4 ms·

Interesting! I got curious about where they come from, so I dug a little. Here is a non-exhaustive breakdown: - 23 from crypto code: https://sourcegraph.com/se

Interesting! I got curious about where they come from, so I dug a little. Here is a non-exhaustive breakdown: - 23 from crypto code: https://sourcegraph.com/search?q=repo:%5Egithub%5C.com/torvalds/linux%24+TODO.*%5C:+case:yes+count:1000+file:crypto/&patternType=regexp https://sourcegraph.com/search?q=repo:%5Egithub%5C.com/torva... - 2380 from driver code: https://sourcegraph.com/search?q=repo:%5Egithub%5C.com/torvalds/linux%24+TODO.*%5C:+case:yes+count:10000+file:%5Edrivers/&patternType=regexp https://sourcegraph.com/search?q=repo:%5Egithub%5C.com/torva... - 73 from ARM arch code: https://sourcegraph.com/search?q=repo:%5Egithub%5C.com/torvalds/linux%24+TODO.*%5C:+case:yes+count:10000+file:%5Earch/arm&patternType=regexp https://sourcegraph.com/search?q=repo:%5Egithub%5C.com/torva... - 43 from x86 arch code: https://sourcegraph.com/search?q=repo:%5Egithub%5C.com/torvalds/linux%24+TODO.*%5C:+case:yes+count:10000+file:%5Earch/x86&patternType=regexp https://sourcegraph.com/search?q=repo:%5Egithub%5C.com/torva... - 114 from other arch code: https://sourcegraph.com/search?q=repo:%5Egithub%5C.com/torvalds/linux%24+TODO.*%5C:+case:yes+count:10000+-file:%5Edrivers/+-file:crypto+-file:Documentation+-file:%5Earch/arm+-file:%5Earch/x86+file:%5Earch/&patternType=regexp https://sourcegraph.com/search?q=repo:%5Egithub%5C.com/torva... - 606 from block IO, FS, networking, and other sources: https://sourcegraph.com/search?q=repo:%5Egithub%5C.com/torvalds/linux%24+TODO.*%5C:+case:yes+count:10000+-file:%5Edrivers/+-file:crypto+-file:Documentation+-file:%5Earch/&patternType=regexp https://sourcegraph.com/search?q=repo:%5Egithub%5C.com/torva...

booleanbetrayal 7y ago
Without auditing, driver code makes sense to me. Guessing there is a lot of vendor churn and therefore a lot of caveats to keep track of ... right up until the point the driver is obsolete and put on life support, never be looked at again.
- simias 7y ago
  Or simply the features are never needed so they are never implemented. There's rarely a perfect match between the hardware capabilities and the OS interface, that's the cost of the abstraction. Sometimes there's no trivial way to expose the feature or it would require too much work for something that's not deemed necessary.
- pm215 7y ago
  Also driver and other hardware-specific code is more likely to have a smaller population of developers to tend and improve it, and a smaller population of users to care about getting it fixed, so minor issues will linger longer. (I had a look at the 'arm' ones, and many of those are in code specific to a particular board or other piece of hardware.)
harryf 7y ago
Herbert XU is apparently still in hiding after receiving threats of a good kicking ... https://sourcegraph.com/github.com/torvalds/linux@fd6988496e79a6a4bdb514a4655d2920209eb85d/-/blob/drivers/crypto/hifn_795x.c#L2030:11 https://sourcegraph.com/github.com/torvalds/linux@fd6988496e...

booleanbetrayal 7y ago

Without auditing, driver code makes sense to me. Guessing there is a lot of vendor churn and therefore a lot of caveats to keep track of ... right up until the point the driver is obsolete and put on life support, never be looked at again.

simias 7y ago
Or simply the features are never needed so they are never implemented. There's rarely a perfect match between the hardware capabilities and the OS interface, that's the cost of the abstraction. Sometimes there's no trivial way to expose the feature or it would require too much work for something that's not deemed necessary.
pm215 7y ago
Also driver and other hardware-specific code is more likely to have a smaller population of developers to tend and improve it, and a smaller population of users to care about getting it fixed, so minor issues will linger longer. (I had a look at the 'arm' ones, and many of those are in code specific to a particular board or other piece of hardware.)

harryf 7y ago

Herbert XU is apparently still in hiding after receiving threats of a good kicking ... https://sourcegraph.com/github.com/torvalds/linux@fd6988496e79a6a4bdb514a4655d2920209eb85d/-/blob/drivers/crypto/hifn_795x.c#L2030:11 https://sourcegraph.com/github.com/torvalds/linux@fd6988496e...