Re: grabbing console (syscons) in kernel
on 13/12/2011 19:49 Andriy Gapon said the following: on 11/12/2011 23:45 Andriy Gapon said the following: There are a few cases when the kernel needs to interact with a user via syscons. This is the cases where the kernel not only spews some output but also expects some input. Some examples are: - asking for a root filesystem specification - entering ddb - asking to press a key for reboot In this cases the kernel implicitly grabs the console for its own use. I'd like to make action more explicit. What do you think about the approach and implementation in the following patches? Thank you! https://gitorious.org/~avg/freebsd/avgbsd/commit/5248b49ebf84d98a0597fa5aa4d813a38f581acc https://gitorious.org/~avg/freebsd/avgbsd/commit/a0849c52242378474bb2eaa41726376fbc4c5bf6 https://gitorious.org/~avg/freebsd/avgbsd/commit/a67515cbd720b16f03ba435ed182966a8a338b15 https://gitorious.org/~avg/freebsd/avgbsd/commit/b8864b68b4c0e26ece065a38301c305833be32eb https://gitorious.org/~avg/freebsd/avgbsd/commit/1017ae425d8abecd7482bd6c6deaaf9f25f5c6cd I was advised that the above links might not be the best way to present the patches for review, so here are them as a single diff file: http://people.freebsd.org/~avg/cngrab.diff I plan to start committing these changes tomorrow. So if this is your area or you want to be sure that nothing gets broken in this area, then please review the changes or let me know your review plans. Post-commit reviews are of course welcome as usual, but pre-commit reviews are preferred. Thank you. P.S. one of the benefits is that a keyboard is put into and out of the polling mode before getting all the required input and after that; not around each character as it is done now in rather twisted way. -- Andriy Gapon ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
Re: grabbing console (syscons) in kernel
on 11/12/2011 23:45 Andriy Gapon said the following: There are a few cases when the kernel needs to interact with a user via syscons. This is the cases where the kernel not only spews some output but also expects some input. Some examples are: - asking for a root filesystem specification - entering ddb - asking to press a key for reboot In this cases the kernel implicitly grabs the console for its own use. I'd like to make action more explicit. What do you think about the approach and implementation in the following patches? Thank you! https://gitorious.org/~avg/freebsd/avgbsd/commit/5248b49ebf84d98a0597fa5aa4d813a38f581acc https://gitorious.org/~avg/freebsd/avgbsd/commit/a0849c52242378474bb2eaa41726376fbc4c5bf6 https://gitorious.org/~avg/freebsd/avgbsd/commit/a67515cbd720b16f03ba435ed182966a8a338b15 https://gitorious.org/~avg/freebsd/avgbsd/commit/b8864b68b4c0e26ece065a38301c305833be32eb https://gitorious.org/~avg/freebsd/avgbsd/commit/1017ae425d8abecd7482bd6c6deaaf9f25f5c6cd I was advised that the above links might not be the best way to present the patches for review, so here are them as a single diff file: http://people.freebsd.org/~avg/cngrab.diff P.S. one of the benefits is that a keyboard is put into and out of the polling mode before getting all the required input and after that; not around each character as it is done now in rather twisted way. -- Andriy Gapon ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
Re: grabbing console (syscons) in kernel
2011/12/11 Andriy Gapon a...@freebsd.org There are a few cases when the kernel needs to interact with a user via syscons. This is the cases where the kernel not only spews some output but also expects some input. Some examples are: - asking for a root filesystem specification - entering ddb - asking to press a key for reboot In this cases the kernel implicitly grabs the console for its own use. I'd like to make action more explicit. What do you think about the approach and implementation in the following patches? Thank you! https://gitorious.org/~avg/freebsd/avgbsd/commit/5248b49ebf84d98a0597fa5aa4d813a38f581acc https://gitorious.org/~avg/freebsd/avgbsd/commit/a0849c52242378474bb2eaa41726376fbc4c5bf6 https://gitorious.org/~avg/freebsd/avgbsd/commit/a67515cbd720b16f03ba435ed182966a8a338b15 https://gitorious.org/~avg/freebsd/avgbsd/commit/b8864b68b4c0e26ece065a38301c305833be32eb https://gitorious.org/~avg/freebsd/avgbsd/commit/1017ae425d8abecd7482bd6c6deaaf9f25f5c6cd P.S. one of the benefits is that a keyboard is put into and out of the polling mode before getting all the required input and after that; not around each character as it is done now in rather twisted way. Does asking for geli password is such case too? For example, I have mouse, which can disappear/reappear while laying still; and when geli asks password, ums0 connect message can appear, which moves line Please enter geli password... to top, and it can be missed; This behavior somehow change? -- Andriy Gapon ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org -- Regards, Alexander Yerenkow ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
Re: grabbing console (syscons) in kernel
on 12/12/2011 00:06 Alexander Yerenkow said the following: 2011/12/11 Andriy Gapon a...@freebsd.org mailto:a...@freebsd.org There are a few cases when the kernel needs to interact with a user via syscons. This is the cases where the kernel not only spews some output but also expects some input. Some examples are: - asking for a root filesystem specification - entering ddb - asking to press a key for reboot In this cases the kernel implicitly grabs the console for its own use. I'd like to make action more explicit. What do you think about the approach and implementation in the following patches? Thank you! https://gitorious.org/~avg/freebsd/avgbsd/commit/5248b49ebf84d98a0597fa5aa4d813a38f581acc https://gitorious.org/~avg/freebsd/avgbsd/commit/a0849c52242378474bb2eaa41726376fbc4c5bf6 https://gitorious.org/~avg/freebsd/avgbsd/commit/a67515cbd720b16f03ba435ed182966a8a338b15 https://gitorious.org/~avg/freebsd/avgbsd/commit/b8864b68b4c0e26ece065a38301c305833be32eb https://gitorious.org/~avg/freebsd/avgbsd/commit/1017ae425d8abecd7482bd6c6deaaf9f25f5c6cd P.S. one of the benefits is that a keyboard is put into and out of the polling mode before getting all the required input and after that; not around each character as it is done now in rather twisted way. Does asking for geli password is such case too? Yes. For example, I have mouse, which can disappear/reappear while laying still; and when geli asks password, ums0 connect message can appear, which moves line Please enter geli password... to top, and it can be missed; This behavior somehow change? No. The patches do not contain user-visible changes. They make things easier for keyboard drivers and hopefully provide a more logical interface to kernel. Their primary target is the input path, they do not introduce any changes to the output path. -- Andriy Gapon ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org