diff options
author | Johannes Berg <johannes.berg@intel.com> | 2019-12-05 09:31:07 +0100 |
---|---|---|
committer | Luca Coelho <luciano.coelho@intel.com> | 2019-12-23 11:33:04 +0200 |
commit | d84a7a654a66eead599cfd4f436d1f921e01074f (patch) | |
tree | 7cae2d5b2e73c5b2981c71d82519210880881e9d /drivers/net/wireless/intel/iwlwifi/pcie/tx-gen2.c | |
parent | 2763bba6328c53c455d8f7f5302b80030551c31b (diff) | |
download | linux-d84a7a654a66eead599cfd4f436d1f921e01074f.tar.gz linux-d84a7a654a66eead599cfd4f436d1f921e01074f.tar.bz2 linux-d84a7a654a66eead599cfd4f436d1f921e01074f.zip |
iwlwifi: pcie: extend hardware workaround to context-info
After more investigation on the hardware side, it appears that the
hardware bug regarding 2^32 boundary reaching/crossing also affects
other uses of the DMA engine, in particular the ones triggered by
the context-info (image loader) mechanism.
It also turns out that the bug only affects devices with gen2 TX
hardware engine, so we don't need to change context info for gen3.
The TX path workarounds are simpler to still keep for both though.
Add the workaround to that code as well; this is a lot simpler as
we have just a single way to allocate DMA memory there.
I made the algorithm recursive (with a small limit) since it's
actually (almost) impossible to hit this today - dma_alloc_coherent
is currently documented to always return 32-bit addressable memory
regardless of the DMA mask for it, and so we could only get REALLY
unlucky to get the very last page in that area.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Diffstat (limited to 'drivers/net/wireless/intel/iwlwifi/pcie/tx-gen2.c')
-rw-r--r-- | drivers/net/wireless/intel/iwlwifi/pcie/tx-gen2.c | 14 |
1 files changed, 2 insertions, 12 deletions
diff --git a/drivers/net/wireless/intel/iwlwifi/pcie/tx-gen2.c b/drivers/net/wireless/intel/iwlwifi/pcie/tx-gen2.c index 56d752beb940..bfb984b2e00c 100644 --- a/drivers/net/wireless/intel/iwlwifi/pcie/tx-gen2.c +++ b/drivers/net/wireless/intel/iwlwifi/pcie/tx-gen2.c @@ -213,16 +213,6 @@ static void iwl_pcie_gen2_free_tfd(struct iwl_trans *trans, struct iwl_txq *txq) } } -/* - * We need this inline in case dma_addr_t is only 32-bits - since the - * hardware is always 64-bit, the issue can still occur in that case, - * so use u64 for 'phys' here to force the addition in 64-bit. - */ -static inline bool crosses_4g_boundary(u64 phys, u16 len) -{ - return upper_32_bits(phys) != upper_32_bits(phys + len); -} - static int iwl_pcie_gen2_set_tb(struct iwl_trans *trans, struct iwl_tfh_tfd *tfd, dma_addr_t addr, u16 len) @@ -238,7 +228,7 @@ static int iwl_pcie_gen2_set_tb(struct iwl_trans *trans, * there's no more space, and so when we know there is enough we * don't always check ... */ - WARN(crosses_4g_boundary(addr, len), + WARN(iwl_pcie_crosses_4g_boundary(addr, len), "possible DMA problem with iova:0x%llx, len:%d\n", (unsigned long long)addr, len); @@ -300,7 +290,7 @@ static int iwl_pcie_gen2_set_tb_with_wa(struct iwl_trans *trans, if (unlikely(dma_mapping_error(trans->dev, phys))) return -ENOMEM; - if (likely(!crosses_4g_boundary(phys, len))) { + if (likely(!iwl_pcie_crosses_4g_boundary(phys, len))) { ret = iwl_pcie_gen2_set_tb(trans, tfd, phys, len); if (ret < 0) |