Tetsu special kernel for the 6000

From OESF

Revision as of 18:25, 24 November 2006 by Phil284 (Talk | contribs)
(diff) ←Older revision | Current revision (diff) | Newer revision→ (diff)
Jump to: navigation, search

source: http://tetsu.homelinux.org/zaurus/kernel/v16a/

Contents

Recent condition report

  • They are 2005/7/8 SL-C3100private v18b-pre3. It built v18a suitable ones of SL-C1000/C3000 from the source of C1000. Because (note) the source code for SL-3100 is not open from SHARP, to reset to the origin, it is necessary to restore NAND backup. Please do not use the person who cannot understand this meaning. 
* SL-C1000 kernel source (2004/4/14) 
* (P01) Clock variable patch of Yamada's (2005/2/22) 
* (P02) Core voltage variable patch of piro (2005/2/22) 
* (P03) Patch for SIGSTOP problem of piro (2004/12/7) 
* (P04) The UNICON patch of naorou (2004/12/7) 
* (P05) Writing brush pressure concern patch ahead and others (2004/12/7) 
* (P06) Touch screen entry patch of piro (2004/12/7) 
* (P07) Keyboard device entry patch of piro (2004/12/7) 
* (P08) And others the security patch of rtc and mremap of む じ ぃ (2004/12/14) 
* (P09) Poly & and others the security patch of ext3, isofs of む じ ぃ (2004/12/17) 
* (P10) The NTFS corresponding conversion patch of poly (2004/12/17) 
* (P11) The Bluetooth up-to-date patch of Marcel Holtmann (2004/12/17) 
* (P12) Bug fix of fbcon of piro (the freezing evasion with console) patch (2004/12/19) 
* (P14) The low latency patch of Andrew Morton (2004/12/23) 
* (P17) Bvdd modularization corresponding patch of piro (2005/2/22) 
* (P18) The battery remaining amount detailed indicatory patch of tattsuu (2005/1/23) 
* The starting message indicatory patch of 59414d41
* Economical electrical patch of Yamada &poly (2005/2/26) 
* The CPAR context switch patch of poly (2005/2/18) 
* The back port of NWFPE from kernel-2.4.30 (2005/5/3) 
* (New) The patch which makes Yamada's SL-C1000 source compulsorily C3100 correspondence (2005/7/8)
  • It is v16a for 2005/7/7 SL-6000. Floating point arithmetic approximately 10% accelerated by the fact that the cord/code whose NWFPE is new is applied. The patch below hitting, it increases. 
* SL-6000 kernel source (2004/4/21) 
* The patch which is made CCCR=163 
* Clock rise patch 1
* Clock rise patch 2
* Clock rise patch 3
* The Bluetooth up-to-date patch of Marcel Holtmann (adjusting to the part SHARP source, correction) 
* The starting message indicatory patch of 59414d41
* Page-out quantitative adjustment patch of coming plate
* And others security patch of む じ ぃ (CAN-2004-0003/CAN-2004-0010/CAN-2004-0109/CAN-2004-0177/CAN-2004-0178 correspondence) 
* (New) The back port of NWFPE from kernel-2.4.30 (2005/7/7)

It is v14e for 2005/7/1SL-B500/C700/C750/C760/C860. Floating point arithmetic approximately 10% accelerated by the fact that the cord/code whose NWFPE is new is applied. In addition, the battery remaining amount detailed indicatory patch of tattsuu (B500 you exclude) it has applied. The patch below hitting, it increases. SL-C860 kernel source (2003/11/7) There is no integrated patch preempt of the ぴ ろ (2004/11/9) The Bluetooth up-to-date patch of Marcel Holtmann (adjusting to the part SHARP source, correction) The starting message indicatory patch of 59414d41 Page-out quantitative adjustment patch of coming plate And others security patch of む じ ぃ (CAN-2004-0003/CAN-2004-0010/CAN-2004-0109/CAN-2004-0177/CAN-2004-0178 correspondence) (New) (P18) The battery remaining amount detailed indicatory patch of tattsuu (2005/1/23) (New) The back port of NWFPE from kernel-2.4.30 (2005/5/3) They are 2005/5/3 SL-C1000/C3000privatev18a. Floating point arithmetic approximately 10% accelerated by the fact that the cord/code whose NWFPE is new is applied. The patch below hitting, it increases. SL-C1000 kernel source (2004/4/14) (P01) Clock variable patch of Yamada's (2005/2/22) (P02) Core voltage variable patch of ぴ ろ (2005/2/22) (P03) Patch for SIGSTOP problem of ぴ ろ (2004/12/7) (P04) The UNICON patch of naorou (2004/12/7) (P05) Writing brush pressure concern patch ahead and others (2004/12/7) (P06) Touch screen entry patch of ぴ ろ (2004/12/7) (P07) Keyboard device entry patch of ぴ ろ (2004/12/7) (P08) And others the security patch of rtc and mremap of む じ ぃ (2004/12/14) (P09) Poly & and others the security patch of ext3, isofs of む じ ぃ (2004/12/17) (P10) The NTFS corresponding conversion patch of poly (2004/12/17) (P11) The Bluetooth up-to-date patch of Marcel Holtmann (2004/12/17) (P12) Bug fix of fbcon of ぴ ろ (the freezing evasion with console) patch (2004/12/19) (P14) The low latency patch of Andrew Morton (2004/12/23) (P17) Bvdd modularization corresponding patch of ぴ ろ (2005/2/22) (P18) The battery remaining amount detailed indicatory patch of tattsuu (2005/1/23) The starting message indicatory patch of 59414d41 Economical electrical patch of Yamada &poly (2005/2/26) The CPAR context switch patch of poly (2005/2/18) (New) The back port of NWFPE from kernel-2.4.30 (2005/5/3) Past past record

Media

When it is in addition to this please teach. 2005/3/31 Linux Zaurus expert(the saw TEX corporation) it was recorded before page 149. Customizing 2005/3/10 Linux Zaurus, the book which you use(the Soviet shim corporation) it was recorded before page 221. 2005/3/1 Linux Zaurus SL-C3000 Super Guide(everyday communication) it was recorded before page 267. 2004/12/25 Zaurus SL-C3000 thorough application manuals(the media TEX publication) it was recorded before page 207. 2004/6/3the book,"hacking Linux Zaurus "from software bankpublishing was sold on 2004 June 21st. I took charge of the section of kernel hack. Because rent of the special kernel is free, when you can buy this book with the money which floats it is delightful, is. SBC store (you can buy here) Cover image small Cover image large (author name is visible) The author introduction with Walrus and Visit (gorgeousness? Writing position) 2004/3/30 Mobile press 2004 spring the number(the technical criticism corporation) it was recorded before page 58. 2004/3/29 Linux Zaurus Super Guide(everyday the communication corporation) it was recorded before page 219. 2004/2/25 Zaurus SL-C860/760/750/700 perfect guides(the media TEX publication) it was recorded before page 222. 2004/2/5 Linux Zaurus wall(the Soviet shim corporation) it was recorded before page 241. 2003/10/28 UNIX MAGAZINE 2003 November edition(ASCII) it was recorded before page 177. 2003/10/8 UNIX USER 2003 November edition(software bank publishing) it was recorded before page 85. 2003/9/22 PDA magazines VOL.4(entering blurring in) it was recorded before page 119.

Irresponsible

Depending upon the use of this kernel, your Zaurus becoming the brick, as for me being nothing to be able to compensate, the あ it does, not driving.

It probably is what

Just a little fumbling the kernel source, it is something which it tries perhaps just a little it moves fast with SL-B500/C700/C750/C760/C860/6000/C1000/C3000/C3100.

License

Being free, it distributes the special kernel, (because of sense). The licenseisGPL.

Replacement procedure of kernel

(1) consolescroll and updater.sh.xxxx are received from the sight of SHARP, (there is no consolescroll for SL-C1000/C3000). As for updater.sh.xxxx file name updater.sh and modification. Copies both parties in the memory card (SL-C1000/C3000 only updater.sh). For SL-B500/C700 consolescroll and updater.sh For SL-C750/C760 consolescroll and updater.sh For SL-C860 consolescroll and updater.sh For SL-6000 consolescroll and updater.sh For SL-C1000 updater.sh (there is no consolescroll in SL-C1000) For SL-C3000 updater.sh (there is no consolescroll in SL-C3000) For SL-C3100 updater.sh (provision release) (2) it chooses the kernel which from description below is agreeable to the type, zImage modifying, it copies file name in the memory card. You pay attention to without being zImage.bin. Does not indicate the extension of the file with such as Windows when it has made the setting which, bin is not indicated and there is a possibility also the て remaining. SL-B500 special kernel (v14e) SL-C700 special kernel (v14e) SL-C750 special kernel (v14e) SL-C760 special kernel (v14e) SL-C860 special kernel (v14e) SL-6000 special kernel (v16a) SL-C1000 special kernel (v18a) SL-C3000 special kernel (v18a) (Danger) SL-C3100 special kernel (v18b-pre3) (only the one which can be reset to the origin with dangerous) --- NAND backup please use (3) verifying that other than consolescroll, updater.sh and zImage is not in the memory card. Being to be dangerous please turn off especially initrd.bin absolutely. After (4) (1) according to procedure of the sight of SHARP it is, but perfection elimination please does not do. Just 7 to execute with procedure 11, then power is turned on normally. Please insert module (5) according to need (option). Especially if there is no inconvenience, installation is unnecessary.

Md5sum of image

Restoration procedure

Case there is a problem something, using the original kernel, description above (3) please redo. If the type which is the update software of SHARP offer (B500, C700, C860) is, please remove the original kernel, from there.

If the type which is not the update software of SHARP offer (C750, C760, 6000) is, please use those which are equal to the original kernel.

/proc file system

With the /proc file system, reference of kernel setting/modification is possible. For example, when liking to refer to the value of CCCR, 

$ cat /proc/cccr_change

When liking to modify CCCR in 241, 

# echo 241 > /proc/cccr_change

And so on, like the usual file it accesses. Modification (entry) root authority is necessary. Principal ones are shown below, but, the kernel patch of piro [1] is collected several is writtenon the page in addition to this. /proc/zaurus/CCCR: The CPU clock of C1000/C3000/C3100 (CCCR value at time of non idling) (read/write) Corresponding type: C1000/C3000/C3100 2000210: CPU clock 416MHz system bus 208MHz LCD104MHz (default) 2000290: CPU clock 520MHz system bus 208MHz LCD104MHz 2000310: CPU clock 624MHz system bus 208MHz LCD104MHz 2000214: CPU clock 520MHz system bus 260MHz LCD 64MHz 2000218: CPU clock 624MHz system bus 312MHz LCD 78MHz A000214: CPU clock 520MHz system bus 260MHz LCD130MHz A000218: CPU clock 624MHz system bus 312MHz LCD156MHz Explanation: At the time of clock rise core voltage please modify with set. Example:

  1. echo 0e > /proc/zaurus/VCORE && echo 2000310 > /proc/zaurus/CCCR

Qclockchange-for-c3000 of the ぴ ろ the recommendation (check is needed forexecution for modification of the clock/verification, with root authority). In addition, if it is BatteryPlus of the ぴ ろ , the clock and core voltage can be changed according to the battery remaining amount. /proc/zaurus/VCORE: The core voltage of C1000/C3000/C3100 (read/write) Corresponding type: C1000/C3000/C3100 00: 0.850V 01: 0.900V 02: 0.950V 03: 1.000V 04: 1.050V 05: 1.100V 06: 1.150V 07: 1.200V 08: 1.250V 09: 1.300V 0a: 1.350V (default: Rating at the time of 416MHz) 0b: 1.400V 0c: 1.450V (rating at the time of 520MHz) 0d: 1.500V 0e: 1.550V (rating at the time of 624MHz) 0f: 1.600V Explanation: At the time of clock rise unless core voltage is increased, it seems that is not stabilized. Example:

  1. echo 0e > /proc/zaurus/VCORE

/proc/cccr_change: CPU clock (CCCR value at time of non idling) (read/write) Corresponding type: B500/C700/C750/C760/C860/6000 241: Value with the normal kernel of B500/C700 242: The initial value for B500/C700 161: The initial value for C750/C760/C860/6000 (the same as normal kernel) 162: If luck is good with C750/C760/C860/6000, whether it moves Explanation: In case of C750/C760/C860/6000, trying trying 162, if useless please use with 161. Here when idling the clock falls due to specification, automatically. Qclockchange of the ぴ ろ the recommendation (check is needed forexecution for modification of the clock/verification, with root authority). In addition, if it is BatteryPlus of the ぴ ろ , the clock and core voltage can be changed according to the battery remaining amount. /proc/driver/w100/fastsysclk: FASTSYSCLK of ATI W100 (read/write) Corresponding type: C700/C750/C760/C860 75: Normal (default) 100: High speed Explanation: The type which loads ATI W100 as a graphic tip/chip (only Cxxx series) validity. When it makes 100, picture drawing accelerates. There are also times when the ghost appears in the picture due to individual difference. When trouble such that the picture is disordered occurs, please use with 75. /proc/power_key_off: Power key suspended mode ON/OFF (read/write) Corresponding type: C700/C750/C760/C860 0: Power key suspended mode OFF (default) 1: Power key suspended mode ON Explanation: For console function /proc/three_button: 3 button modes ON/OFF (read/write) Corresponding type: B500/C700/C750/C760/C860 0: 3 button mode OFF (default) 1: 3 button mode ON Explanation: For console function /proc/tspressure (with SL-C1000/C3000/C3100 /proc/zaurus/tspressure): Writing brush pressure perception ON/OFF (read/write) Corresponding type: B500/C700/C750/C760/C860/C1000/C3000/C3100 0: Writing brush pressure perception OFF (default) 1: Writing brush pressure perception ON Explanation: To use writing brush pressure perception function, the writing brush pressure corresponding application like PetitePeinture is necessary. /proc/driver/w100/rotation: Hardware rotation (read/write) Corresponding type: C700/C750/C760/C860 0: Standard (default) 180: Top and bottom opposition Explanation: For console? Example:

  1. echo 180 > /proc/driver/w100/rotation
  2. echo 0 > /proc/driver/w100/rotation

/proc/version: Version verification (read) Corresponding type: B500/C700/C750/C760/C860/6000/C1000/C3000/C3100 Explanation: Only read-out. Like v13a@tetsu.homelinux.org version is indicated.

FAQ

Whole

  • (Q1-1) Without being dangerous? Hard not being broken?
  • (A1-1) As for me responsibility does not come off. So far you do not hear a story that hard was broken. There was also a thing which () breaks the contents of the built-in flash software depending upon version, but this kind of version is open stopping. Mistaking at the time of kernel replacement job, when you write initrd.bin of the other type, it meaning that restoration is difficult please note sufficiently.


  • (Q1-2) The kernel where sooner or later equal speed comes out of also SHARP release with being done.
  • (A1-2) It is unreasonable. Because it is not possible to recommend the method of using outside rating of the processor the ま っ you question as a manufacturer is.


  • (Q1-3) It isn't not to replace the kernel and the て as for clock rise it to be possible?
  • (A1-3) It is unreasonable. Because the kind of cord/code which inside the kernel is not other than the stipulated clock has entered is. When it is the kernel which applies after 10/27 edition patches of the ぴ ろ it reaches the point where the clock can be modified.


  • (Q1-4) It meaning that the trouble, yyyy with my SL-xxxx occurs, it does, but.
  • (A1-4) If you can teach concretely, you think. Whether or not you can correct, you do not understand. You refer to also Q1-11.


  • (Q1-5) Even in B500 and C700 why using the source for C750/C760/C860? If it is for B500/C700, it doesn't mean that the source for B500/C700 is better?
  • (A1-5) Because the cord/code which corresponds to light/write back cash has entered into only the source for C750/C760/C860 is. Using the source for B500/C700 temporarily, if it does light/write back cash correspondence, becoming equal to the source for C750/C760/C860, stripe now the sushi and the use memory quantity do not decrease and/or do not do. Because also the source of 6000 it is older than the source of C860, the ぴ ろ finite difference of 6000 the port did in the source of C860.


  • (Q1-6) You can use even in C750/C760/C860/C1000/C3100?
  • (A1-6) You prepared, but always it is not the case that it does operation verification.


  • (Q1-7) It is good reprinting binary and the patch? It is good stretching link? It is good making the article?
  • (A1-7) Please freely. The license of the software is GPL.


  • (Q1-8) If which version is used, you do not know whether it is good.
  • (A1-8) It meaning that newer one is improved, you recommend.


  • (Q1-9) The kernel writes and does not substitute.
  • (A1-9) Please verify whether file name has become zImage. When entry succeeds, in Success and indication are ã‚‹ expectation.


  • (Q1-10) As for the value of CCCR either one being fast?
  • (A1-10) When it is other than C1000/C3000/C3100, speaking very roughly, 241 < 161 < 242 < 162 is. However, with 161 and 242, because bus speed of 161 is faster than 242, whether using Katazi with is reversed (161 is good).


  • (Q1-11) We would like to report trouble. We would like to convey opinion and proposition etc..
  • (A1-11) Substantially Linux Zaurus development memo/the individual build kernel - the individual build kernel distributionsight has become that kind of place.

Detailed exposition


  • (Q2-1) The touch screen does not stabilize.
  • (A2-1) "Echo 3230 > please try trying /dev/ts". Please try changing 3230 in between 0 - 5000 with individual difference (ぴ ろ and information from uch). After v11g also negative value is setting possible. When tscalib of the ぴろ is used, it may be found appropriate value simply, is.


  • (Q2-2) After v11b, Bluetooth (BlueZ) became unable to use.
  • (A2-2) Whenbluez-zaurus_2.3-2.4.18c-preempt_arm.ipk is used, BlueZ moves with プリエンプション as effective.

Furthermore, from v14 as for プリエンプション it has become invalid.


  • (Q2-3) It cannot modify CCCR.
  • (A2-3) Because the modification of CCCR is dangerous, after taking backup, please do. The terminal (the console) being similar, at root authority, "echo xxx > /proc/cccr_change" is. "Echo xxx > /proc/cpu/registers/CCCR" is not.


  • (Q2-4) Writing brush pressure is not perceived.
  • (A2-4) Unless "echo 1 > it makes /proc/tspressure", writing brush pressure perception function does not become effective. What it makes invalid is "echo 0 > /proc/tspressure". With SL-C1000/C3000/C3100 it is /proc/zaurus/tspressure.


  • (Q2-5) As for method of using the USB storage.
  • (A2-5) USB storage function is used in the page SL-C760 of humorumviewing. You refer to also Q2-6.


  • (Q2-6) USB storage function does not move well. USB network does not move after the USB storage functional utilizing, trouble occurs e.g., it becomes not be able to restart.
  • (A2-6) The patch of usbd-storeage and please install both parties of USB module.

In case of v14x: The patch of usbd-storage USB module for non preemptive kernel In case of v13x: The patch of usbd-storage USB module for preemptive kernel


The patch of usbd-storage (non preemptive kernel/preemptive kernel common use) Usbd-storage-patch_preempt-1_arm.ipk (preempt it has been attached to file name, but you can use even with the non preemptive kernel) 

USB module for the non preemptive kernel (v14x) -- the ぴ ろ work (mppe module it includes) 
For B500 
For C700 
For C750 
For C760 
For C860
  • (Q2-7) As for method of using mppe.
  • (A2-7) The page of zryviewing.

Only pppd replaces to those of the page of zry. Ppp_mppe.o and ppp_deflate.o module are installed. It tried making the package. Mppe-modules-1_arm.ipk In addition


  • (Q3-1) The method speed rising more was found.
  • (A3-1) Please teach.


  • (Q3-2) This is enormous. We would like to spread in everyone.
  • (A3-2) Please favorite.


  • (Q3-3) As for this the rag it is. That you do not use in everyone, you will say you try to spread but.
  • (A3-3) Please favorite.


Technical information

PXA27x operational frequency

PXA270 is installed in SL-C1000/C3000/C3100 

Turbo-mode frequency = 13MHz * L * N
Run-mode frequency   = 13MHz * L 
System-bus frequency = 13MHz * L/B 
B = 1 (when in fast-bus mode) 
B = 2 (when not in fast-bus mode)

CCCR     L   N   B  Turbo-mode System-bus LCD-freq 
2000210  16  2   1  416MHz     208MHz     104MHz (C1000/C3000/C3100 default) 
2000290  16  2.5 1  520MHz     208MHz     104MHz 
2000310  16  3   1  624MHz     208MHz     104MHz
2000214  20  2   1  520MHz     260MHz      64MHz
2000218  24  2   1  624MHz     312MHz      78MHz
2000214  20  2   1  520MHz     260MHz     130MHz
2000218  24  2   1  624MHz     312MHz     156MHz
Refer to Intel (R) PXA27x Processor FamilyDeveloper's MANUAL for details.

PXA25x operational frequency

PXA250 is installed in SL-B500/C700. PXA255 is installed in SL-C750/C760/C860/6000.

Crystal Frequency = 3.6864MHz Memory Frequency = Crystal Frequency * L Run Mode Frequency = Memory Frequency * M Turbo Mode Frequency = Run Mode Frequency * N PXbus Frequency = Run Mode Frequency/2 CCCR L M N Turbo-mode System-bus 241 27 2 2 398MHz 100MHz (B500/C700 default) 242 32 2 2 472MHz 118MHz 161 27 4 1 398MHz 199MHz (C750/C760/C860/6000 default) 162 32 4 1 472MHz 236MHz 163 36 4 1 531MHz 264MHz Refer to Intel(R) PXA255 Applications Processors Developer 's Manual for details.

Kernel Configuration The SL-B500 = poodle-j, the SL-C700 = corgi, the SL-C750 = shepherd-j, the SL-C760 = husky-j, the SL-C860 = boxer-j, the SL-6000 = tosa-j, the SL-C3000 = spitz-j and the one for the SL-C1000 = akita-j. The one for the SL-C750 is included in the kernel source. The configuration file is corrected based on this file as follows.


Functional table This is from my best recall of memory

;;
Function B500 special kernel (v14x) C700 special kernel (v14x) C750/ C760/ C860 special kernel (v14x) 6000 special kernel (v16x) C1000/C3000/C3100 special kernel (v17x/v18x)
Clock rise Yes
CCCR= 242		 Yes
CCCR= 242		 Proc
CCCR= 161		 Yes
CCCR= 163		 Proc
CCCR= 2000210
Core voltage variable No No No No Proc
SIGSTOP problem coping Yes Yes Yes No Yes
Console Japanese indication Yes Yes Yes No Yes
Writing brush pressure perception Proc Proc Proc Not yet
tosa-presure-wm97xx-2.4.19.diff		 Proc
Touch screen entry Yes Yes Yes No Yes
Keyboard entry Yes Yes Yes No Yes
Hardware rotation
No Proc Proc No No
3 button simultaneous pushing Proc Proc Proc No -
LCD Bufferable No No No No No
Fbcon fix No Yes Yes No Yes
WB cash
Yes Yes - - -
Mouse 3 button emulation
Yes Yes Yes No No
Power key suspention - Proc Proc - -
Ppp-mppe Yes Yes Yes No -
Security patch Yes Yes Yes Yes Yes
Preemptive kernel No No No No No
LOCK break No No No No No
FCSE Yes Yes - - -
NO SHARP LOGO Yes YesYes Yes Yes
MILD SWAP Yes Yes Yes Yes Proc
FASTSYSCLK No Proc Proc No No
Bluetooth up-to-date patch Yes Yes Yes Yes Yes
NTFS corresponding conversion patch No No No No Yes
Key repeat enabling patch - - - - -
Low latency - - - - Yes
Idle POWER saving - - - - Yes
PXA270 Overlay - - - - Yes
NWFPE Yes Yes Yes Yes Yes
Detailed battery No Yes Yes No Yes


山田哲靖 webmaster@tetsu.homelinux.org

Retrieved from "http://www.oesf.org/index.php?title=Tetsu_special_kernel_for_the_6000"
Personal tools