On Tue, Jan 02, 2007 at 12:04:49PM -0600, Brian Ford wrote: >As suggested here: > >http://cygwin.com/ml/cygwin/2006-12/msg00911.html > >2007-01-02 Brian Ford <[EMAIL PROTECTED]> > > * fhandler.cc (fhandler_base::fstat): Use system page size (64k) > as the st_blksize prefered I/O size for improved performance. > * fhandler_disk_file.cc (fhandler_base::fstat_helper): Likewise.
I don't see how replacing the constant "S_BLKSIZE" with what seems to be an unrelated getpagesize () makes a lot of sense. Assuming that this is a good idea, should S_BLKSIZE be changed directly? cgf >Index: fhandler.cc >=================================================================== >RCS file: /cvs/src/src/winsup/cygwin/fhandler.cc,v >retrieving revision 1.273 >diff -u -p -r1.273 fhandler.cc >--- fhandler.cc 11 Dec 2006 18:55:28 -0000 1.273 >+++ fhandler.cc 2 Jan 2007 17:55:07 -0000 >@@ -1328,7 +1328,7 @@ fhandler_base::fstat (struct __stat64 *b > buf->st_uid = geteuid32 (); > buf->st_gid = getegid32 (); > buf->st_nlink = 1; >- buf->st_blksize = S_BLKSIZE; >+ buf->st_blksize = getpagesize (); > time_as_timestruc_t (&buf->st_ctim); > buf->st_atim = buf->st_mtim = buf->st_ctim; > return 0; >Index: fhandler_disk_file.cc >=================================================================== >RCS file: /cvs/src/src/winsup/cygwin/fhandler_disk_file.cc,v >retrieving revision 1.200 >diff -u -p -r1.200 fhandler_disk_file.cc >--- fhandler_disk_file.cc 21 Dec 2006 10:59:47 -0000 1.200 >+++ fhandler_disk_file.cc 2 Jan 2007 17:55:07 -0000 >@@ -436,7 +436,7 @@ fhandler_base::fstat_helper (struct __st > else > buf->st_ino = get_namehash (); > >- buf->st_blksize = S_BLKSIZE; >+ buf->st_blksize = getpagesize (); > > if (nAllocSize >= 0LL) > /* A successful NtQueryInformationFile returns the allocation size
