happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

llama : Extend fallback, fix fileno for dio file, exclude case that mmap uses dio file (#18887)

This commit is contained in:
Julius Tischbein 2026-01-18 17:35:57 +01:00 committed by GitHub
parent 293a1565dc
commit 287a33017b
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 17 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
src/llama-mmap.cpp
View File

@ -265,7 +265,8 @@ struct llama_file::impl {
continue; // Interrupted by signal, retry
}
// Fallback to std::fread in case the DMA controller cannot access the buffer
if (errno == EFAULT) {
if (errno == EFAULT || errno == EINVAL) {
LLAMA_LOG_WARN("%s: Falling back to buffered IO due to %s\n", __func__, strerror(errno));
auto curr_off = tell();
close(fd);
fd = -1;
@ -384,6 +385,9 @@ int llama_file::file_id() const {
#ifdef _WIN32
return _fileno(pimpl->fp);
#else
if (pimpl->fd != -1) {
return pimpl->fd;
}
#if defined(fileno)
return fileno(pimpl->fp);
#else

18
src/llama-model-loader.cpp
View File

@ -539,12 +539,18 @@ llama_model_loader::llama_model_loader(
files.emplace_back(new llama_file(fname.c_str(), "rb", use_direct_io));
contexts.emplace_back(ctx);
use_direct_io = use_direct_io && files.back()->has_direct_io();
// Disable mmap in case Direct I/O is enabled and available
if (use_direct_io && use_mmap) {
use_mmap = false;
LLAMA_LOG_WARN("%s: direct I/O is enabled, disabling mmap\n", __func__);
if (use_mmap && use_direct_io) {
if (files.back()->has_direct_io()) {
// Disable mmap, as DirectIO is available
use_mmap = false;
LLAMA_LOG_WARN("%s: direct I/O is enabled, disabling mmap\n", __func__);
} else {
// Disable DirectIO and reopen file using std::fopen for mmap
use_direct_io = false;
files.pop_back();
files.emplace_back(new llama_file(fname.c_str(), "rb", false));
LLAMA_LOG_WARN("%s: direct I/O is not available, using mmap\n", __func__);
}
}
// Save tensors data offset of the main file.