I have 3 Easy2Boot USB drives with many ISOs on it. There seems to be a problem with my external USB SSD drive (which is in caddy) and my external 80gb SATA HDD in a different caddy. The 2TB My Passprot USb HDD seems OK with all versions (as far as I can tell so far)

I have tried various versions: 2014-08-19 OK 2014-08-12 OK 2014-09-18 BAD

The problem occurs when booting from my Asus EeePC 904HD and on the USB HDD caddy also under VBox+VMUB.

Problems vary - e.g. only some ISO files are detected (or no ISO files are detected and added to the menu), or if I reload the /menu.lst file, hd0,0 becomes inaccessible

e.g. in command console root >> Disk I/O Error ls (hd0,0) >> Disk I/I Error

something is broken or at least not compatible with 0.4.5c and earlier versions of 0.4.6a!

To reproduce, make an E2B USB HDD/SSD drive. Add 0.4.6a 2014-09-18 grldr to the root. Add some ISOs to _ISO\MAINMENU test on a variety of systems

Something is also odd with 0.4.5c latest version - similar problems

e.g. copy ISO to Easy2Boot _ISO\ANTIVIRUS\clonzilla.iso boot in Vbox type Ctrl-A for Antivirus menu

Error 30: Invalid argument

works OK with 2014-09-08 version

Please try this first

I think problem with 0.4.5c may be caused by incompatibility with older versions regarding use of double-quotes

set HPAD=\x20\x20 echo "%HPAD:~0,1%"

"%HPAD:~0,1%" (new versions)

"\" (old versions)

https://github.com/chenall/grub4dos/commit/0618759937ea55e192a4e515daa4d16458379a2b

Think problem started here?

@Steve6375 not problem here

did you try grub4dos-0.4.5c-2014-09-21.7z ?

Yes, that is the version that fails for HPAD test

try

if "%HPAD:~0,1%"==" " echo fred

ok.

OK - that fixed the problem with 0.4.5c.

0.4.6a

grub4dos-0.4.6a-2014-09-22.7z - looks good!

dpms works xp no disks found issue fixed if "%HPAD:~0,1%"==" " echo fred issue fixed

Still a problem with both 0.4.5c and 0.4.6a Asus EeePC External USB 3 caddy + 230GB SSD

set a=1 2 if "%a%"=="" echo fred

Error 30: Invalid argument

but OK if use WD Passport USB 2TB HDD with EeePC or 80GB external USB hdd with EeePC. All OK on VBox.

Very strange!!!

If CheckAccess.g4b is not called then no problem with SSD on EeePC. I add extra code to fond error

code exits batch file after XXXXXXXXXXXXX4 and never gets to pause 4

EeePC has 137GB bug - so it will return an error on cat line.

echo Last Sector of last partition of %driv% is %lba% %redir% echo -n Checking last sector is accessible... %redir% cat --length=16 --hex %driv%%%lba%+1 > nul && set result=1

pause pre4 if "%a%"=="" echo XXXXXXXXXXXXX4 pause 4

grldr 0.4.5c 20140908 works OK and does not exit - it runs whole batch file

HDD NTFS

Fails on CheckAccess.g4b (if remove > nul from cat line) Fatal! Inconsistent data read from (0x80)488391961+127 Fatal! Inconsistent data read from (0x80)488392056+1 Fatal! Inconsistent data read from (0x80)488391961+127 Fatal! Inconsistent data read from (0x80)488392056+1 pause pre4

if add 0x4000000 to lba then it does NOT fail! cat --length=16 --hex (hd0)555500920+1 (remove > nul from cat line so can see errors).....

Fatal! Inconsistent data read from (0x80)555500794+127 Fatal! Inconsistent data read from (0x80)555500928+1 Fatal! Inconsistent data read from (0x80)555500794+127 Fatal! Inconsistent data read from (0x80)555500928+1 pause pre4

adding 0x10000 >> passes adding 0x1000 >> FAILS

SSD Disk has Last LBA 488,397,167

if use cat --hex and then if "%a%"=="" echo fred

If access is made to < Last LBA 488,397,167 then we get error situation. If access is made > Last LBA 488,397,167 then does not cause error situation (presumably BIOS knows LBA is > end of disk and does not actually attempt to access disk at all).

So this indicates bad registers returned by BIOS on disk error??

This could be similar to Issue 198?? https://code.google.com/p/grub4dos-chenall/issues/detail?id=198&can=1&start=100

Is there which version can be used?

Please upload the memory info..using grub4dos displaymem

grab1.txt using 22/09/2014

0.4.6a 2014-08-19 OK 2014-08-12 OK 2014-09-18 BAD

0.4.5c 20140908 OK

can you please try below version? grub4dos-0.4.5c-2014-09-16.7z grub4dos-0.4.6a-2014-09-12.7z grub4dos-0.4.6a-2014-09-08.7z grub4dos-0.4.6a-2014-09-06.7z grub4dos-0.4.6a-2014-09-03.7z

grub4dos-0.4.5c-2014-09-16.7z OK but screws up later grub4dos-0.4.6a-2014-09-12.7z OK but screws up later grub4dos-0.4.6a-2014-09-08.7z OK but screws up later grub4dos-0.4.6a-2014-09-06.7z OK but screws up later grub4dos-0.4.6a-2014-09-03.7z OK but screws up later

Problem with 'screw up later' is that

?_BOOT variable is now junk!

echo %?_BOOT%

displays a load of dc A5

characters

i.e. I booted to grub4dos menu I ran checkaccess.g4b as a menu item First time it correctly reported BIOS BAD After that it would not run because ?_BOOT was garbage

Here is my test menu for the problem(s)

title Check BIOS Disk Access\n Checks BIOS can access all of disk.\n Some BIOSes cannot access a USB disk past 137GB /%grub%/CheckAccess.g4b errorcheck on set a=1 2 if "%a%"=="" echo fred pause errorcheck off boot

I think this is similar to Issue 198. Please try new and use @boot instead ?_BOOT

grub4dos-0.4.5c-2014-09-26.7z OK

if I use @boot

Actually there is still a problem!

The CheckAccess.g4b runs OK, but when it returns I get an Error 27: Unrecognised command on the command following the call to /checkaccess.g4b (see menu in post #19)

if I use /%grub%/CheckAccess.g4b ;; echo s

then it echos 's' and I get no error

/%grub%/CheckAccess.g4b && echo s >> no error (no s)

/%grub%/CheckAccess.g4b || echo s >> no error (s is dispayed)

Have you try new?

yes! See #21 and #22

Attached my test files - copy to root

I use Asus 904HA EeePC Boot from external 230GB Samsung SSD

CheckAccess runs OK but it returns I get a Error 27: Unrecognised command

and it does not run the rest of the lines (i.e. echo 1 etc etc and pause is not run). It then returns to the menu.

also, when it returns to the menu, if I do a c for command line

echo %?_BOOT%

displays a load of garbage characters - so ?_BOOT has been corrupted.

P.S. When I say CheckAccess runs OK, I mean it runs and says the BIOS is buggy, as expected.

Same test using 2TB USB 3.0 WD MyPassport drive shows no problem on same Asus 904HA EeePC in same USB port. i.e. 2TB USB drive >> CheckAccess reports a Bad BIOS Then echoes 1 ?_BOOT=(cd) @boot=(hd0,0)

so it works correctly!

This is strange? 230GB SSD drive shows a problem (formatted as NTFS or FAT32), but 2TB HDD does not (even though both report BAD BIOS in CheckAccess.g4b and both tested on same EeePC)!

Note: if I remove setlocal from CheckAccess.g4b then still get same problem. So it is nothing to do with setlocal. Other variables are still present if I do a set command, but ?_BOOT is corrupted.

Comment deleted

I see your attach of grldr version is old.

Please download this version. http://grub4dos.chenall.net/downloads/grub4dos-0.4.5c-2014-09-26/

and there maybe bug in bios,Please try modify variable_base_addr(0x307FF4).

e.g. write 0x307ff4 0xC00000 && set *

There is little difference with the new 2014-09-26 and the old 2014-09-26.

With new version ?_BOOT is lost after running the menu, with older version it is corrupted (full of NNNN). I still get the 'Error 27: Unrecognised command' error.

If I use write 0x307ff4 0xC00000 && set * then ?_BOOT is lost, so I used write 0x307ff4 0xC00000 && set ?_BOOT=%?_BOOT%

In this case, I still get 'Error 27: Unrecognised command' after CheckAccess.g4b is run and 'echo 1' and the other lines do not execute. ?_BOOT returns (hd0,0).

If test with 0.4.6a 2014-09-23 I get same problem. i.e. CheckAccess reports Bad BIOS and then get 'Error 27: Unrecognised command' error. echo 1 and other commands not executed. Also ?_BOOT is corrupted (full of NNNNN...)

I tested with grub4dos-0.4.5c-2014-09-04.7z safe_int13 version

Same symptom as #32

I tested older version of 0.4.5c

It seems bug was introduced in 2013-03-20 version

FAIL = same symptom as #32

2014-08-19 FAIL 2013-06-30 FAIL 2014-05-16 FAIL 2013-04-19 FAIL 2013-03-29 FAIL 2013-03-20 FAIL 2013-03-12 OK 2012-12-31 OK 2012-03-21 OK

These are two test files I used for checking the problem

More testing - If I comment out cat --length=16 --hex %driv%%%lba%+1 > nul && set result=1

then code works, so problem is caused by disk access returning error + change in 2013-03-20??

If I set errorcheck off before cat --length=16 --hex %driv%%%lba%+1 line then 2013-03-20 almost works except that ?_BOOT is still corrupted.

so bug in 2013-03-20 must be to do with error status being returned by command.

If I use 2013-03-12 then all works and ?_BOOT is not corrupt.

I can not found problem in code. Try this please.

Same problem! Error 27: Unrecognized command and ?_BOOT corrupted

(Test3.zip from #35 used for test on Asus EeePC 904HA)

It works perfectly if I use 23-03-2012 !!!

Can you make a special version which removed the ;; code so that it is the same as 23-03-2012 but with all newer changes and I will test?

removed the ;; code support.

That still fails in exactly the same way!

So what other changes were there between 2013-03-20 FAIL 2013-03-12 OK

???

Please can you check the download in #40 - is it correct version with ;; removed?

How come

echo 1 ;; echo 2

works???

Sorry!this please.

That version also fails the test and echo 1 ;; echo 2 does not work. So the problem must be somewhere else???

https://github.com/chenall/grub4dos/commit/e2b5ce08d11687f7ec21ffc989113ea501f1c69a#diff-3f10e611bd6c650328f5c927c5b50bdaR184

Both of those work! What did you change?

just delete below line. if (cmd[0] == cmd[1] && cmd[2] != 0x20) continue; Please try this again.

That fails in the same way as before.

Thanks for you test,I don't know why.

Please try below again ,hope both can work.

test1 WORKS test2 FAILS

Ok,Thank you for your patience tested again..

workaround for this issue is remove following two lines.

if (cmd[0] == cmd[1] && cmd[2] != 0x20) continue;

the new version is uploaded http://grub4dos.chenall.net

Something odd with this new version

In VBOX I get FATAL: int13_hardisk_ext: function 42. can't use 64bits lba

On Asus EeePC, Checkaccess.g4b (in E2B, not test version) stops on 'Checking last sector is accessible...' for about 2 minutes before continuing.

???

grub4dos-0.4.5c-2014-10-12-test1.7z also fails with the same FATAL:: int13 error.

grub4dos-0.4.6a-2014-10-12.7z seems OK but not 0.4.5c

if run from USB 32Gb flash drive under VBox, grub4dos-0.4.6a-2014-10-12.7z does hang on DPMS2 in Easy2Boot (0.4.5c does not hang and works, but get FATAL: error if booting from 2TB USb HDD under Vbox).

I think issue with VBox and 2 minute delay on EeePC was caused by accessing sectors past the end of the disk. The code in CheckAccess.g4b uses lines like

set /a s1=*0x6001c6 & 0xffffffff >>2

This seems to work (i.e. divides number by 4) with some disks but not others (depending on size of partitions?)! Changing code to remove the space to

set /a s1=*0x6001c6 & 0xffffffff>>2

works OK and always divides by 4. So on some drives, the lba sector to access was x4 greater than the end of the disk and this was causing the issues in VBox and EeePC - 2 minute delay.

So 0.4.5c looks OK.

Also tested 0.4.6a 2014-10-15 and this now works with DPMS - so problems look fixed. I will continue to test but so far both versions look OK... Thanks!

If has any other problems please report issues at github.(because Googlecode is hard to access from china)

https://github.com/chenall/grub4dos/issues

so far testing is good! No problems. Will you be releasing this new version and 0.4.6a version?

FYI - This problem is only fixed by using

write 0x307ff4 0xC00000 && set ?_BOOT=%?_BOOT%

on first line of menu.lst

If this line is omitted then ?_BOOT is corrupted to all NNNNNs with sedillas as before.

Also, if I use the write 0x307ff4 0xC00000 && set ?_BOOT=%?_BOOT% line in a VM or any system, then on first boot, I get splashimage 800x600 as usual. However, if I reboot VBox or QEMU then it does not change to 800x600 graphicsmode and no splashimage! This is when using the 'reboot' command. Really strange!

I tried different locations - e.g. 0x1200000 and same result - boot OK first time but on reboot -> no graphicsmode!