adjust = gsi->version virt + adjust + GSI_INTER_EE_SRC_CH_IRQ_OFFSET) * Reverse the offset adjustment for inter-EE register offsets */ Iowrite32(0, gsi->virt + GSI_CNTXT_SRC_IEOB_IRQ_MSK_OFFSET) Iowrite32(0, gsi->virt + GSI_CNTXT_GLOB_IRQ_EN_OFFSET) Gsi_irq_type_update(gsi, -206,10 +204,9 static void gsi_irq_setup(struct gsi *gsi) Static void gsi_irq_setup(struct gsi *gsi) * Turn off all GSI interrupts initially */ +++ -195,8 +195,6 static void gsi_irq_type_disable(struct gsi *gsi, enum gsi_irq_type_id type_id) That's an adjusted copy of the "raw" pointer, and that is used forĭrivers/net/ipa/gsi.c | 26 +++++++++++-ĭrivers/net/ipa/gsi_reg.h | 19 ++++++++++++-ģ files changed, 25 insertions(+), 23 deletions(-)ĭiff -git a/drivers/net/ipa/gsi.c b/drivers/net/ipa/gsi.c Rather than adjusting *that* pointer, we maintain a separate pointer Used to access the two registers that do *not* sit in the "adjusted" Memory is mapped and used to unmap the memory. This patch fixes that error, but it does so by maintaining one "raw" The bug is that the pointer is not "de-adjusted" as it Handled by adjusting the I/O virtual address pointer after it hasīeen mapped. Registers is adjusted downward by a fixed amount. The fix implements GSI memoryįor IPA version 4.5 and above, the address space for almost all GSI This patch actually fixes a bug, though it doesn't affect the two ` (3 subsequent siblings) 4 siblings, 0 replies 6+ messages in thread 22:33 ` net: ipa: don't report EPROBE_DEFER error Alex Elder 22:33 net: ipa: some more cleanup Alex 22:33 ` Alex Elder * net: ipa: use a separate pointer for adjusted GSI memory Net: ipa: introduce ipa_table_hash_support() Net: ipa: fix register write command validation Net: ipa: don't report EPROBE_DEFER error Net: ipa: use a separate pointer for adjusted GSI memory The last two just create helper functions to improve readability. The third fixes some errors that show up when IPA_VALIDATION is The second just quiets some unnecessary noise during early probe. Mapped) address, and the other will have been adjusted if necessary It maintain a new GSI address pointer one is the "raw" (original The first was motivated by a bug found that would affect IPA v4.5. This is another fairly innocuous set of cleanup patches. ` (4 more replies) 0 siblings, 5 replies 6+ messages in threadĬc: elder, evgreen, bjorn.andersson, cpratapa, subashab, netdev, 22:33 ` net: ipa: use a separate pointer for adjusted GSI memory Alex Elder Net: ipa: some more cleanup LKML Archive on help / color / mirror / Atom feed * net: ipa: some more cleanup 22:33 Alex Elder
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |