On 07/01/2015 12:19 PM, John Snow wrote: > requires: 1434470575-21625-1-git-send-email-js...@redhat.com > 1435016308-6150-1-git-send-email-js...@redhat.com > [PATCH v2 0/4] ahci: misc fixes/tests for 2.4 > [PATCH v2 00/16] ahci: ncq cleanup, part 1 > > This chunk gets NCQ migration and and resume support working. > > There's still some left to do, particularly around error handling > and FIS semantics, but this should get us most of the way there. > > === > v2: > === > > Key: > [----] : patches are identical > [####] : number of functional differences between upstream/downstream patch > [down] : patch is downstream-only > The flags [FC] indicate (F)unctional and (C)ontextual differences, > respectively > > 001/15:[0043] [FC] 'ide: add limit to .prepare_buf()' > 002/15:[----] [--] 'ahci: stash ncq command' > 003/15:[----] [--] 'ahci: assert is_ncq for process_ncq' > 004/15:[----] [--] 'ahci: refactor process_ncq_command' > 005/15:[----] [--] 'ahci: factor ncq_finish out of ncq_cb' > 006/15:[down] 'ahci: add rwerror=stop support for ncq' > 007/15:[----] [--] 'ahci: correct types in NCQTransferState' > 008/15:[----] [--] 'ahci: correct ncq sector count' > 009/15:[----] [--] 'qtest/ahci: halted NCQ test' > 010/15:[----] [-C] 'ahci: add cmd header to ncq transfer state' > 011/15:[0004] [FC] 'ahci: add get_cmd_header helper' > 012/15:[0006] [FC] 'ahci: ncq migration' > 013/15:[----] [--] 'ahci: Do not map cmd_fis to generate response' > 014/15:[----] [--] 'qtest/ahci: halted ncq migration test' > 015/15:[0001] [FC] 'ahci: fix sdb fis semantics' > > 01: Fixed limit parameter to int32_t to match existing implicit cap > Removed is_write parameter. Nobody used it. > Implemented PRD limit for BMDMA > 06: Squashed what was patch 06/07. > Adjusted where we clear ncq_tfs->halt. > 11: Fixed port/slot bounds checking > Removed ncq_tfs section from ahci_state_post_load > moved ahead of what's now #12. > 12: Picked up the truncated sections from what's now #11 > Rely on get_cmd_header to do more error checking for us. > 16: Comment explaining the IRQ mechanics for SDBFIS. > > ________________________________________________________________________________ > > For convenience, this branch is available at: > https://github.com/jnsnow/qemu.git branch ahci-ncq-s2 > https://github.com/jnsnow/qemu/tree/ahci-ncq-s2 > > This version is tagged ahci-ncq-s2-v2: > https://github.com/jnsnow/qemu/releases/tag/ahci-ncq-s2-v2 > > John Snow (15): > ide: add limit to .prepare_buf() > ahci: stash ncq command > ahci: assert is_ncq for process_ncq > ahci: refactor process_ncq_command > ahci: factor ncq_finish out of ncq_cb > ahci: add rwerror=stop support for ncq > ahci: correct types in NCQTransferState > ahci: correct ncq sector count > qtest/ahci: halted NCQ test > ahci: add cmd header to ncq transfer state > ahci: add get_cmd_header helper > ahci: ncq migration > ahci: Do not map cmd_fis to generate response > qtest/ahci: halted ncq migration test > ahci: fix sdb fis semantics > > hw/ide/ahci.c | 334 > +++++++++++++++++++++++++++++++++--------------------- > hw/ide/ahci.h | 9 +- > hw/ide/core.c | 15 ++- > hw/ide/internal.h | 4 +- > hw/ide/macio.c | 2 +- > hw/ide/pci.c | 21 +++- > tests/ahci-test.c | 38 +++++-- > 7 files changed, 269 insertions(+), 154 deletions(-) >
Thanks, applied to my IDE tree: https://github.com/jnsnow/qemu/commits/ide https://github.com/jnsnow/qemu.git --js