Open Source & Linux Lab

It's better when it's simple

User Tools

Site Tools


etc:users:jcmvbkbc:linux-xtensa:esp32s3

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
etc:users:jcmvbkbc:linux-xtensa:esp32s3 [2024/02/23 13:36] jcmvbkbcetc:users:jcmvbkbc:linux-xtensa:esp32s3 [2024/03/15 21:07] (current) jcmvbkbc
Line 37: Line 37:
   * perf stat (heavy patching is needed to build it for nommu). -D1 is needed to properly enable events (perf relies on ability to run code after the fork but before the exec in the child process to manage events on systems with mmu, -D1 looks like a good workaround for nommu case).   * perf stat (heavy patching is needed to build it for nommu). -D1 is needed to properly enable events (perf relies on ability to run code after the fork but before the exec in the child process to manage events on systems with mmu, -D1 looks like a good workaround for nommu case).
   * c++ exceptions, c cleanup routines, forced stack unwinding, unwinding over signal frames.   * c++ exceptions, c cleanup routines, forced stack unwinding, unwinding over signal frames.
 +  * TLS and NPTL. Some corner cases still need attention though.
  
 ===== Things that don't work ===== ===== Things that don't work =====
  
   * about 290 failing tests in the gcc testsuite, most related to TLS and linuxthreads limitations. About 300 failing tests in the g++ testsuite.   * about 290 failing tests in the gcc testsuite, most related to TLS and linuxthreads limitations. About 300 failing tests in the g++ testsuite.
-  * NPTL. Not yet. Need to define relevant TLS relocations, teach the compiler to use them, the assembler to understand them and the linker to relax them. 
   * mmap with MAP_FIXED flag. By design of the nommu linux, but it seems to me that it doesn't have to be like that. It's usually not a big deal, but that's the reason there's no module information in the /proc/*/maps other than for the ld.so and the executable.   * mmap with MAP_FIXED flag. By design of the nommu linux, but it seems to me that it doesn't have to be like that. It's usually not a big deal, but that's the reason there's no module information in the /proc/*/maps other than for the ld.so and the executable.
   * tcpdump and libpcap in general. It tries to mmap the packet socket and it's missing a few things (mm/nommu.c doesn't know what capabilities to assign to S_IFSOCK files, the socket file needs to have get_unmapped_area callback defined), but most importantly the packet_mmap code heavily relies on the presence of mmu. Although it seems that it could be worked around.   * tcpdump and libpcap in general. It tries to mmap the packet socket and it's missing a few things (mm/nommu.c doesn't know what capabilities to assign to S_IFSOCK files, the socket file needs to have get_unmapped_area callback defined), but most importantly the packet_mmap code heavily relies on the presence of mmu. Although it seems that it could be worked around.
etc/users/jcmvbkbc/linux-xtensa/esp32s3.1708684611.txt.gz · Last modified: 2024/02/23 13:36 by jcmvbkbc