1 Need to process the bad block inode *before* doing the inode scan. 2 3 Also check to see if the first block of the inode table is not on the 4 bad block scan, and fix that. We need to check for an inaccurate 5 blocks, and fix them before we start doing anything else with the 6 filesystem! 7 8 --------------------------------------------------- 9 User request: 10 11 BTW: Could you please add some sort of deleted and possibly corrupted file 12 and inode list to e2fsck report. There should be filenames deleted 13 from directory inodes, files with duplicate blocks e.t.c. 14 It's pretty annoying to filter this information from e2fsck output 15 by hand :- 16 17 ------------------------------------------ 18 19 Add a "answer Yes always to this class of question" response. 20 21 ---------------------------------- 22 23 ext2fs_flush() should return a different error message for primary 24 versus backup superblock flushing, so that mke2fs can print an 25 appropriate error message. 26 27 --------------------------------- 28 Date: Mon, 08 Mar 1999 21:46:14 +0100 29 From: Sergio Polini <s.polini (a] mclink.it> 30 31 32 I'm reading the sorce code of e2fsck 1.14. 33 In pass2.c, lines 352-357, I read: 34 35 if ((dirent->name_len & 0xFF) > EXT2_NAME_LEN) { 36 if (fix_problem(ctx, PR_2_FILENAME_LONG, &cd->pctx)) { 37 dirent->name_len = EXT2_NAME_LEN; 38 dir_modified++; 39 } 40 } 41 42 I think that I'll never see any messages about too long filenames, 43 because "whatever & 0xFF" can never be "> 0xFF". 44 Am I wrong? 45 -------------------------------------- 46 47 Add chmod command to debugfs. 48 49 ------------------------------------------ 50 51 Date: Tue, 18 Jan 2000 17:54:53 -0800 (PST) 52 From: Alan Blanchard <alan (a] abraxas.to> 53 To: tytso (a] MIT.EDU 54 Subject: DEBUGFS - thanks and a feature idea 55 Content-Type: TEXT/PLAIN; charset=US-ASCII 56 57 Theodore: 58 59 First, let me thank you for writing debugfs. Recently, my Linux box 60 (RH 6.0, 400 MHz PIII, on a DSL line) was hacked into. The intruder did 61 an "rm -Rf" on a 34 GB drive with about 5GB of data on it. I was able to 62 restore essentially the entire thing with debugfs and a bit of C code and Perl. 63 Actually, I could have done the entire thing with debugfs and Perl, but I 64 thought it would be too slow. 65 66 During this exercise, I noticed that one small feature was lacking that would 67 have made my job a bit easier. The length of a deleted directory is 68 reported as 0, hence debugfs won't dump the contents of the directory to a 69 file using the "dump" command. The only thing that saved me was that the 70 list of disk blocks is not zeroed out. I was able to dump the contents of the 71 directories by using debugfs to get the relevant block numbers, then 72 using dd to get the actual data. 73 74 If debugfs had a feature where it ignored the size of a directory reported by 75 the inode and instead just dumped all the blocks, it would have facilited 76 things a bit. This seems like a very easy feature to add. 77 78 Again, thanks for writing debugfs (and all the other Linux stuff you've written!). 79 80 Cheers, 81 Alan Blanchard 82 alan (a] abraxas.to 83 84 85 ------------------------------------------------------------------- 86 87 Date: Fri, 21 Jan 2000 14:07:12 -0800 88 From: "H. Peter Anvin" <hpa (a] www.transmeta.com> 89 Subject: mkfs -cc and fsck -c 90 91 b) An option to mkfs to zero the partition. Yes, it can be done with 92 dd, but it would be a nicer way of doing it. 93 94 ------------------------------------------------------------------ 95 96 Add support for in ext2fs_block_iterate() for a returning the 97 compressed flag blocks to block_iterate. Change default to not return 98 EXT2_COMPRESSED_BLKADDR. Change e2fsck to pass this flag in. 99 100 (The old compression patches did this by default all the time, which 101 is bad, since it meant e2fsck never saw the EXT2_COMPRESSED_BLKADDR 102 flagword. 103 104 ------------------------------------------------------------ 105 106 E2fsck should offer to clear all the blocks in an indirect block, not 107 the entire inode, so there's better recovery for when an indirect 108 block gets trashed. 109 110 111 ------------------------------------------------------------- 112 113 From: Yann Dirson - LOGATIQUE <Yann.Dirson (a] France.Sun.COM> 114 Date: Thu, 2 Mar 2000 13:52:13 +0100 (MET) 115 116 During my experiments on the broken system, I noticed the following in 117 the badblocks program (which I'm aware is not designed for IDE drives) 118 - I'd probably have already fixed them if my home system was up :( 119 120 * the syntax summary documents 2nd arg as blocks_count, which should 121 probably read something like end_count. 122 123 * testing past end of device is not detected, and lists those blocks 124 as bad, whereas they simply do not exist. 125 126 127 I think I'll probably add a "max count" option to findsuper(8), so 128 that I do not have to wait for the whole disk to be scanned when the 129 system had to be launched with "init=/bin/sh", in which case Ctrl-[CZ] 130 and friends appear to be absolutely ignored. 131 132 133 Somewhat unrelated, I just noticed the 134 http://web.mit.edu/tytso/www/linux/ext2.html could be updated: 135 136 - could mention SGI xfs (http://oss.sgi.com/projects/xfs/ - they just 137 release 0.03 snapshot) 138 139 ---------------------------------------------------------------- 140 141 Return-Path: <tytso (a] MIT.EDU> 142 Date: Thu, 10 Feb 2000 13:20:14 -0500 143 From: "Theodore Y. Ts'o" <tytso (a] MIT.EDU> 144 To: R.E.Wolff (a] BitWizard.nl 145 In-Reply-To: Rogier Wolff's message of Thu, 10 Feb 2000 08:46:30 +0100 (MET), 146 <200002100746.IAA24573 (a] cave.bitwizard.nl> 147 Subject: Re: e2fsck request for enhancement. 148 Phone: (781) 391-3464 149 150 Date: Thu, 10 Feb 2000 08:46:30 +0100 (MET) 151 From: R.E.Wolff (a] BitWizard.nl (Rogier Wolff) 152 153 Lately, while trying to recover a broken disk, my system froze (twice, 154 until I tried something else) while copying the disk. 155 156 So I had a file of about 50Mb that was growing frantically at the 157 moment of the crash. 158 159 e2fsck, then finds an indirect block that is completely bogus. It 160 starts by asking me if it's ok to clear a few of the referenced 161 blocks. I say yes. Then it comes to the conclusion: 162 163 too many invalid blocks. Clear inode? 164 165 and then I get the option to delete the whole file. Not to truncate 166 the file to a "working" size. 167 168 169 I'd MUCH rather have e2fsck say something like: 170 171 inode 1234 references an invalid block 134345454. Hmm. 172 inode 1234 references 567 out of 50176 invalid blocks, 173 all near the end. Truncate file to 49152 blocks? 174 175 Here you can see that of the 1024 blocks near the end of the file, 176 only 567 were detected as invalid. However now 48Mb of the file will 177 be recovered, instead of thrown away. 178 179 That's a good point. Actually, the right thing is for e2fsck to offer 180 to clear all of the bad blocks in a particular indirect block. I don't 181 know how hard it would be to do that, but I'll put it on my e2fsprogs 182 TODO list. 183 184 - Ted 185 186 --------------------------------------------------------------- 187 From e2fsprogs Debian TODO file as of 1.10-13. 188 189 * Maybe make -dbg packages. Look at how others do it. 190 191 --------------------------------------------------------------- 192 193 Add --lba option to debian icheck command, and have ways of making it 194 easier to translate LBA to filesystem block numbers. 195 196 ------------------------------------------------------- 197 198 199 200 List of projects for e2fsprogs: 201 202 203 1) Make debugfs's "ncheck <inode>" command list all of the pathnames 204 to an inode, not just only the first link to the inode which is found. 205 (A good "intro to libext2fs programming interfaces project) 206 207 Difficulty: Low Priority: Low 208 209 2) Use a code coverage tool such as Rational's PureCoverage to see 210 what kind of code coverage we have for e2fsck, and try to add test 211 cases to increase the code coverage for e2fsck. 212 213 Difficulty: Medium Priorty: Low 214 215 3) Use a code coverage tool such as Rational's PureCoverage to see 216 what kind of code coverage we have for resize2fs, and try to add test 217 cases to increase the code coverage for resize2fs. 218 219 Difficulty: Medium Priorty: Medium 220 221 4) Create a new I/O manager (i.e., test_io.c, unix_io.c, et.al.) which 222 layers on top of an existing I/O manager which provides copy-on-write 223 functionality. This COW I/O manager takes will take two open I/O 224 managers, call them "base" and "changed". The "base" I/O manager is 225 opened read/only, so any changes are written instead to the "changed" 226 I/O manager, in a compact, non-sparse format containing the intended 227 modification to the "base" filesystem. 228 229 This will allow resize2fs to figure out what changes need to made to 230 extend a filesystem, or expand the size of inodes in the inode table, 231 and the changes can be pushed the filesystem in one fell swoop. (If 232 the system crashes; the program which runs the "changed" file can be 233 re-run, much like a journal replay. My assumption is that the COW 234 file will contain the filesystem UUID in a the COW superblock, and the 235 COW file will be stored in some place such as /var/state/e2fsprogs, 236 with an init.d file to automate the replay so we can recover cleanly 237 from a crash during the resize2fs process.) 238 239 Difficulty: Medium Priority: Medium 240 241 5) Create a new I/O manager (i.e., test_io.c, unix_io.c, et.al.) which 242 layers on top of an existing I/O manager which provides an "undo" 243 functionality. This undo I/O manager takes will take two open I/O 244 managers, call them "base" and "undo". The "base" I/O manager is be 245 opened read/write, and when any writes are sent to the I/O manager, 246 the I/O manager will check the "undo" I/O manager, using a file format 247 identical to the one found in (5) above. 248 249 This is useful for allowing e2fsck to create an "undo" file, which 250 would make things like "e2fsck -y" much safer. 251 252 Difficulty: Low (once 5 is done) Priority: Low 253 254 6) Modify resize2fs so that it can relocate and reorganize the 255 filesystem in the following ways: (1) increase the inode size, so that 256 an existing filesystem can use the EA-in-inode kernel patch, (2) 257 reserve blocks in the resize inode to allow for on-line resizing. Use 258 the COW I/O manager described in (5) in order to provide robustness in 259 case of a crash during the resize/reorganization operation. 260 261 Difficulty: High Priority: Medium 262 263 7) Review the EA-in-inode patches to e2fsck for correctness/code 264 cleanliness. (I will probably have to do this myself -- Ted) 265 266 Difficulty: High Priorty: Medium 267 268 8) Add support for extent maps to e2fsprogs. I need to review the 269 extent maps first/in parallel. 270 271 Difficulty: High Priority: Medium 272 273 ---------------------------------- 274 275 Need to deal with the case where the resize inode overlaps with the 276 bad blocks inode. 277 278
