Merge remote-tracking branch 'origin/master' into dev-talesam

Author: talesam

README.md

@@ -13,7 +13,7 @@ Smart dynamic swap management for Linux, written in Rust.
 
 ## Installation
 
-### Arch Linux / BigLinux
+### Arch Linux / BigLinux / Manjaro...
 ```bash
 cd pkgbuild
 makepkg -si

include/swap-default.conf

@@ -18,19 +18,35 @@ swap_mode=auto
 
 ################################################################################
 # Zswap Settings (used in zswap+swapfc mode)
+# This space is used for uncompressed memory. Since we have around a 3x compression ratio,
+# a system with 8 GB of RAM and 45% compressed memory usage can reach roughly 10.8 GB of
+# swap in RAM.
+#
+# However, keep in mind that 3.6 GB of RAM is no longer actually being used as regular RAM,
+# but rather as compressed cache. In practice, this is equivalent to around 15.2 GB total
+# (4.4 GB of RAM + 10.8 GB of swap compressed in RAM).
+#
+# 45% is a high value, but the kernel automatically manages when it is better to move swap
+# to storage, freeing up RAM, or when it is better to simply deallocate this memory.
+#
+# In addition, the system can automatically create swap files, allowing it to manage when
+# to move swap to storage. This makes it possible to use a few more gigabytes of swap when needed.
 ################################################################################
 
 zswap_compressor=zstd            # Compression algo: zstd, lz4, lzo
-zswap_max_pool_percent=60        # Max RAM for compressed pool
+zswap_max_pool_percent=45        # Max RAM for compressed pool
 zswap_zpool=zsmalloc             # Allocator
 zswap_shrinker_enabled=1         # Proactively move cold pages to disk
 zswap_accept_threshold=90        # Restart accepting pages when pool drops to X%
 
 ################################################################################
 # Zram Settings (used in zram modes)
+# Unlike zswap, with zram the reserved space corresponds to the virtual size of the
+# swap being created, not to a usage limit based on a percentage of physical memory.
+# Therefore, 90% in zram does not represent a higher virtual swap value than 45% in zswap.
 ################################################################################
 
-zram_size=80%                    # Size of zram device (e.g., 50%, 1G)
+zram_size=90%                    # Size of zram device (e.g., 50%, 1G)
 zram_alg=zstd                    # Compression algo
 zram_prio=32767                  # Priority (highest)
 
@@ -46,28 +62,25 @@ zram_writeback_threshold=50      # Trigger when zram is X% full
 ################################################################################
 
 # Progressive Scaling: Start small, grow as needed
-# e.g. 256M -> 256M -> 512M -> 512M... doubles every 4 files
 swapfc_enabled=1                 # Enable swap files
 swapfc_path=/swapfc/swapfile     # Directory and filename prefix
-swapfc_chunk_size=256M           # Base size
-swapfc_max_chunk_size=32G        # Max single file size
-swapfc_max_count=28              # Max number of files
+swapfc_chunk_size=512M           # Base size
+swapfc_max_chunk_size=64G        # Max single file size ( Theoretically be up to 16 TB or more )
+swapfc_max_count=32              # Max number of files ( Kernel limit is 32 )
 swapfc_scaling_step=4            # Double size every X files
+                                 # Doubles every 4 files e.g. 512M -> 512M -> 512M -> 512M -> 1024M -> 1024M...
 
 # Triggers
 swapfc_free_ram_perc=35          # Create swap when free RAM < X%
 swapfc_free_swap_perc=25         # Create more when free swap < X%
 swapfc_remove_free_swap_perc=55  # Remove files when free swap > X%
 
-################################################################################
-# Advanced / Experimental
-################################################################################
-
 # MGLRU Anti-Thrashing (Kernel 6.1+)
 # Protect working set for X ms. 0 to disable.
 mglru_min_ttl_ms=1000
 
 # Sparse Files
 # Create files that only use disk space when actually written to.
-# Highly recommended for Zswap mode.
-# swapfc_use_sparse_disable=1    # Uncomment to disable (pre-allocate all space)
+# With pre-allocated space there should be a performance gain, however
+# more storage space will be reserved for swap, so the performance gain may not be noticeable.
+# swapfc_use_sparse_disable=1    # Uncomment to pre-allocate all space