bug-hurd
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Clisp problem with reserved ranges during startup


From: Flávio Cruz
Subject: Clisp problem with reserved ranges during startup
Date: Wed, 5 Oct 2022 01:53:17 -0400

Hi

I have compiled Clisp in Debian GNU/Hurd but get a warning message during Clisp startup since the mmapable address ranges configured for UNIX_HURD seem to be incorrect:

Warning: reserving address range 0x18060000...0xbfffffff that contains memory mappings. clisp might crash later!
Memory dump:
  0x0 - 0xfff
  0x1000 - 0x39fff
  0x3a000 - 0x3bfff
  0x3c000 - 0x3cfff
  0x3d000 - 0x3dfff
  0x3e000 - 0x103dfff
  ... <- many ranges here
  0x830c000 - 0x832efff
  0x18000000 - 0x1805ffff
  0x20000000 - 0x20020fff
  0x20021000 - 0x20021fff
  0x20022000 - 0x27ffffff

In src/lispbibl.d we have the following:

/* On Hurd/i386:
       MMAP_FIXED_ADDRESS_HIGHEST_BIT = 30
       CODE_ADDRESS_RANGE   = 0x00000000UL or 0x08000000UL
       MALLOC_ADDRESS_RANGE = 0x08000000UL
       SHLIB_ADDRESS_RANGE  = 0x01000000UL
       STACK_ADDRESS_RANGE  = 0x01000000UL
       Addresses >= 0xC0000000UL are not mmapable.
       There is room from 0x11000000UL to 0xBFFFFFFFUL, but let's keep some
       distance. */
    #define MAPPABLE_ADDRESS_RANGE_START 0x18000000UL
    #define MAPPABLE_ADDRESS_RANGE_END   0xBFFFFFFFUL

Either something changed in Clisp that would cause memory to be allocated at 0x18000000UL or maybe the memory layout has changed. 

Dumping the memory addresses for a bare bones C program shows a very similar memory layout so I suspect the definitions are incorrect. Can anybody confirm that and where to look?

Otherwise using 0x28000000UL for MAPPABLE_ADDRESS_RANGE_START will work.

Thanks

reply via email to

[Prev in Thread] Current Thread [Next in Thread]