Skip to contentSkip to conversation

Do we care about 4k stacks?

uml-devel Rob Landley 3 participants 4 messages Nov 21, 2005 – Nov 23, 2005 User-mode Linux (UML)

lore
Share & tools
lore
#0 lore raw

Just curious if this is at all relevant to us, in light of the "yank 8k stacks" patch proposed...

$ make checkstack
objdump -d vmlinux $(find . -name '*.ko') | \
perl /home/landley/linux-2.6.15-rc1/scripts/checkstack.pl i386
0xa010ddcf buffered_vfprintf:                           33168
0xa010deca buffered_vfprintf:                           33168
0xa00e5d5b buffered_vfprintf:                           8416
0xa00e5e43 buffered_vfprintf:                           8416
0xa0121ffb __res_vinit:                                 8256
0xa012293a __res_vinit:                                 8256
0xa011199c phys_pages_info:                             8212
0xa01117bf __get_nprocs:                                8208
0xa010a426 _IO_vfwprintf:                               4376
0xa00ebe33 _IO_wfile_seekoff:                           4280
0xa00ec0b9 _IO_wfile_seekoff:                           4280
0xa00f8626 __unix_grantpt:                              4252
0xa00f848f grantpt:                                     4188
0xa01103c6 execle:                                      4140
0xa0005119 make_uml_dir:                                4124
0xa00051bd make_uml_dir:                                4124
0xa0023dc0 make_tempfile:                               4124
0xa0023e5a make_tempfile:                               4124
0xa00e01d7 openpty:                                     4124
0xa011cd40 _dl_get_origin:                              4112
0xa011cdc4 _dl_get_origin:                              4112

Rob

-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc. Get Certified Today Register for a JBoss Training Course. Free Certification Exam for All Training Attendees Through End of 2005. For more info visit:
http://ads.osdn.com/?ad_id=7628&alloc_id=16845&op=click
_______________________________________________
User-mode-linux-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

#1 lore raw
Quoted 2 lines · Rob Landley

Just curious if this is at all relevant to us, in light of the "yank 8k stacks" patch proposed...

That's just on i386. Each arch defines its own stack size, and UML uses 16K. Not that that's good, so we might be following i386 at some point, at least to 8K.

Quoted 7 lines
$ make checkstack
objdump -d vmlinux $(find . -name '*.ko') | \
perl /home/landley/linux-2.6.15-rc1/scripts/checkstack.pl i386
0xa010ddcf buffered_vfprintf:                           33168
0xa010deca buffered_vfprintf:                           33168
0xa00e5d5b buffered_vfprintf:                           8416
0xa00e5e43 buffered_vfprintf:                           8416

I don't think these are of concern, but feel free to double-check :-)

If you look at the code, you'll see some care in making sure that certain things (like printf) run only on the initial process stack, which is expanded as needed by the host, and not on a UML kernel, which isn't.

Jeff

-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc. Get Certified Today Register for a JBoss Training Course. Free Certification Exam for All Training Attendees Through End of 2005. For more info visit:
http://ads.osdn.com/?ad_id=7628&alloc_id=16845&op=click
_______________________________________________
User-mode-linux-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

#2 lore raw
In reply to Rob Landley
Quoted 6 lines · Rob Landley

Just curious if this is at all relevant to us, in light of the "yank 8k stacks" patch proposed...

$ make checkstack
objdump -d vmlinux $(find . -name '*.ko') | \
perl /home/landley/linux-2.6.15-rc1/scripts/checkstack.pl i386

That's _SO_ much? Ouch... I now understand a lot more why Al Viro removed printf from UML usage (I tried to use printf rather than printk to report that a console channel is disabled; I later solved the problem with the kernel sprintf + a direct write() on stdout).

Quoted 6 lines
0xa010ddcf buffered_vfprintf:                           33168
0xa010deca buffered_vfprintf:                           33168
0xa00e5d5b buffered_vfprintf:                           8416
0xa00e5e43 buffered_vfprintf:                           8416
0xa0121ffb __res_vinit:                                 8256
0xa012293a __res_vinit:                                 8256
Quoted 8 lines
0xa011199c phys_pages_info:                             8212
0xa01117bf __get_nprocs:                                8208
0xa010a426 _IO_vfwprintf:                               4376
0xa00ebe33 _IO_wfile_seekoff:                           4280
0xa00ec0b9 _IO_wfile_seekoff:                           4280
0xa00f8626 __unix_grantpt:                              4252
0xa00f848f grantpt:                                     4188
0xa01103c6 execle:                                      4140

Ok, the two below functions are plain dumb - there's a MAXPATHLEN (i.e. PATH_MAX i.e. 4096) array on the stack, and they could happily be fixed.

All the rest seems plain glibc code and thus unfixable.

Quoted 10 lines
0xa0005119 make_uml_dir:                                4124
0xa00051bd make_uml_dir:                                4124
0xa0023dc0 make_tempfile:                               4124
0xa0023e5a make_tempfile:                               4124
0xa00e01d7 openpty:                                     4124
0xa011cd40 _dl_get_origin:                              4112
0xa011cdc4 _dl_get_origin:                              4112

Rob

Signature · 24 lines
Inform me of my mistakes, so I can keep imitating Homer Simpson's "Doh!".
Paolo Giarrusso, aka Blaisorblade (Skype ID "PaoloGiarrusso", ICQ 215621894)
http://www.user-mode-linux.org/~blaisorblade


	

	
		
___________________________________ 
Yahoo! Mail: gratis 1GB per i messaggi e allegati da 10MB 
http://mail.yahoo.it



-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.  Get Certified Today
Register for a JBoss Training Course.  Free Certification Exam
for All Training Attendees Through End of 2005. For more info visit:
http://ads.osdn.com/?ad_id=7628&alloc_id=16845&op=click
_______________________________________________
User-mode-linux-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
#3 lore raw
Quoted 4 lines · Blaisorblade

Ok, the two below functions are plain dumb - there's a MAXPATHLEN (i.e. PATH_MAX i.e. 4096) array on the stack, and they could happily be fixed.

All the rest seems plain glibc code and thus unfixable.

I can link against uClibc and see what the result is... :)

Rob

-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc. Get Certified Today Register for a JBoss Training Course. Free Certification Exam for All Training Attendees Through End of 2005. For more info visit:
http://ads.osdn.com/?ad_id=7628&alloc_id=16845&op=click
_______________________________________________
User-mode-linux-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel