csullivan commented on code in PR #13381:
URL: https://github.com/apache/tvm/pull/13381#discussion_r1023139210
##########
src/runtime/hexagon/hexagon_buffer.cc:
##########
@@ -237,8 +236,15 @@ void hexagon_buffer_copy_across_regions(const BufferSet&
dest, const BufferSet&
// Finally, do the memory copies.
for (const auto& copy : macro_copies) {
- int error_code = hexagon_user_dma_1d_sync(copy.dest, copy.src,
copy.num_bytes);
- CHECK_EQ(error_code, 0);
+ qurt_mem_cache_clean(reinterpret_cast<qurt_addr_t>(copy.dest),
copy.num_bytes,
+ QURT_MEM_CACHE_INVALIDATE, QURT_MEM_DCACHE);
+ qurt_mem_cache_clean(reinterpret_cast<qurt_addr_t>(copy.src),
copy.num_bytes,
+ QURT_MEM_CACHE_INVALIDATE, QURT_MEM_DCACHE);
+ memcpy(copy.dest, copy.src, copy.num_bytes);
+ qurt_mem_cache_clean(reinterpret_cast<qurt_addr_t>(copy.dest),
copy.num_bytes,
+ QURT_MEM_CACHE_INVALIDATE, QURT_MEM_DCACHE);
+ qurt_mem_cache_clean(reinterpret_cast<qurt_addr_t>(copy.src),
copy.num_bytes,
+ QURT_MEM_CACHE_INVALIDATE, QURT_MEM_DCACHE);
Review Comment:
Thank you! We will definitely need to revisit this down the road so leaving
bread crumbs to help us reload the caches once we are ready will be very
helpful!
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscr...@tvm.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
