to correct me or add suggestions as I'm not (yet...) an expert on Elphel cameras. I think the first step is to make an exhaustive list of the hardware blocks that need modifications and then detail what need to be done in each block. --[[User:JeremViewsurf\|JeremViewsurf]] 03:48, 19 February 2008 (CST)		Ok, I started a new page on the Theora project in 353 cameras. I started with a global approach and a list of modules that need to be changed or replaced on the 353 model. Feel free to correct me or add suggestions as I'm not (yet...) an expert on Elphel cameras. I think the first step is to make an exhaustive list of the hardware blocks that need modifications and then detail what need to be done in each block. --[[User:JeremViewsurf\|JeremViewsurf]] 03:48, 19 February 2008 (CST)

JeremViewsurf at 13:32, 17 April 2008

2008-04-17T13:32:14Z

← Older revision		Revision as of 13:32, 17 April 2008
Line 1:		Line 1:
	Ok, I started a new page on the Theora project in 353 cameras. I started with a global approach and a list of modules that need to be changed or replaced on the 353 model. Feel free to correct me or add suggestions as I'm not (yet...) an expert on Elphel cameras. I think the first step is to make an exhaustive list of the hardware blocks that need modifications and then detail what need to be done in each block. --[[User:JeremViewsurf\|JeremViewsurf]] 03:48, 19 February 2008 (CST)		--[[User:JeremViewsurf\|JeremViewsurf]] 08:32, 17 April 2008 (CDT)Ok, I started a new page on the Theora project in 353 cameras. I started with a global approach and a list of modules that need to be changed or replaced on the 353 model. Feel free to correct me or add suggestions as I'm not (yet...) an expert on Elphel cameras. I think the first step is to make an exhaustive list of the hardware blocks that need modifications and then detail what need to be done in each block. --[[User:JeremViewsurf\|JeremViewsurf]] 03:48, 19 February 2008 (CST)


Line 200:		Line 200:

	I realized that I did not check the frame dimensions and I tried to simulate 1280x1024 frames (init_dimensions(9,7)). I changed to init_dimensions(1,1) thus setting nstx and nsty to 1. I now have output on ch7 after 2 ms, thanks. --[[User:JeremViewsurf\|JeremViewsurf]] 04:07, 14 April 2008 (CDT)		I realized that I did not check the frame dimensions and I tried to simulate 1280x1024 frames (init_dimensions(9,7)). I changed to init_dimensions(1,1) thus setting nstx and nsty to 1. I now have output on ch7 after 2 ms, thanks. --[[User:JeremViewsurf\|JeremViewsurf]] 04:07, 14 April 2008 (CDT)



			----



			I carried out the synthesis of the design with ISE, I am able to get the .bit file with some warnings (old constraints file from 353 : some timings not met, some clocks warnings and obviously unused signals) but no error. I started to have a look at the driver and made a first version of the definition file x3x3.h (see CVS). Some macros at the end may be obsolete as they command the 353 memory controller. --[[User:JeremViewsurf\|JeremViewsurf]] 08:32, 17 April 2008 (CDT)

JeremViewsurf at 09:07, 14 April 2008

2008-04-14T09:07:48Z

← Older revision		Revision as of 09:07, 14 April 2008
Line 198:		Line 198:

	Jeremy, what frame size do you try to simulate ? You should try on a minimal: 32x128. Look on the 333 Theora testbench for more info. --[[User:Polto\|Alexandre.Poltorak]] 14:28, 11 April 2008 (CDT)		Jeremy, what frame size do you try to simulate ? You should try on a minimal: 32x128. Look on the 333 Theora testbench for more info. --[[User:Polto\|Alexandre.Poltorak]] 14:28, 11 April 2008 (CDT)

			I realized that I did not check the frame dimensions and I tried to simulate 1280x1024 frames (init_dimensions(9,7)). I changed to init_dimensions(1,1) thus setting nstx and nsty to 1. I now have output on ch7 after 2 ms, thanks. --[[User:JeremViewsurf\|JeremViewsurf]] 04:07, 14 April 2008 (CDT)

Polto at 19:28, 11 April 2008

2008-04-11T19:28:47Z

← Older revision		Revision as of 19:28, 11 April 2008
Line 196:		Line 196:

	--[[User:JeremViewsurf\|JeremViewsurf]] 07:53, 11 April 2008 (CDT)		--[[User:JeremViewsurf\|JeremViewsurf]] 07:53, 11 April 2008 (CDT)

			Jeremy, what frame size do you try to simulate ? You should try on a minimal: 32x128. Look on the 333 Theora testbench for more info. --[[User:Polto\|Alexandre.Poltorak]] 14:28, 11 April 2008 (CDT)