[opensuse-kernel] Re: [LOG] kernel-source branch users/ohering/master/for-next created. ceec9f0
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 12/1/15 9:24 AM, Olaf Hering wrote: > The branch users/ohering/master/for-next has been created at > ceec9f0a394b865d7a4ab0b0dd264805f7ede727 (commit) > > - Log > ----------------------------------------------------------------- > > commit ceec9f0a394b865d7a4ab0b0dd264805f7ede727 Author: Olaf > Hering <ohering@suse.de> Date: Tue Dec 1 15:22:12 2015 +0100 > > ppc64: Revert from 64K to 4K PAGE_SIZE to fix firewire and nouveau > on G5 > > ---------------------------------------------------------------------- - - > > > This seems like a pretty big change just to fix two drivers. Isn't the correct solution to fix the drivers? In SLE12, we only support ppc64le, but in openSUSE, are we certain that there aren't ppc64 big-endian users out there and that this change to fix drivers they might not even use is acceptable? - -Jeff - -- Jeff Mahoney SUSE Labs -----BEGIN PGP SIGNATURE----- Version: GnuPG/MacGPG2 v2.0.19 (Darwin) Comment: GPGTools - http://gpgtools.org iQIcBAEBAgAGBQJWXa4aAAoJEB57S2MheeWyi7YQAKHtW4Bq8Aiv4ohouHl+GZKf S5vWSavRuDYYCgREgXMkuKnBTzRgoTjepNdrb8gffkOWl2/pH5sOn71KNyl5Ies6 6nNzJlJB+JFGG073H7ZLuHPOa3yiAP1/wsBDPPYBbofc2X3JtUNnq5ppPWhTKDcw cuEqEz6ePsUjjgeB8HhnRxdjtZ9SS5qUeGEIKULSuuBbgoDpSXRhEXNdc36Mvi84 NMl8wzTw/gAhFMy/41K4HBWLp32Z4L6rqcoXhYGnMpVeP4u6bDQelfuhUGYYTeVG 0OQbi6KUgDeZ88UnLu5z+1ZJddsjsOIGXJzb8eNmeM/2dL6HLE793DJovX29X73T PCVKn+rbUG+du6dQuoeBLTwsLatsJPZliiO7QxKu2TSTBm1+VdD0gLly1VVCwT1i JLi7qovvp2v67RsJc0bmKvVj8FGRwu5uMhKL8ggqfBWRfXRHCSMdN06lb9Wwqf7R R4AFv9xYqB4kW8/sl3WKzO+qmAXhzbLCdE+qUKxoF0u+RUbHBJbaMJCu+V6Q+ye8 +35VoHMoDqYdz71c0WYambF8NNeqlX1YbzZJ4DURF/FdrY5/QaMaXPjC3o3HaVbn d5l86cUM7fjasQ9xIRxJK7EvSmMrY4CfpIbML2QMaXfngwBX81gKIMFd9CP82W58 9OesJ3bil1vh8LsrQ7Yi =/FUP -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
On Tue, Dec 01, Jeff Mahoney wrote:
This seems like a pretty big change just to fix two drivers. Isn't the correct solution to fix the drivers?
If there are people out there who are able to fix this. The firewire bug was open for years on kernel.org.
In SLE12, we only support ppc64le, but in openSUSE, are we certain that there aren't ppc64 big-endian users out there and that this change to fix drivers they might not even use is acceptable?
What wil they lose if the kernel starts to use a different pagesize, compared to the gain? Olaf -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 12/1/15 9:43 AM, Olaf Hering wrote:
On Tue, Dec 01, Jeff Mahoney wrote:
This seems like a pretty big change just to fix two drivers. Isn't the correct solution to fix the drivers?
If there are people out there who are able to fix this. The firewire bug was open for years on kernel.org.
In SLE12, we only support ppc64le, but in openSUSE, are we certain that there aren't ppc64 big-endian users out there and that this change to fix drivers they might not even use is acceptable?
What wil they lose if the kernel starts to use a different pagesize, compared to the gain?
Did you look at the fallout from changing the page size to 4k? There will be significant performance impact on larger systems. Transparent huge pages are disabled now. The kernel will need to maintain 16 times as many page table entries as with 4k pages. There's a reason we chose the 64k page size ages ago and I don't think two broken drivers is a good reason to throw performance out the window especially since that hardware won't exist on any non-Apple hardware. - -Jeff - -- Jeff Mahoney SUSE Labs -----BEGIN PGP SIGNATURE----- Version: GnuPG/MacGPG2 v2.0.19 (Darwin) Comment: GPGTools - http://gpgtools.org iQIcBAEBAgAGBQJWXbNMAAoJEB57S2MheeWym3AQAJkFszRzyVs+OIKdVc6jEOdE CWCOPbhpL0SK5WpL+rZAVR3uia6hYWxvLa4aXrUUKNev7fTYwJDySMDTI4Aal2km eF0D/ejgbRmz7yGM1v/jMYlHOIrEuMOYFJYZR0kzWUEgJkcoAzmGCnySQxFtemIt GaAWkcH71bZOrCGP1Fvsgm2XH+tDDhtluwuE0p65p0dgBvYkcpAM9ip7M5Hup3kb Ro3d71w29zpWtYuOH+r+ynVFjFbx5jrXtaNAilIECP2rZdMUwTylEIfSadVcbL/u HfrZsii3xgG++oV89iCID3dqv0ibXWBL1rVfUQPZuBSJ3CiVov8x25A9jJ3rs/zJ 18kCp1Jlnp9Qc/vtFPwFXev8m9YSHe0H8MlyoS6B1EcyTLtBEzL7KynZIpLBUEgK QHHQ+vH740hHRs2Xf3IWj69zMcNhy+65sTOxA6HUuO4ZdCFsynsS0KMR5XzKGwuY AomisVdKSzAVvrrzGzZJUJ236WcIesIF7hk746orM7V62PQtuodJOohQzDUZbMNw bXM5xTjkO/aeTBDln//CgFMgXkksAP40rgYFZVtQONUAqEiWPkC8DtBQy2eWEpBM /jvnX8f8lRv+mTb6MZ/h1fqDzq1m1Zp+plKtq9DWtS18IN1rRzwCqgdVx56ybye3 QVjrkk7vLQfbFBFi9vIi =aIXX -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
On Tue, Dec 01, Jeff Mahoney wrote:
Did you look at the fallout from changing the page size to 4k? There will be significant performance impact on larger systems. Transparent huge pages are disabled now. The kernel will need to maintain 16 times as many page table entries as with 4k pages. There's a reason we chose the 64k page size ages ago and I don't think two broken drivers is a good reason to throw performance out the window especially since that hardware won't exist on any non-Apple hardware.
Yes, but will anyone notice this theoretical impact, in TW? Olaf -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 12/1/15 9:50 AM, Olaf Hering wrote:
On Tue, Dec 01, Jeff Mahoney wrote:
Did you look at the fallout from changing the page size to 4k? There will be significant performance impact on larger systems. Transparent huge pages are disabled now. The kernel will need to maintain 16 times as many page table entries as with 4k pages. There's a reason we chose the 64k page size ages ago and I don't think two broken drivers is a good reason to throw performance out the window especially since that hardware won't exist on any non-Apple hardware.
Yes, but will anyone notice this theoretical impact, in TW?
It's not theoretical and, in any case, the onus is on you as the submitter of the change to show that it doesn't. - -Jeff - -- Jeff Mahoney SUSE Labs -----BEGIN PGP SIGNATURE----- Version: GnuPG/MacGPG2 v2.0.19 (Darwin) Comment: GPGTools - http://gpgtools.org iQIcBAEBAgAGBQJWXbQrAAoJEB57S2MheeWy3yUP/iepKtn+Egiq2IgoHwax14SY bb0ijI+jqFy4HZF9w+pfK7R8H7YfsvNeI0iAmY0KwvsSqjFau6TvmCK7KNzEqRSp PE2G1BTGpW/bX+6Ni2PwO9wKIH5e+9R9CcQ2tHGlf2l3IF5Mu0Ds2Qb3td3pjUdb 7jG7vwzK1IVPFHttD1M9oM5Hjdp+QKV+qjeoQbpTVlf7YKTQ6xi4HJwuyPG7n0fd Z4essrUKK2PqKW3NwrQPHZjthYrBDQ7NbIEwpIOhmtpl/wRYP9kWIbtn4RIhD6xD taIlGnj6fgnFZfDV1VhLYGmvrWPA7L09FlyRRhiNQu7tLbqMBEt3BDDZz0ZdpJOw Xu9KO8DKWZkS+Eiix21DlKwHXbVUxAXxS1fCHZpuyXOrnE12nYwOo6ahXxCMfwDG WRtMI6w87b+JMcTr5+DXu+FvnlS2RZ3KwL4j3EZuhm9++OS/EAzzspeUJs/iaT4M i1GTT5otANNhWfnAsOI7FTrbuKLeN/1waLweecY9JPW1BPaJvElGlrw8VH8opcfI OLlRbkD47nF8GL6zb3CWSD5CQPNFcwQ9D7Wn+TyAgmL5WKkGAFl+B/OF2BBIykpj zTIwFI/ebZQHKCOdX0Hmu9SleX+NcDJ1nLfH6yglhNTn4sx0kJZ6ftnqOVEhFyBY 97hvxqhJQTjrUhvN4kRX =iJfM -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org
participants (2)
-
Jeff Mahoney -
Olaf Hering