:)

[lsp199308]

No description provided.

[borntohonk]

also remove lines:
https://github.com/impeeza/sys-patch/blob/master/overlay/src/main.cpp#L106
https://github.com/impeeza/sys-patch/blob/master/overlay/src/main.cpp#L132

edit lines, save
git add overlay/src/main.cpp
git commit --amend --no-edit
git push -f

or just new commit message and push

[borntohonk]

also comment:

on 18.1.0 https://github.com/impeeza/sys-patch/blob/master/sysmod/src/main.cpp#L256

this hits function 0x74310 instead of 0x743BC

and must be changed to:

    { "nocntchk", "0x94081C00121F050071..0054", -3, 0, bl_cond, ret0_patch, ret0_applied, true, MAKEHOSVERSION(10,0,0), MAKEHOSVERSION(18,1,0) },

[borntohonk]

I mirrored this logic here:
borntohonk/Switch-Ghidra-Guides@4efdfc3
also here:
borntohonk@dbe97b6

[borntohonk]

also i notice @impeeza is uploading .7z release, why?

do "make dist" and it will create sys-patch.zip

https://github.com/impeeza/sys-patch/blob/master/Makefile#L51-L56

[lsp199308]

also comment:

on 18.1.0 https://github.com/impeeza/sys-patch/blob/master/sysmod/src/main.cpp#L256

this hits function 0x74310 instead of 0x743BC

and must be changed to:

    { "nocntchk", "0x94081C00121F050071..0054", -3, 0, bl_cond, ret0_patch, ret0_applied, true, MAKEHOSVERSION(10,0,0), MAKEHOSVERSION(18,1,0) },

{ "10-18", "0x40f9...9408.0012.050071", 2, 0, bl_cond, ret0_patch, ret0_applied, true, MAKEHOSVERSION(10,2,0), MAKEHOSVERSION(18,1,0) },//fw10.2.0-fw18.1.0
{ "19", "0x40f9...94..40b9..0012", 2, 0, bl_cond, ret0_patch, ret0_applied, true, MAKEHOSVERSION(19,0,0) },

[lsp199308]

{ "1-9A", "0x........0036..00b4..40b9", 2, 0, bl_cond, ret0_patch, ret0_applied, true, MAKEHOSVERSION(1,0,0), MAKEHOSVERSION(9,2,0) },
{ "1-9B", "0x....0094.0210911f000072", 2, 0, bl_cond, ret0_patch, ret0_applied, true, MAKEHOSVERSION(1,0,0), MAKEHOSVERSION(9,2,0) },
{ "1-9C", "0x.40f9...94081c00121f050071", 2, 0, bl_cond, ret0_patch, ret0_applied, true, MAKEHOSVERSION(1,0,0), MAKEHOSVERSION(9,2,0) },
//补丁：1f2003d5
{ "1-9D", "0x.97..0036881e42b9", 2, 0, tbz_cond, nop_patch, nop_applied, true, MAKEHOSVERSION(1,0,0), MAKEHOSVERSION(9,2,0) },
//10.0.0
//补丁：e0031f2a
{ "10A", "0x0091c5b10194fd7b43a9f44f42a9", 2, 0, bl_cond, ret0_patch, ret0_applied, true, MAKEHOSVERSION(10,0,0), MAKEHOSVERSION(10,1,0) },
{ "10-18A", "0x40f9...9408.0012.050071", 2, 0, bl_cond, ret0_patch, ret0_applied, true, MAKEHOSVERSION(10,2,0), MAKEHOSVERSION(18,1,0) },
{ "19A", "0x40f9...94..40b9..0012", 2, 0, bl_cond, ret0_patch, ret0_applied, true, MAKEHOSVERSION(19,0,0) },

//补丁：1f2003d5
{ "10D", "0xe80a009008613b91280000f9c0035fd6", 2, 0, tbz_cond, nop_patch, nop_applied, true, MAKEHOSVERSION(10,0,0), MAKEHOSVERSION(10,1,0) },
{ "10-16D", "0x0036.......71..0054..4839", -2, 0, tbz_cond, nop_patch, nop_applied, true, MAKEHOSVERSION(10,0,0), MAKEHOSVERSION(16,1,0) },
{ "17-19D", "0x.94..0036.258052", 2, 0, tbz_cond, nop_patch, nop_applied, true, MAKEHOSVERSION(17,0,0), },//fw17-fw19

};

[lsp199308]

This doesn't apply to EXFAT, exfat I couldn't find the wildcard, so this item applies to 17+

[borntohonk]

{ "10-18", "0x40f9...9408.0012.050071", 2, 0, bl_cond, ret0_patch, ret0_applied, true, MAKEHOSVERSION(10,2,0), MAKEHOSVERSION(18,1,0) },//fw10.2.0-fw18.1.0

40 f9 .. .. .. 94 08 .. 00 12 .. 05 00 71 + 0x2 does hit the correct offset on 18.1.0 at least

{ "19", "0x40f9...94..40b9..0012", 2, 0, bl_cond, ret0_patch, ret0_applied, true, MAKEHOSVERSION(19,0,0) },

40 f9 .. .. .. 94 .. .. 40 b9 .. .. 00 12 + 0x2 also hits correct offset for 19.0.0

[impeeza]

also i notice @impeeza is uploading .7z release, why?

do "make dist" and it will create sys-patch.zip

https://github.com/impeeza/sys-patch/blob/master/Makefile#L51-L56

Sorry about that was so slept, editing now.

[impeeza]

Hello there, Thanks a lot for all the work you are investing on.

This is a great optimization. It's ready to be mereged? it looks so nice!

[lsp199308]

This is the unfixed area:
I've uploaded the unpatched binary so that someone can find the wildcards
You can use ida or ghidra to look for wildcards.
#FS 6.0.0-ExFAT.
[FS:330553F6B5FB55C4]
.nosigchk=0:0x7C934:0x4:07DFFF97,E0031F2A
.nosigchk=0:0x7C9A8:0x4:8E3E0094,E0031F2A
.nosigchk=0:0xF678C:0x4:C0030036,1F2003D5
.nosigchk=0:0x81784:0x4:73D40194,E0031F2A

#FS 7.0.0-ExFAT
[fs:2cce659cec536a8e]
.nosigchk=0:0x7FFA4:0x4:33D7FF97,E0031F2A
.nosigchk=0:0x7FFDC:0x4:31430094,E0031F2A
.nosigchk=0:0xFDB94:0x4:C0030036,1F2003D5
.nosigchk=0:0x85E18:0x4:C6DF0194,E0031F2A

#FS 8.0.0-ExFAT
[fs:dbd941c0c53c52cc]
.nosigchk=0:0x81884:0x4:BBD6FF97,E0031F2A
.nosigchk=0:0x818BC:0x4:51440094,E0031F2A
.nosigchk=0:0xFFF54:0x4:C0030036,1F2003D5
.nosigchk=0:0x87978:0x4:DEE10194,E0031F2A

#FS 8.1.0-ExFAT
[FS:B4CAE1F24965D92E]
.nosigchk=0:0x81884:0x4:BBD6FF97,E0031F2A
.nosigchk=0:0x818BC:0x4:51440094,E0031F2A
.nosigchk=0:0xFFF54:0x4:C0030036,1F2003D5
.nosigchk=0:0x87978:0x4:DEE10194,E0031F2A

#FS 14.0.0-ExFAT
[FS:D488D1F29217355C]
.nosigchk=0:0x13354C:0x4:A0070036,1F2003D5
.nosigchk=0:0x79C00:0x4:2DE70294,E0031F2A

#FS 15.0.0-ExFAT
[fs:34c0d9ed6ad1873d]
.nosigchk=0:0x06EBA4:0x4:40F50294,E0031F2A
.nosigchk=0:0x12BD7C:0x4:A0040036,1F2003D5

#FS 16.0.0-ExFAT
[FS:CFAB450C2C539DA9]
.nosigchk=0:0x06F704:0x4:E9900394,E0031F2A
.nosigchk=0:0x15380C:0x4:60040036,1F2003D5

#FS 16.0.3-ExFAT
[fs:62c65efd9abf7c43]
.nosigchk=0:0x06F754:0x4:E9900394,E0031F2A
.nosigchk=0:0x15385C:0x4:60040036,1F2003D5

Uploading exfat.zip…

[borntohonk]

This is the unfixed area: I've uploaded the unpatched binary so that someone can find the wildcards You can use ida or ghidra to look for wildcards.

So I have a very hot opinion here.

Maintaining compatibility down to 10.0.0 is good enough.

10.0.0 released april 14 2020.

4 years ago.

most important is current firmware, and maybe down to 17.0.0 (all are supported currently, no reason to do much more work)

[lsp199308]

This is the unfixed area: I've uploaded the unpatched binary so that someone can find the wildcards You can use ida or ghidra to look for wildcards.

So I have a very hot opinion here.

Maintaining compatibility down to 10.0.0 is good enough.

10.0.0 released april 14 2020.

4 years ago.

most important is current firmware, and maybe down to 17.0.0 (all are supported currently, no reason to do much more work)
You're right. I couldn't agree more.

[impeeza]

@borntohonk correctme if I am wrong.

using as starting point IPS patch (or boot.ini enties) is no easy to create SYS-Patch patterns, they are two different approachs.

[borntohonk]

@borntohonk correctme if I am wrong.

using as starting point IPS patch (or boot.ini enties) is no easy to create SYS-Patch patterns, they are two different approachs.

ips patches contain offsets and what patch it writes to the offset, you could just as easily open the binary in ghidra, go to the offset, see whats there and create a pattern (diffing several firmwares give better patterns,)

(same applies to patches.ini with hekate, nothing there cannot be looked up)

alternatively you can just do with one and instead pattern arm instructions, and then some static bytes

[impeeza]

@borntohonk Is This PR ready to be merged?

[borntohonk]

@borntohonk Is This PR ready to be merged?

🤷 it's not my PR, but it's about as ready as it can be i guess

[lsp199308]

Okay, I'm sure it's fine.

[impeeza]

will review this the next week I come back to home and let you to know.

[borntohonk]

i personally prefer having the functions labelled over "9a 9b 9c 9d", but the existing labels weren't consistent either.

[impeeza]

Merged on PR9

[impeeza]

This changes was merged on the PR9, Thanks a lot for all the work.